Application in Electrical Power Systems

*Application of Expert Systems - Theoretical and Practical Aspects*

[15] Satuf E et al. Situation awareness measurement of an ecological

interface designed to operator support during alarm floods. International Journal of Industrial Ergonomics. Elsevier. 2016:179-192. DOI: 10.1016/j.

ergon.2016.01.002

**66**

**69**

**Chapter 6**

**Abstract**

**1. Introduction**

Intelligent System for the

Physicochemical Tests

*and Bruno Augusto Trevisam*

Estimation of Gases Dissolved

in Insulating Mineral Oil from

The objective of this work was to make the modeling through artificial neural networks of the gas concentrations dissolved in insulating mineral oil from the results of physicochemical tests. In this case, a mapping between the data of physicochemical tests and gas chromatography was obtained by means of artificial neural networks. The proposed approach proved to be efficient to identify the amount of gases, taking the following attributes as input: color degree, density, dielectric rigidity, interfacial tension, power factor of the insulating oil, neutralization index, and water level. In addition, artificial neural networks provide not only a new methodology to support decisions but also satisfactory results comparatively

*Ivan Nunes da Silva, Rogério Andrade Flauzino,* 

*Danilo Hernane Spatti, Renato Pagotto Bossolan* 

to actual analyses when referring to the estimation of gases.

consists essentially of insulating paper and insulating mineral oil.

**Keywords:** intelligent systems, artificial neural networks, power transformer

The power transformer is an extremely important and expensive equipment consisting, in short, of two or three windings, the core, and the insulation system. The windings are arranged such that the magnetic leakage flux is dispersed as little as possible. The core is the medium by which the magnetic flux finds a path of low reluctance and preferably flows through it. Among these elements, core, windings, and the transformer tank, there is also the insulation system. The insulation system

During its operation, various wear and aging processes may occur in the transformer insulation system. Examples can be hot spots, overheats, and overvoltages, among other phenomena that change the transformer insulation system [1]. Thus, the reduction in the lifetime of transformers is directly linked to the deterioration of the dielectric materials used in their manufacture. In this sense, there is an important motivating aspect for the development of supervision and preventive maintenance programs in order to provide an increase in the useful life of the equipment and, consequently, a better management of them under the command of the electric

#### **Chapter 6**

## Intelligent System for the Estimation of Gases Dissolved in Insulating Mineral Oil from Physicochemical Tests

*Ivan Nunes da Silva, Rogério Andrade Flauzino, Danilo Hernane Spatti, Renato Pagotto Bossolan and Bruno Augusto Trevisam*

#### **Abstract**

The objective of this work was to make the modeling through artificial neural networks of the gas concentrations dissolved in insulating mineral oil from the results of physicochemical tests. In this case, a mapping between the data of physicochemical tests and gas chromatography was obtained by means of artificial neural networks. The proposed approach proved to be efficient to identify the amount of gases, taking the following attributes as input: color degree, density, dielectric rigidity, interfacial tension, power factor of the insulating oil, neutralization index, and water level. In addition, artificial neural networks provide not only a new methodology to support decisions but also satisfactory results comparatively to actual analyses when referring to the estimation of gases.

**Keywords:** intelligent systems, artificial neural networks, power transformer

#### **1. Introduction**

The power transformer is an extremely important and expensive equipment consisting, in short, of two or three windings, the core, and the insulation system. The windings are arranged such that the magnetic leakage flux is dispersed as little as possible. The core is the medium by which the magnetic flux finds a path of low reluctance and preferably flows through it. Among these elements, core, windings, and the transformer tank, there is also the insulation system. The insulation system consists essentially of insulating paper and insulating mineral oil.

During its operation, various wear and aging processes may occur in the transformer insulation system. Examples can be hot spots, overheats, and overvoltages, among other phenomena that change the transformer insulation system [1]. Thus, the reduction in the lifetime of transformers is directly linked to the deterioration of the dielectric materials used in their manufacture. In this sense, there is an important motivating aspect for the development of supervision and preventive maintenance programs in order to provide an increase in the useful life of the equipment and, consequently, a better management of them under the command of the electric utilities. In fact, the electricity market is becoming increasingly competitive, and the high costs involved in maintenance require the development of processes in order to lower its costs, increase asset performance, as well as extend its useful life [2].

Among the methodologies used to identify failures in power transformers immersed in insulating mineral oil, we highlight those based on the monitoring of their electrical parameters, acoustic monitoring, and the evaluation of dissolved gas concentrations in oil. On the other hand, the condition associated with the insulating paper present in the power transmission transformer coils is one of the main responsible for the lifetime of this equipment. Winding insulation paper can deteriorate abruptly, resulting in unpredictable failure of these transformers, particularly if in contact with oxygen, humidity, and metal contaminants. All these possible scenarios can be avoided if the degradation process is discovered in time [3].

For identifying failures from the dissolved gases in the oil, several chromatographic tests are then performed, which result in their respective concentrations. The examination of such dissolved gases is based on the premise that failures from partial discharges, high energy discharges, corona effect, overheating, etc. react with the insulating mineral oil causing alterations in their physicochemical properties, which are responsible for the release of gases that eventually react with the oil and, consequently, are dissolved within it. Therefore, the dissolved gas analysis (DGA) in insulating mineral oil is one of the main methods used to diagnose the insulation situation of the power transmission transformer windings. In addition, proper interpretation of DGA results is one of the most significant procedures for detecting fault types, as well as for identifying the process of degradation of its insulation [4, 5].

In addition to the chromatographic tests, with which it is possible to obtain the volumetric values of the main dissolved gases in the oil, there are also physicochemical tests that allow to identify several physical and chemical characteristics present in the insulating mineral oil. An important feature associated with physicochemical tests is that they can all be performed at the location where the transformers are installed, and part of them can be performed with the equipment in service, i.e., without interruption in the supply of electricity. Moreover, when power transmission transformers are in service, they may be subject to a number of environmental factors, such as heat and humidity, which must lead to an increasing aging process over time. Consequently, the electrical properties of the insulating mineral oil (resistivity, dielectric losses, relative permittivity, and breakdown strength), as well as the chemical (acidity and humidity) and mechanical properties (tensile strength and viscosity), are also modified, which directly impact its useful life [6, 7].

Therefore, it is emphasized here the motivation of a system for estimation of gases dissolved in oil from the results of physicochemical tests. Thus, the objective would then be to adapt existing methodologies based on the concentration of dissolved gases to, based on physicochemical tests, predict possible incipient failures in power transformers.

Since the estimation of these results is a complex problem, which does not have a mathematical formula capable of relating both tests, an interesting alternative is the use of computational intelligence techniques. Among these techniques, we can highlight the artificial neural networks, which do not require detailed knowledge about the relationship between the input space and the output space, as they are expert in dealing with nonlinear mappings [8].

It should be emphasized that the great novelty of this work is the precise estimation of the concentration of gases that are dissolved in insulating mineral oil, through neural networks, from the results of physicochemical tests. For the estimation of gases, we then use data from physicochemical tests performed on insulating oil taken from transformers in operation. The developed technique allowed the

**71**

*Intelligent System for the Estimation of Gases Dissolved in Insulating Mineral Oil…*

estimation of these gases without the need to perform the chromatographic test. Consequently, the amount of power outages will be considerably reduced, and the associated costs in determining the gas concentration in the mineral oil will follow

For this purpose, the organization of this article was divided in five sections. In Section 2, several aspects related to the main experimental tests performed on the insulating mineral oil are described. In Section 3, the intelligent system based on artificial neural networks for mapping the relationship between both physicochemical and chromatographic tests is presented. The main results produced by intelligent system in order to validate developed methodology are reported in Section 4. In Section 5, we present the conclusions about the applicability of the

**2. Aspects related to the experimental tests in insulating mineral oil**

Depending on the type of petroleum used for fractional distillation, the insulating mineral oil may have paraffinic base (paraffinic oil) or naphthenic base (naphthenic oil). Its classification is based on the results of the percentage of paraffinic (PC), naphthenic (NC), and aromatic (AC) carbons. This classification can be obtained by using infrared spectroscopy techniques by determining the amount of paraffin carbons. Oils with PC lower than 50% are considered naphthenic, while those with PC equal to or greater than 56% are classified as paraffinic; between 50

When different mineral oils are mixed, the properties of the resulting mixture will be an average of the properties of each of the components, since the mixed oils are of good quality. If one of these components is of poor quality, the resulting oil

Determining the physicochemical properties of insulating mineral oil is of fundamental importance to ensure the operating conditions of transformers and to maintain or extend the lifetime of these equipments [9, 10]. In order to prevent premature degradation, which can be translated into financial loss, it is therefore necessary to monitor the physicochemical properties of the insulating mineral oil. Through the analysis of physicochemical characteristics, it is then possible to evaluate the quality of both the oil and the transformer itself. These analyses are of great importance for the maintenance of electrical equipments, as they allow planned interventions to correct any defects and, in some cases, early failure

To test the insulating mineral oil, it is necessary that the transformer is working, because in operation the oil goes through an aging process. This aging is due to the demand for temperature increase, to the action of oxygen, and to contact with materials present in its construction, resulting in deterioration of oil properties. This deterioration generates by-products that promote its acceleration, that is, it generates a chain reaction where the insulating mineral oil loses its insulating properties and, as a result, the cellulose degrades generating sludge. The process that governs the oxidation of hydrocarbons is the peroxidation mechanism (peroxide formation), where there is the formation of intermediate products, which may be alcohols, aldehydes, and ketones depending on the species that originated them. In **Table 1**, it showed the main experimental tests and their respective standards,

which are usually recommended to trace the transformer isolation conditions. Therefore, the various tests performed on the insulating mineral oil in use allow to diagnose some problems, such as hot spots, overheating, and leaks, as well as

informing about the insulating and thermal quality of the respective oil.

*DOI: http://dx.doi.org/10.5772/intechopen.91807*

the same downward trend.

proposed technique.

and 56% are intermediate oils.

will be of poor quality.

diagnoses [11, 12].

*Intelligent System for the Estimation of Gases Dissolved in Insulating Mineral Oil… DOI: http://dx.doi.org/10.5772/intechopen.91807*

estimation of these gases without the need to perform the chromatographic test. Consequently, the amount of power outages will be considerably reduced, and the associated costs in determining the gas concentration in the mineral oil will follow the same downward trend.

For this purpose, the organization of this article was divided in five sections. In Section 2, several aspects related to the main experimental tests performed on the insulating mineral oil are described. In Section 3, the intelligent system based on artificial neural networks for mapping the relationship between both physicochemical and chromatographic tests is presented. The main results produced by intelligent system in order to validate developed methodology are reported in Section 4. In Section 5, we present the conclusions about the applicability of the proposed technique.

#### **2. Aspects related to the experimental tests in insulating mineral oil**

Depending on the type of petroleum used for fractional distillation, the insulating mineral oil may have paraffinic base (paraffinic oil) or naphthenic base (naphthenic oil). Its classification is based on the results of the percentage of paraffinic (PC), naphthenic (NC), and aromatic (AC) carbons. This classification can be obtained by using infrared spectroscopy techniques by determining the amount of paraffin carbons. Oils with PC lower than 50% are considered naphthenic, while those with PC equal to or greater than 56% are classified as paraffinic; between 50 and 56% are intermediate oils.

When different mineral oils are mixed, the properties of the resulting mixture will be an average of the properties of each of the components, since the mixed oils are of good quality. If one of these components is of poor quality, the resulting oil will be of poor quality.

Determining the physicochemical properties of insulating mineral oil is of fundamental importance to ensure the operating conditions of transformers and to maintain or extend the lifetime of these equipments [9, 10]. In order to prevent premature degradation, which can be translated into financial loss, it is therefore necessary to monitor the physicochemical properties of the insulating mineral oil. Through the analysis of physicochemical characteristics, it is then possible to evaluate the quality of both the oil and the transformer itself. These analyses are of great importance for the maintenance of electrical equipments, as they allow planned interventions to correct any defects and, in some cases, early failure diagnoses [11, 12].

To test the insulating mineral oil, it is necessary that the transformer is working, because in operation the oil goes through an aging process. This aging is due to the demand for temperature increase, to the action of oxygen, and to contact with materials present in its construction, resulting in deterioration of oil properties. This deterioration generates by-products that promote its acceleration, that is, it generates a chain reaction where the insulating mineral oil loses its insulating properties and, as a result, the cellulose degrades generating sludge. The process that governs the oxidation of hydrocarbons is the peroxidation mechanism (peroxide formation), where there is the formation of intermediate products, which may be alcohols, aldehydes, and ketones depending on the species that originated them.

In **Table 1**, it showed the main experimental tests and their respective standards, which are usually recommended to trace the transformer isolation conditions. Therefore, the various tests performed on the insulating mineral oil in use allow to diagnose some problems, such as hot spots, overheating, and leaks, as well as informing about the insulating and thermal quality of the respective oil.

*Application of Expert Systems - Theoretical and Practical Aspects*

utilities. In fact, the electricity market is becoming increasingly competitive, and the high costs involved in maintenance require the development of processes in order to lower its costs, increase asset performance, as well as extend its useful life [2]. Among the methodologies used to identify failures in power transformers immersed in insulating mineral oil, we highlight those based on the monitoring of their electrical parameters, acoustic monitoring, and the evaluation of dissolved gas concentrations in oil. On the other hand, the condition associated with the insulating paper present in the power transmission transformer coils is one of the main responsible for the lifetime of this equipment. Winding insulation paper can deteriorate abruptly, resulting in unpredictable failure of these transformers, particularly if in contact with oxygen, humidity, and metal contaminants. All these possible scenarios can be avoided if the degradation process is discovered in time [3]. For identifying failures from the dissolved gases in the oil, several chromatographic tests are then performed, which result in their respective concentrations. The examination of such dissolved gases is based on the premise that failures from partial discharges, high energy discharges, corona effect, overheating, etc. react with the insulating mineral oil causing alterations in their physicochemical properties, which are responsible for the release of gases that eventually react with the oil and, consequently, are dissolved within it. Therefore, the dissolved gas analysis (DGA) in insulating mineral oil is one of the main methods used to diagnose the insulation situation of the power transmission transformer windings. In addition, proper interpretation of DGA results is one of the most significant procedures for detecting fault types, as well as for identifying the process of degradation of its

In addition to the chromatographic tests, with which it is possible to obtain the volumetric values of the main dissolved gases in the oil, there are also physicochemical tests that allow to identify several physical and chemical characteristics present in the insulating mineral oil. An important feature associated with physicochemical tests is that they can all be performed at the location where the transformers are installed, and part of them can be performed with the equipment in service, i.e., without interruption in the supply of electricity. Moreover, when power transmission transformers are in service, they may be subject to a number of environmental factors, such as heat and humidity, which must lead to an increasing aging process over time. Consequently, the electrical properties of the insulating mineral oil (resistivity, dielectric losses, relative permittivity, and breakdown strength), as well as the chemical (acidity and humidity) and mechanical properties (tensile strength

and viscosity), are also modified, which directly impact its useful life [6, 7].

Therefore, it is emphasized here the motivation of a system for estimation of gases dissolved in oil from the results of physicochemical tests. Thus, the objective would then be to adapt existing methodologies based on the concentration of dissolved gases to, based on physicochemical tests, predict possible incipient failures in

Since the estimation of these results is a complex problem, which does not have a mathematical formula capable of relating both tests, an interesting alternative is the use of computational intelligence techniques. Among these techniques, we can highlight the artificial neural networks, which do not require detailed knowledge about the relationship between the input space and the output space, as they are

It should be emphasized that the great novelty of this work is the precise estimation of the concentration of gases that are dissolved in insulating mineral oil, through neural networks, from the results of physicochemical tests. For the estimation of gases, we then use data from physicochemical tests performed on insulating oil taken from transformers in operation. The developed technique allowed the

**70**

insulation [4, 5].

power transformers.

expert in dealing with nonlinear mappings [8].


#### **Table 1.**

*Experimental tests and technical standards applied in insulating mineral oil.*

On the other hand, chromatographic tests, also known as dissolved gas analysis (DGA), are diagnostic tools for the detection and evaluation of incipient failures in insulating mineral oil. When a transformer failure occurs, the insulation system will provide chemical degradation, which will generate the production of various gases that will be present in the oil. These concentrations of gases are related to a type of transformer failure.

The main gases analyzed by DGA are listed in **Table 2**. In normal operation, the presence of these gases in the transformer is also normal, i.e., it does not indicate the existence of failures. The only exception is carbon dioxide (CO2), whose presence indicates a failure inside the equipment.

A transformer failure can result in the variation of the standard concentration of these gases in the oil. The increase in the gas concentration in the transformer indicates the occurrence of a failure that results in the insulation oil saturation. With this saturation, gas is released from the oil by modifying the characteristics of


**73**

applicability.

**mineral oil**

interconnected by artificial synapses.

those produced by chromatographic tests.

and *m* neurons in its output layer..

*Intelligent System for the Estimation of Gases Dissolved in Insulating Mineral Oil…*

the transformer. The amount of gas released depends on the oil temperature and the type of gas. Such gas production can be classified into three groups, namely, polarization, corona effect, and electric arc. This classification is based on the severity with which power is released during the failure. The largest and smallest amount of energy released is associated with the electric arc and corona, respectively [28].

In relation to the polarization, the gases released in the oil at low temperature are methane and ethylene, while at high temperature are methane, ethylene, ethane, and hydrogen. In cellulose, the gases generated at both high and low temperatures are carbon monoxide and dioxide. Regarding the corona effect, the gas produced in the oil is hydrogen, and the gases released by cellulose are hydrogen, monoxide, and carbon dioxide. For the electric arc, the gases released in this case are hydrogen,

Related to temperature, laboratory research shows that CO, CO2, and water originate when cellulose is overheated to a temperature of 140°C. Pyrolysis, which is destruction at temperatures above 250°C, produces more CO gas than CO2. In this case, there is also the formation of water, coal, and tar. At temperatures above 500°C, methane (CH4), ethane (C2H6), ethylene (C2H4), CO2, and water originate when O2 is present.

For the extraction of gases from the insulating mineral oil to be correct, it is then necessary that there are no bubbles in the collection vessel. The sample should be slightly warmed or injected into a degassing apparatus. The sample is subjected to vacuum, and the gases are collected in a graduated burette, which is taken to the chromatograph. The chromatograph is basically an apparatus used for chemical analysis of substances, capable of separating the various components of the sample. Thus, from values obtained in physicochemical tests performed on the insulating mineral oil, we have then developed an intelligent system capable of estimating

**3. Artificial neural network for estimating gases dissolved in insulating** 

Computational models based on artificial neural networks are inspired by the knowledge we have about the nervous system of humans, which has the ability to learn from experience. They are composed of artificial neurons, which are also

Artificial neural networks can be applied to various engineering- and sciencerelated problems. One of the areas with the greatest potential for their applicability is the universal approximation of functions, whose purpose is to represent functional relationships between inputs and outputs of typically nonlinear systems.

They can be applied to solve problems coming from various areas of knowledge, such as engineering, medicine, chemistry, and physics, where the possible solutions

One of the most commonly used architectures in artificial neural networks is that known as multilayer perceptron (MLP) [8], which will be also used in this application to map the relationship between results from physicochemical tests and

In addition, MLP networks are also characterized by the high possibilities of application in various types of problems related to the most different areas of knowledge, which is also considered as one of the most widely used in terms of

**Figure 1** illustrates an MLP topology consisting of three neural layers, which has *n* signals in the input layer, with *n*1 neurons (first layer), *n*2 neurons (second layer),

to the problems presented are difficult to obtain by conventional techniques.

*DOI: http://dx.doi.org/10.5772/intechopen.91807*

methane, ethane, ethylene, and acetylene.

the gas values present in the insulating oil of the transformer.

#### **Table 2.**

*Gases evaluated by DGA technique.*

*Intelligent System for the Estimation of Gases Dissolved in Insulating Mineral Oil… DOI: http://dx.doi.org/10.5772/intechopen.91807*

the transformer. The amount of gas released depends on the oil temperature and the type of gas. Such gas production can be classified into three groups, namely, polarization, corona effect, and electric arc. This classification is based on the severity with which power is released during the failure. The largest and smallest amount of energy released is associated with the electric arc and corona, respectively [28].

In relation to the polarization, the gases released in the oil at low temperature are methane and ethylene, while at high temperature are methane, ethylene, ethane, and hydrogen. In cellulose, the gases generated at both high and low temperatures are carbon monoxide and dioxide. Regarding the corona effect, the gas produced in the oil is hydrogen, and the gases released by cellulose are hydrogen, monoxide, and carbon dioxide. For the electric arc, the gases released in this case are hydrogen, methane, ethane, ethylene, and acetylene.

Related to temperature, laboratory research shows that CO, CO2, and water originate when cellulose is overheated to a temperature of 140°C. Pyrolysis, which is destruction at temperatures above 250°C, produces more CO gas than CO2. In this case, there is also the formation of water, coal, and tar. At temperatures above 500°C, methane (CH4), ethane (C2H6), ethylene (C2H4), CO2, and water originate when O2 is present.

For the extraction of gases from the insulating mineral oil to be correct, it is then necessary that there are no bubbles in the collection vessel. The sample should be slightly warmed or injected into a degassing apparatus. The sample is subjected to vacuum, and the gases are collected in a graduated burette, which is taken to the chromatograph. The chromatograph is basically an apparatus used for chemical analysis of substances, capable of separating the various components of the sample.

Thus, from values obtained in physicochemical tests performed on the insulating mineral oil, we have then developed an intelligent system capable of estimating the gas values present in the insulating oil of the transformer.

#### **3. Artificial neural network for estimating gases dissolved in insulating mineral oil**

Computational models based on artificial neural networks are inspired by the knowledge we have about the nervous system of humans, which has the ability to learn from experience. They are composed of artificial neurons, which are also interconnected by artificial synapses.

Artificial neural networks can be applied to various engineering- and sciencerelated problems. One of the areas with the greatest potential for their applicability is the universal approximation of functions, whose purpose is to represent functional relationships between inputs and outputs of typically nonlinear systems.

They can be applied to solve problems coming from various areas of knowledge, such as engineering, medicine, chemistry, and physics, where the possible solutions to the problems presented are difficult to obtain by conventional techniques.

One of the most commonly used architectures in artificial neural networks is that known as multilayer perceptron (MLP) [8], which will be also used in this application to map the relationship between results from physicochemical tests and those produced by chromatographic tests.

In addition, MLP networks are also characterized by the high possibilities of application in various types of problems related to the most different areas of knowledge, which is also considered as one of the most widely used in terms of applicability.

**Figure 1** illustrates an MLP topology consisting of three neural layers, which has *n* signals in the input layer, with *n*1 neurons (first layer), *n*2 neurons (second layer), and *m* neurons in its output layer..

*Application of Expert Systems - Theoretical and Practical Aspects*

**Tests Technical standards** Aniline point ABNT–NBR–11343 [13] Chromatography analysis ABNT–NBR–7070 [14, 15] Cinematic viscosity ABNT–NBR–10441 [16] Color ASTM–D1500 [17] Density ABNT–NBR–7148 [18]

Factor of losses at 90°C IEC–247 [21] Index of total acidity ASTM–D974 [22] Interfacial tension ABNT–NBR–6234 [23] Oxidation stability ABNT–NBR–10504 [24] Point of splendor and combustion ABNT–NBR–11341 [25] Polymerization level related to insulating paper ABNT–NBR–8148 [26] Tenor of water ABNT–NBR–5755 [27]

*Experimental tests and technical standards applied in insulating mineral oil.*

On the other hand, chromatographic tests, also known as dissolved gas analysis (DGA), are diagnostic tools for the detection and evaluation of incipient failures in insulating mineral oil. When a transformer failure occurs, the insulation system will provide chemical degradation, which will generate the production of various gases that will be present in the oil. These concentrations of gases are related to a type of

Dielectric rigidity IEC–156 [19] and ABNT–NBR–6869 [20]

The main gases analyzed by DGA are listed in **Table 2**. In normal operation, the presence of these gases in the transformer is also normal, i.e., it does not indicate the existence of failures. The only exception is carbon dioxide (CO2), whose presence

A transformer failure can result in the variation of the standard concentration of these gases in the oil. The increase in the gas concentration in the transformer indicates the occurrence of a failure that results in the insulation oil saturation. With this saturation, gas is released from the oil by modifying the characteristics of

**Dissolved gas Chemical formula**

Acetylene C2H2 Carbon dioxide CO2 Carbon monoxide CO Ethane C2H6 Ethylene C2H4 Hydrogen H2 Methane CH4 Nitrogen N2 Oxygen O2

**72**

**Table 2.**

*Gases evaluated by DGA technique.*

transformer failure.

**Table 1.**

indicates a failure inside the equipment.

Each of the neurons of this topology illustrated in **Figure 1** can be represented according to the terminology adopted in **Figure 2**.

In purely mathematical terms, the internal processing carried out by each MLP neural network neuron can be expressed as follows:

$$\mathcal{Y} \quad = \text{ g}\left(\sum\_{i=1}^{n} w\_i \cdot \boldsymbol{\varkappa}\_i \star \boldsymbol{b}\right) \tag{1}$$

where *xi* are the neuron inputs, *wi* represents the synaptic weight (artificial synapse) belonging to *i* th input value, θ refers to the activation threshold, *u* indicates activation potential, *y* represents the output response produced by the artificial neuron, and *g*(.) expresses the activation function that must be both differentiable and continuous throughout its domain, which are usually represented by the logistic activation function or by the hyperbolic tangent.

The training process of the neural network consists of applications of ordered steps that are necessary to tune the synaptic weights (*wi*) and thresholds (θ) associated to its neurons, with the ultimate goal of generalizing solutions to be produced by its outputs, whose responses are representative of the physical system to which they are mapping. The learning method used here for training MLP network was

**Figure 1.** *Illustration of an artificial neural network with multilayer perceptron architecture.*

**75**

**Figure 3.**

*Estimation of gas levels in the insulating mineral oil.*

*Intelligent System for the Estimation of Gases Dissolved in Insulating Mineral Oil…*

that based on the Levenberg-Marquardt algorithm, whose detailed steps are pre-

The training of this network was performed in a supervised manner, which consists in having available, considering each sample of the input signals, the respective desired outputs, i.e., each training example is represented by input signals and their respective corresponding outputs. In this case, from a database supplied by electric utilities, the network inputs were the parameters related to the physicochemical tests (**Table 1**) performed on several isolating mineral oil samples, while the desired outputs are the respective values of those gases obtained by chromatographic tests

Therefore, the MLP network used for this purpose was composed of 10 neurons (first layer) and 20 neurons (second layer), and the activation function adopted for

According to the previously presented sections, several isolating mineral oil samples were also presented to the MLP network in order to select the best architecture for the problem mapping, whose results are reported in the next figures.

**Figure 3** shows the estimated levels of gases dissolved in oil. Physicochemical experimental tests data were inserted in the MLP network inputs in order to produce, subsequently, the estimation of all gases dissolved in the insulating mineral oil. Network responses were compared to the values obtained from the chromatographic analyses. It is noteworthy that the MLP network responses are in accordance

Based upon **Figure 4**, it is possible to check the MLP network capacity of making

Estimation of oxygen (O2) obtained from 20 samples, which were not included in the training data, is shown in **Figure 5**. Considering this figure, it is possible to verify that the MLP network responses are very near to those obtained from the

adaptation and generalization about the analyzed data, whose results were com-

pared to actual values of H2 obtained from experimental tests.

*DOI: http://dx.doi.org/10.5772/intechopen.91807*

(**Table 2**) using the DGA technique.

with the gas concentration.

chromatographic tests.

all artificial neurons was the hyperbolic tangent.

**4. Experimental results and validation**

sented in [29, 30].

**Figure 2.** *Schematic diagram of artificial neuron used in MLP neural network.*

*Intelligent System for the Estimation of Gases Dissolved in Insulating Mineral Oil… DOI: http://dx.doi.org/10.5772/intechopen.91807*

that based on the Levenberg-Marquardt algorithm, whose detailed steps are presented in [29, 30].

The training of this network was performed in a supervised manner, which consists in having available, considering each sample of the input signals, the respective desired outputs, i.e., each training example is represented by input signals and their respective corresponding outputs. In this case, from a database supplied by electric utilities, the network inputs were the parameters related to the physicochemical tests (**Table 1**) performed on several isolating mineral oil samples, while the desired outputs are the respective values of those gases obtained by chromatographic tests (**Table 2**) using the DGA technique.

Therefore, the MLP network used for this purpose was composed of 10 neurons (first layer) and 20 neurons (second layer), and the activation function adopted for all artificial neurons was the hyperbolic tangent.

#### **4. Experimental results and validation**

*Application of Expert Systems - Theoretical and Practical Aspects*

according to the terminology adopted in **Figure 2**.

neural network neuron can be expressed as follows:

activation function or by the hyperbolic tangent.

*y* = *g*

synapse) belonging to *i*

Each of the neurons of this topology illustrated in **Figure 1** can be represented

In purely mathematical terms, the internal processing carried out by each MLP

*wi* ⋅ *xi* + *b*

th input value, θ refers to the activation threshold, *u* indicates

) (1)

(∑ *i*=1 *n*

activation potential, *y* represents the output response produced by the artificial neuron, and *g*(.) expresses the activation function that must be both differentiable and continuous throughout its domain, which are usually represented by the logistic

where *xi* are the neuron inputs, *wi* represents the synaptic weight (artificial

The training process of the neural network consists of applications of ordered steps that are necessary to tune the synaptic weights (*wi*) and thresholds (θ) associated to its neurons, with the ultimate goal of generalizing solutions to be produced by its outputs, whose responses are representative of the physical system to which they are mapping. The learning method used here for training MLP network was

**74**

**Figure 2.**

**Figure 1.**

*Schematic diagram of artificial neuron used in MLP neural network.*

*Illustration of an artificial neural network with multilayer perceptron architecture.*

According to the previously presented sections, several isolating mineral oil samples were also presented to the MLP network in order to select the best architecture for the problem mapping, whose results are reported in the next figures.

**Figure 3** shows the estimated levels of gases dissolved in oil. Physicochemical experimental tests data were inserted in the MLP network inputs in order to produce, subsequently, the estimation of all gases dissolved in the insulating mineral oil. Network responses were compared to the values obtained from the chromatographic analyses. It is noteworthy that the MLP network responses are in accordance with the gas concentration.

Based upon **Figure 4**, it is possible to check the MLP network capacity of making adaptation and generalization about the analyzed data, whose results were compared to actual values of H2 obtained from experimental tests.

Estimation of oxygen (O2) obtained from 20 samples, which were not included in the training data, is shown in **Figure 5**. Considering this figure, it is possible to verify that the MLP network responses are very near to those obtained from the chromatographic tests.

**Figure 3.** *Estimation of gas levels in the insulating mineral oil.*

**Figure 4.** *Estimated levels of H2 and comparative analysis.*

**Figure 5.** *Estimated levels of O2 and comparative analysis.*

Regarding oxygen (O2), it is noticed that the MLP network generalization about physicochemical analysis is very reasonable. Therefore, it is appropriate to estimate such parameters.

Estimation of carbon monoxide (CO) obtained from 20 samples, which were not included in the training data, is shown in **Figure 6**. Considering this figure, it is possible to confirm that network computed results are very near to those obtained from the chromatographic analysis.

In addition, it is noted that generalization of CO gas values carried out by the neural network based upon physicochemical analysis is positive. Therefore, it is also appropriate to estimate such gas values.

Estimation of carbon dioxide (CO2) obtained from 20 samples, which were not included in the training data, is shown in **Figure 7**. Considering this figure, it is

**77**

from the chromatographic analysis.

*Estimated levels of CO2 and comparative analysis.*

such parameters.

**Figure 7.**

**Figure 6.**

*Intelligent System for the Estimation of Gases Dissolved in Insulating Mineral Oil…*

possible to check that network computed results are also very near to those obtained

To summarize, it is verified that the generalization of CO2 gas values based upon physicochemical analysis is positive. Therefore, it is also appropriate to estimate

The following figures illustrate the results provided by the MLP neural network

**Figure 8** illustrates the relationship between CO2 gas levels with their colors, and these attributes are tabulated from 0.5 (for new oil) to 3.0 (for deteriorated oil). We have here checked that the relationship between CO2 levels decreases as the color

relating the estimated gases in relation to the physicochemical parameters.

*DOI: http://dx.doi.org/10.5772/intechopen.91807*

*Estimated levels of CO and comparative analysis.*

*Intelligent System for the Estimation of Gases Dissolved in Insulating Mineral Oil… DOI: http://dx.doi.org/10.5772/intechopen.91807*

**Figure 6.** *Estimated levels of CO and comparative analysis.*

*Application of Expert Systems - Theoretical and Practical Aspects*

Regarding oxygen (O2), it is noticed that the MLP network generalization about physicochemical analysis is very reasonable. Therefore, it is appropriate to estimate

Estimation of carbon monoxide (CO) obtained from 20 samples, which were not included in the training data, is shown in **Figure 6**. Considering this figure, it is possible to confirm that network computed results are very near to those obtained

In addition, it is noted that generalization of CO gas values carried out by the neural network based upon physicochemical analysis is positive. Therefore, it is also

Estimation of carbon dioxide (CO2) obtained from 20 samples, which were not included in the training data, is shown in **Figure 7**. Considering this figure, it is

**76**

such parameters.

**Figure 5.**

**Figure 4.**

*Estimated levels of H2 and comparative analysis.*

from the chromatographic analysis.

*Estimated levels of O2 and comparative analysis.*

appropriate to estimate such gas values.

**Figure 7.** *Estimated levels of CO2 and comparative analysis.*

possible to check that network computed results are also very near to those obtained from the chromatographic analysis.

To summarize, it is verified that the generalization of CO2 gas values based upon physicochemical analysis is positive. Therefore, it is also appropriate to estimate such parameters.

The following figures illustrate the results provided by the MLP neural network relating the estimated gases in relation to the physicochemical parameters.

**Figure 8** illustrates the relationship between CO2 gas levels with their colors, and these attributes are tabulated from 0.5 (for new oil) to 3.0 (for deteriorated oil). We have here checked that the relationship between CO2 levels decreases as the color

#### **Figure 8.**

*Relationship between CO2 gas levels as a function of the oil color.*

level increases. This behavior is related to the heating of the insulating mineral oil, which consequently also increases its oxidation level.

On the other hand, as illustrated in **Figure 9**, it is reported that CO2 gas levels increase linearly as the oil density value also increases. This behavior is related to abrupt changes in temperature over time.

**Figure 10** shows that the relationship between CO2 gas levels and the neutralization index of the insulating mineral oil decreases nonlinearly. This behavior may be related to oil acidity levels, since the increase of CO2 has direct implications on its neutralization levels.

**Figure 11**, in contrast to **Figure 9**, has illustrated that the relationship of CO gas levels as a function of insulating mineral oil density decreases linearly, as the behavior of both gases is independent of each other.

Already in **Figure 12**, we can see that the relationship between CO gas levels with their interfacial tension values remains almost constant until tension limit of 25 dyn/cm, and from this critical point, the CO gas levels drop sharply, as the less water present in the oil also implies higher interfacial tension.

**79**

**Figure 10.**

**Figure 11.**

*Intelligent System for the Estimation of Gases Dissolved in Insulating Mineral Oil…*

**Figure 13** now shows that CO gas levels as a function of water quantity are also inversely proportional. This behavioral profile is probably associated with temperature, as its decrease also implies a reduction in water evaporation and, consequently,

Finally, as shown in **Figure 14**, it is reported that the relationship of O2 gas levels as a function of oil density has a decreasing profile. Such behavior may be due to

From the results presented in this section, we demonstrate the possibility of mapping, by means of artificial neural networks, concentration values of gases dissolved in insulating mineral oil from the results of physicochemical tests. To estimate the gases, data from physicochemical tests performed on insulating mineral oil

The developed technique allowed the gas estimations in insulating mineral oil and without need to perform the chromatographic tests. Thus, the amount

increases its concentration in the insulating mineral oil.

*Relationship between CO gas levels as a function of the oil density.*

taken from operating transformers were used.

changes in acidity levels of insulating oil as density increases.

*Relationship between CO2 gas levels as a function of the oil neutralization index.*

*DOI: http://dx.doi.org/10.5772/intechopen.91807*

**Figure 9.** *Relationship between CO2 gas levels as a function of the oil density.*

*Intelligent System for the Estimation of Gases Dissolved in Insulating Mineral Oil… DOI: http://dx.doi.org/10.5772/intechopen.91807*

#### **Figure 10.**

*Application of Expert Systems - Theoretical and Practical Aspects*

which consequently also increases its oxidation level.

*Relationship between CO2 gas levels as a function of the oil color.*

behavior of both gases is independent of each other.

*Relationship between CO2 gas levels as a function of the oil density.*

water present in the oil also implies higher interfacial tension.

abrupt changes in temperature over time.

neutralization levels.

**Figure 8.**

level increases. This behavior is related to the heating of the insulating mineral oil,

On the other hand, as illustrated in **Figure 9**, it is reported that CO2 gas levels increase linearly as the oil density value also increases. This behavior is related to

**Figure 10** shows that the relationship between CO2 gas levels and the neutralization index of the insulating mineral oil decreases nonlinearly. This behavior may be related to oil acidity levels, since the increase of CO2 has direct implications on its

**Figure 11**, in contrast to **Figure 9**, has illustrated that the relationship of CO gas levels as a function of insulating mineral oil density decreases linearly, as the

Already in **Figure 12**, we can see that the relationship between CO gas levels with their interfacial tension values remains almost constant until tension limit of 25 dyn/cm, and from this critical point, the CO gas levels drop sharply, as the less

**78**

**Figure 9.**

*Relationship between CO2 gas levels as a function of the oil neutralization index.*

#### **Figure 11.**

*Relationship between CO gas levels as a function of the oil density.*

**Figure 13** now shows that CO gas levels as a function of water quantity are also inversely proportional. This behavioral profile is probably associated with temperature, as its decrease also implies a reduction in water evaporation and, consequently, increases its concentration in the insulating mineral oil.

Finally, as shown in **Figure 14**, it is reported that the relationship of O2 gas levels as a function of oil density has a decreasing profile. Such behavior may be due to changes in acidity levels of insulating oil as density increases.

From the results presented in this section, we demonstrate the possibility of mapping, by means of artificial neural networks, concentration values of gases dissolved in insulating mineral oil from the results of physicochemical tests. To estimate the gases, data from physicochemical tests performed on insulating mineral oil taken from operating transformers were used.

The developed technique allowed the gas estimations in insulating mineral oil and without need to perform the chromatographic tests. Thus, the amount

#### **Figure 12.**

*Relationship between CO gas levels as a function of the oil interfacial tension.*

of power supply interruptions can be considerably reduced, and the associated costs in determining the concentration of gases in the mineral oil follow this same downward trend.

#### **5. Conclusion**

An approach based on artificial neural networks was presented in this article, and its target was to estimate gas levels dissolved in insulating oil. Based upon data obtained from physicochemical analyses, decisions for transformer maintenance may be supported by the network; it can even indicate whether chromatographic analyses are necessary.

The artificial neural network approach proved to be efficient for estimating the gas levels dissolved in insulating oil. Based upon input data provided by

**81**

*Intelligent System for the Estimation of Gases Dissolved in Insulating Mineral Oil…*

physicochemical analyses, estimation of any gas obtained from chromatographic analyses becomes possible (without physicochemical analysis). Furthermore, artificial neural network approach was developed in order to set up the relationship

Finally, the developed neural approach is innovative, and it is also the first one that performs the estimation of dissolved gases in insulating mineral oil from the results of physicochemical tests. Therefore, in order to authenticate the developed methodology, the results provided by the artificial neural network were compared

The authors thank the ANEEL Research and Development Program under

between attributes which did not have any apparent relationship.

*Relationship between O2 gas levels as a function of the oil density.*

with those actual values obtained by chromatographic tests.

**Acknowledgements**

**Figure 14.**

contract number PD-0068-0037/2016.

*DOI: http://dx.doi.org/10.5772/intechopen.91807*

*Intelligent System for the Estimation of Gases Dissolved in Insulating Mineral Oil… DOI: http://dx.doi.org/10.5772/intechopen.91807*

**Figure 14.** *Relationship between O2 gas levels as a function of the oil density.*

physicochemical analyses, estimation of any gas obtained from chromatographic analyses becomes possible (without physicochemical analysis). Furthermore, artificial neural network approach was developed in order to set up the relationship between attributes which did not have any apparent relationship.

Finally, the developed neural approach is innovative, and it is also the first one that performs the estimation of dissolved gases in insulating mineral oil from the results of physicochemical tests. Therefore, in order to authenticate the developed methodology, the results provided by the artificial neural network were compared with those actual values obtained by chromatographic tests.

#### **Acknowledgements**

*Application of Expert Systems - Theoretical and Practical Aspects*

*Relationship between CO gas levels as a function of the oil interfacial tension.*

of power supply interruptions can be considerably reduced, and the associated costs in determining the concentration of gases in the mineral oil follow this same

*Relationship between CO gas levels as a function of the water concentration in the insulating mineral oil.*

An approach based on artificial neural networks was presented in this article, and its target was to estimate gas levels dissolved in insulating oil. Based upon data obtained from physicochemical analyses, decisions for transformer maintenance may be supported by the network; it can even indicate whether chromatographic

The artificial neural network approach proved to be efficient for estimating the gas levels dissolved in insulating oil. Based upon input data provided by

**80**

downward trend.

**Figure 13.**

**Figure 12.**

**5. Conclusion**

analyses are necessary.

The authors thank the ANEEL Research and Development Program under contract number PD-0068-0037/2016.

*Application of Expert Systems - Theoretical and Practical Aspects*

#### **Author details**

Ivan Nunes da Silva1 \*, Rogério Andrade Flauzino1 , Danilo Hernane Spatti1 , Renato Pagotto Bossolan2 and Bruno Augusto Trevisam2

1 University of São Paulo, São Carlos, SP, Brazil

2 São Paulo State Electric Power Transmission Company, ISA/CTEEP, São Paulo, SP, Brazil

\*Address all correspondence to: insilva@sc.usp.br

© 2020 The Author(s). Licensee IntechOpen. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/ by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

**83**

2017

*Intelligent System for the Estimation of Gases Dissolved in Insulating Mineral Oil…*

[9] Khuntia SR, Rueda JL, Bouwman S, Van der Meijden MAMM. A literature survey on asset management in electrical power [transmission and distribution] system. International Transactions on Electrical Energy Systems. 2016;**26**:2123-2133

[10] Zhong J, Li W, Wang C, Yu J, Xu R. Determining optimal inspection intervals in maintenance considering equipment aging failures. IEEE Transactions on Power Systems.

[11] Liu C, Huang G, Zhang K, Wen F, Salam MA, Ang SP. Asset management in power systems. In: 10th International Conference on Advances in Power System Control, Operation &

Management, Hong Kong, China; 2015.

[12] Shah SWA, Mahmood MN, Das N. Strategic asset management framework for the improvement of large scale PV power plants in Australia. In: Australasian Universities Power Engineering Conference, Brisbane,

[13] ABNT–NBR–11343. Petroleum products - Determination of aniline point and mixed aniline point. Brazilian Association of Technical

[14] ABNT–NBR–7070. Sampling of gases and mineral insulating oil of electrical equipments and free and solved gases analysis. Brazilian Association of Technical Standards;

[15] Haykin S. Neural Networks and Learning Machines. 3rd Ed. Upper Saddle River, USA: Prentice Hall; 2008

[16] ABNT–NBR–10441. Petroleum products – Transparent and opaque liquids – Determination of kinematic

Australia; 2016. pp. 1-5

Standards; 2003

2006

2017;**32**:1474-1482

pp. 1-5

*DOI: http://dx.doi.org/10.5772/intechopen.91807*

regeneration of oils insulating minerals. Modern Electricity. 1996;**25**:39-49

[2] Liang Z, Parlikad A. A Markovian

maintenance. International Journal of Electrical Power & Energy Systems.

[3] Koksal A, Ozdemir A. Improved transformer maintenance plan for reliability centred asset management of power transmission system. IET Generation, Transmission and Distribution. 2016;**10**(8):1976-1983

[4] Wattakapaiboon W, Pattanadech N. The state of the art for dissolved gas analysis based on interpretation techniques. In: International Conference on Condition Monitoring and Diagnosis (CMD), Xi'an, China; 2016. pp. 60-63

[5] Cox R. Categorizing transformer faults via dissolved gas analysis. In: 19th IEEE International Conference on Dielectric Liquids, Manchester, UK;

[6] Ariffin MM, Ishak MT, Hamid MHA, Katim NIA, Ishak AM, Azis N. Ageing effect of vegetable oils impregnated paper in transformer application. In: International Conference on High Voltage Engineering and Power Systems, Sanur, Indonesia; 2017. pp. 183-187

[7] Ariffin MM, Ishak MT, Hamid MHA,

Comparative studies of the stability of various fluids under electrical discharge and thermal stresses. IEEE Transactions on Dielectrics and Electrical Insulation.

[8] Silva IN, Spatti DH, Flauzino RA, Liboni LHB, Alves SFR. Artificial Neural Networks – A Practical Course. 1st ed. Zurich, Switzerland: Springer;

Katim NIA, Ishak AM, Azis N.

2015;**22**(5):2491-2499

[1] Ferreira LG. Procedures for

model for power transformer

**References**

2018;**99**:175-182

2017. pp. 1-3

*Intelligent System for the Estimation of Gases Dissolved in Insulating Mineral Oil… DOI: http://dx.doi.org/10.5772/intechopen.91807*

#### **References**

*Application of Expert Systems - Theoretical and Practical Aspects*

**82**

Brazil

**Author details**

Ivan Nunes da Silva1

Renato Pagotto Bossolan2

1 University of São Paulo, São Carlos, SP, Brazil

\*Address all correspondence to: insilva@sc.usp.br

provided the original work is properly cited.

\*, Rogério Andrade Flauzino1

and Bruno Augusto Trevisam2

2 São Paulo State Electric Power Transmission Company, ISA/CTEEP, São Paulo, SP,

© 2020 The Author(s). Licensee IntechOpen. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/ by/3.0), which permits unrestricted use, distribution, and reproduction in any medium,

, Danilo Hernane Spatti1

,

[1] Ferreira LG. Procedures for regeneration of oils insulating minerals. Modern Electricity. 1996;**25**:39-49

[2] Liang Z, Parlikad A. A Markovian model for power transformer maintenance. International Journal of Electrical Power & Energy Systems. 2018;**99**:175-182

[3] Koksal A, Ozdemir A. Improved transformer maintenance plan for reliability centred asset management of power transmission system. IET Generation, Transmission and Distribution. 2016;**10**(8):1976-1983

[4] Wattakapaiboon W, Pattanadech N. The state of the art for dissolved gas analysis based on interpretation techniques. In: International Conference on Condition Monitoring and Diagnosis (CMD), Xi'an, China; 2016. pp. 60-63

[5] Cox R. Categorizing transformer faults via dissolved gas analysis. In: 19th IEEE International Conference on Dielectric Liquids, Manchester, UK; 2017. pp. 1-3

[6] Ariffin MM, Ishak MT, Hamid MHA, Katim NIA, Ishak AM, Azis N. Ageing effect of vegetable oils impregnated paper in transformer application. In: International Conference on High Voltage Engineering and Power Systems, Sanur, Indonesia; 2017. pp. 183-187

[7] Ariffin MM, Ishak MT, Hamid MHA, Katim NIA, Ishak AM, Azis N. Comparative studies of the stability of various fluids under electrical discharge and thermal stresses. IEEE Transactions on Dielectrics and Electrical Insulation. 2015;**22**(5):2491-2499

[8] Silva IN, Spatti DH, Flauzino RA, Liboni LHB, Alves SFR. Artificial Neural Networks – A Practical Course. 1st ed. Zurich, Switzerland: Springer; 2017

[9] Khuntia SR, Rueda JL, Bouwman S, Van der Meijden MAMM. A literature survey on asset management in electrical power [transmission and distribution] system. International Transactions on Electrical Energy Systems. 2016;**26**:2123-2133

[10] Zhong J, Li W, Wang C, Yu J, Xu R. Determining optimal inspection intervals in maintenance considering equipment aging failures. IEEE Transactions on Power Systems. 2017;**32**:1474-1482

[11] Liu C, Huang G, Zhang K, Wen F, Salam MA, Ang SP. Asset management in power systems. In: 10th International Conference on Advances in Power System Control, Operation & Management, Hong Kong, China; 2015. pp. 1-5

[12] Shah SWA, Mahmood MN, Das N. Strategic asset management framework for the improvement of large scale PV power plants in Australia. In: Australasian Universities Power Engineering Conference, Brisbane, Australia; 2016. pp. 1-5

[13] ABNT–NBR–11343. Petroleum products - Determination of aniline point and mixed aniline point. Brazilian Association of Technical Standards; 2003

[14] ABNT–NBR–7070. Sampling of gases and mineral insulating oil of electrical equipments and free and solved gases analysis. Brazilian Association of Technical Standards; 2006

[15] Haykin S. Neural Networks and Learning Machines. 3rd Ed. Upper Saddle River, USA: Prentice Hall; 2008

[16] ABNT–NBR–10441. Petroleum products – Transparent and opaque liquids – Determination of kinematic viscosity and calculation of dynamic viscosity. Brazilian Association of Technical Standards; 2007

[17] ASTM–D1500. Standard test method for ASTM color of petroleum products (ASTM color scale). American Society for Testing and Materials; 2017

[18] ABNT–NBR–7148. Petroleum and petroleum products – Determination of density, relative density and °API – Hydrometer method. Brazilian Association of Technical Standards; 2013

[19] IEC–156. Insulating liquids – Determination of the breakdown voltage at power frequency – Test method. International Electrotechnical Commission; 1995

[20] ABNT–NBR–6869. Electric insulating liquids – Determination of the dielectric breakdown voltage (disk electrodes). Brazilian Association of Technical Standards; 1989

[21] IEC–247. Measurement of relative permittivity, dielectric dissipation factor and d.c. resistivity of insulating liquids. International Electrotechnical Commission; 1979

[22] ASTM–D974. Standard test method for acid and base number by colorindicator titration. American Society for Testing and Materials; 2014

[23] ABNT–NBR–6234. Mineral insulating oil – Determination of interfacial tension of oil-water – Test method. Brazilian Association of Technical Standards; 2015

[24] ABNT–NBR–10504. Mineral insulating oil – Determination of oxidation stability – Test method. Brazilian Association of Technical Standards; 1988

[25] ABNT–NBR–11341. Petroleum products – Determination of the flash and fire points by Cleveland open cup. Brazilian Association of Technical Standards; 2004

[26] ABNT–NBR–8148. Measurement of the average viscosimetric degree of polymerization of new and aged electrical papers and paperboards. Brazilian Association of Technical Standards; 2000

[27] ABNT–NBR–5755. Determination of water in insulating liquids (Method of Karl Fischer). Brazilian Association of Technical Standards; 1984

[28] Milasch M. Maintenance of Transformers in Insulating Liquid. Ed. São Paulo, Brazil: Edgar Blucher; 1984

[29] Hagan MT, Menhaj MB. Training feedforward networks with the Marquardt algorithm. IEEE Transactions on Neural Networks. 1994;**5**(6):989-993

[30] Shinde SB, Sayyad SS. Cost sensitive improved Levenberg Marquardt algorithm for imbalanced data. In: IEEE International Conference on Computational Intelligence and Computing Research, Chennai, India; 2017. pp. 1-4

**85**

**Chapter 7**

**Abstract**

**1. Introduction**

systems can be listed as follows:

• Equipment lifetime

• Operating conditions

• External conditions

• Operational safety risks to be assumed

considered, such as:

Expert Systems

*Renato Bossolan and Bruno Vitti*

Efficient Asset Management

*Danilo Spatti, Luisa H.B. Liboni, Marcel Araújo,* 

**Keywords:** expert systems, asset management, power systems

Practices for Power Systems Using

Electric power companies have high financial costs due to poor asset management practices. Therefore, it is crucial to use decision-making processes to decrease the global costs of an active asset and to extend its lifetime to a maximum. Asset management programs, which are frequently used to tackle optimization problems, aim to guide the use of the physical assets of a business, mainly by optimizing their lifetime. Efficient asset management practices establish operation and maintenance for each equipment, from the time the equipment is acquired until the appropriate time for its replacement. So, based on these assumptions, we propose a method to assist asset management decisionmaking in the electric power companies, which is embodied by computer software.

The big challenge involving asset management in electric systems is to seek a solution that enables the electric sector to reconcile interests with the environmental, economic, market, technological, regulatory and corporate constraints.

Still, the first parameters to be considered in the asset management of electrical

Additionally, directly or indirectly, operational context aspects must also be

• Requirement level to be met or minimum performance requirement

• Type of operation requests (existence of redundancies or equipment in standby)

#### **Chapter 7**

*Application of Expert Systems - Theoretical and Practical Aspects*

and fire points by Cleveland open cup. Brazilian Association of Technical

[26] ABNT–NBR–8148. Measurement of the average viscosimetric degree of polymerization of new and aged electrical papers and paperboards. Brazilian Association of Technical

[27] ABNT–NBR–5755. Determination of water in insulating liquids (Method of Karl Fischer). Brazilian Association of

Standards; 2004

Standards; 2000

Technical Standards; 1984

[28] Milasch M. Maintenance of Transformers in Insulating Liquid. Ed. São Paulo, Brazil: Edgar Blucher; 1984

feedforward networks with the Marquardt algorithm. IEEE Transactions on Neural Networks.

1994;**5**(6):989-993

2017. pp. 1-4

[29] Hagan MT, Menhaj MB. Training

[30] Shinde SB, Sayyad SS. Cost sensitive

improved Levenberg Marquardt algorithm for imbalanced data. In: IEEE International Conference on Computational Intelligence and Computing Research, Chennai, India;

viscosity and calculation of dynamic viscosity. Brazilian Association of

[17] ASTM–D1500. Standard test method for ASTM color of petroleum products (ASTM color scale). American Society for Testing and Materials; 2017

[18] ABNT–NBR–7148. Petroleum and petroleum products – Determination of density, relative density and °API – Hydrometer method. Brazilian Association of Technical Standards;

[19] IEC–156. Insulating liquids – Determination of the breakdown voltage at power frequency – Test method. International Electrotechnical

[20] ABNT–NBR–6869. Electric insulating liquids – Determination of the dielectric breakdown voltage (disk electrodes). Brazilian Association of

[21] IEC–247. Measurement of relative permittivity, dielectric dissipation factor and d.c. resistivity of insulating liquids. International Electrotechnical

[22] ASTM–D974. Standard test method for acid and base number by colorindicator titration. American Society for

Technical Standards; 1989

Testing and Materials; 2014

[23] ABNT–NBR–6234. Mineral insulating oil – Determination of interfacial tension of oil-water – Test method. Brazilian Association of Technical Standards; 2015

[24] ABNT–NBR–10504. Mineral insulating oil – Determination of oxidation stability – Test method. Brazilian Association of Technical

[25] ABNT–NBR–11341. Petroleum products – Determination of the flash

Commission; 1995

Commission; 1979

Technical Standards; 2007

2013

**84**

Standards; 1988

## Efficient Asset Management Practices for Power Systems Using Expert Systems

*Danilo Spatti, Luisa H.B. Liboni, Marcel Araújo, Renato Bossolan and Bruno Vitti*

### **Abstract**

Electric power companies have high financial costs due to poor asset management practices. Therefore, it is crucial to use decision-making processes to decrease the global costs of an active asset and to extend its lifetime to a maximum. Asset management programs, which are frequently used to tackle optimization problems, aim to guide the use of the physical assets of a business, mainly by optimizing their lifetime. Efficient asset management practices establish operation and maintenance for each equipment, from the time the equipment is acquired until the appropriate time for its replacement. So, based on these assumptions, we propose a method to assist asset management decisionmaking in the electric power companies, which is embodied by computer software.

**Keywords:** expert systems, asset management, power systems

#### **1. Introduction**

The big challenge involving asset management in electric systems is to seek a solution that enables the electric sector to reconcile interests with the environmental, economic, market, technological, regulatory and corporate constraints.

Still, the first parameters to be considered in the asset management of electrical systems can be listed as follows:


Additionally, directly or indirectly, operational context aspects must also be considered, such as:


With a large amount of information currently available from assets, it is critical that standalone or semi-autonomous machine learning techniques should be able to extract knowledge and increase the performance or robustness of a system or process. These algorithms and techniques are vast and are also used in non-database applications—for example; such methods can directly interact with the environment to accomplish the tasks discussed above [1, 2].

The computational tools and techniques based on machine learning help to solve problems related to the extraction of information and knowledge from data. For example, machine learning applied to data mining is capable of performing specific tasks such as:


The purpose of this chapter is to present a case study involving the processing of asset databases of an electrical transmission system in order to create subsidies for the development of an efficient asset management system.

### **2. Database studies**

The databases in this case study are composed of data from power transformers. Such databases contain nominal values from the devices, data from laboratory tests, and maintenance data. The database consists of 6929 records, with the attributes shown in **Table 1**.

Data preprocessing methods consists of database conditioning, inconsistent data correction and the analysis of the data through relational graphs and probability distributions in order to extract knowledge and useful information.

**Figure 1** shows a histogram of the transformers with respect to their manufacturing date and **Figure 2** shows a histogram of the transformers with respect to their age. **Table 2** shows the records for the voltage attribute of the transformers.

By using data from maintenance records, one can classify maintenance tasks as Preventive Maintenance and Corrective Maintenance. One of the critical analysis

**87**

**Figure 2.**

**Figure 1.**

**Table 1.**

*Efficient Asset Management Practices for Power Systems Using Expert Systems*

Age of the transformer when corrective maintenance was performed

Manufacturer Year of construction Time in operation Feeder phase Did achieve its lifetime according to regulation policies? Power Location Voltage

Level of priority of the maintenances Preventive maintenance dates Technical maintenance team Corrective maintenance dates

*Histogram (count percentage) of the transformers with respect to their manufacturing date.*

*Histogram (count percentage) of the transformers with respect to their age.*

*DOI: http://dx.doi.org/10.5772/intechopen.89766*

*Attributes from the power transformers database.*

#### *Efficient Asset Management Practices for Power Systems Using Expert Systems DOI: http://dx.doi.org/10.5772/intechopen.89766*


#### **Table 1.**

*Application of Expert Systems - Theoretical and Practical Aspects*

With a large amount of information currently available from assets, it is critical that standalone or semi-autonomous machine learning techniques should be able to extract knowledge and increase the performance or robustness of a system or process. These algorithms and techniques are vast and are also used in non-database applications—for example; such methods can directly interact with the environ-

The computational tools and techniques based on machine learning help to solve problems related to the extraction of information and knowledge from data. For example, machine learning applied to data mining is capable of performing specific

• Feature selection (selection of the most critical variables in a process, system,

The purpose of this chapter is to present a case study involving the processing of asset databases of an electrical transmission system in order to create subsidies for

The databases in this case study are composed of data from power transformers. Such databases contain nominal values from the devices, data from laboratory tests, and maintenance data. The database consists of 6929 records, with the attributes

Data preprocessing methods consists of database conditioning, inconsistent data correction and the analysis of the data through relational graphs and probability

By using data from maintenance records, one can classify maintenance tasks as Preventive Maintenance and Corrective Maintenance. One of the critical analysis

• Environmental standards

• Equipment life cycle

• Maintenance logistics

tasks such as:

• Pattern recognition

• Data processing

• Data clustering

or database)

• Regressions

• Data sorting

**2. Database studies**

shown in **Table 1**.

transformers.

• Regulatory standards and legislation

ment to accomplish the tasks discussed above [1, 2].

the development of an efficient asset management system.

distributions in order to extract knowledge and useful information.

**Figure 1** shows a histogram of the transformers with respect to their manufacturing date and **Figure 2** shows a histogram of the transformers with respect to their age. **Table 2** shows the records for the voltage attribute of the

**86**

*Attributes from the power transformers database.*

#### **Figure 1.**

*Histogram (count percentage) of the transformers with respect to their manufacturing date.*

#### **Figure 2.**

*Histogram (count percentage) of the transformers with respect to their age.*

#### *Application of Expert Systems - Theoretical and Practical Aspects*


#### **Table 2.**

*Voltage attribute from the power transformers database.*

#### **Figure 3.**

*Histogram (count percentage) of the transformers with respect to preventive and corrective maintenance, taking into account if the lifetime was achieved (1) or not (0).*

that must be done by an efficient asset management system is to take into account the lifetime of the power transformer and the maintenance procedures held during its lifetime [3, 4]. Such analysis is shown in **Figure 3**.

#### **3. Statistical distribution analysis**

One can verify, in **Figure 4**, the relationship between the probability of maintenance as a function of the lifetime of the power transformers. In this figure, the cumulative distribution functions of preventive and corrective maintenance are shown and have, clearly, a bathtub shape (**Figure 5**) [5].

The curves in **Figure 4** raise a management question, and, therefore, explain the importance of such analysis: the more preventive maintenance, the less corrective

**89**

in **Figures 7** and **8**.

**Figure 4.**

**Figure 5.**

Manufacturer 2?

as can be seen in **Figures 9** and **10**.

*Efficient Asset Management Practices for Power Systems Using Expert Systems*

maintenance? This hypothesis can be tested by a correlation analysis, which indicates a Correlation between preventive and corrective maintenance of −0.9367, i.e.,

One can observe that preventive maintenance in equipment such as power transformers is made in a more uniform way throughout its operating lifespan. This result corroborates with the fact that this kind of equipment is crucial and should

To explore the hypothesis mentioned above, two manufacturers were chosen, namely A and B, and the following analysis was performed. The results are shown

These analyses can be extended for other voltage classes, such as 345 and 440 kV,

By analyzing these curves, some questions can be raised concerning asset management in order to help decision-making. For example, in this scenario, would it be possible to purchase more equipment with this specific nominal voltage from

the two types of maintenances are firmly negatively correlated.

be robust and faultless, as is shown in **Figure 6**.

*Maintenance rate with respect to the age of the assets.*

*DOI: http://dx.doi.org/10.5772/intechopen.89766*

*Cumulative probability function of the power transformers.*

*Efficient Asset Management Practices for Power Systems Using Expert Systems DOI: http://dx.doi.org/10.5772/intechopen.89766*

**Figure 4.** *Cumulative probability function of the power transformers.*

**Figure 5.** *Maintenance rate with respect to the age of the assets.*

maintenance? This hypothesis can be tested by a correlation analysis, which indicates a Correlation between preventive and corrective maintenance of −0.9367, i.e., the two types of maintenances are firmly negatively correlated.

One can observe that preventive maintenance in equipment such as power transformers is made in a more uniform way throughout its operating lifespan. This result corroborates with the fact that this kind of equipment is crucial and should be robust and faultless, as is shown in **Figure 6**.

To explore the hypothesis mentioned above, two manufacturers were chosen, namely A and B, and the following analysis was performed. The results are shown in **Figures 7** and **8**.

By analyzing these curves, some questions can be raised concerning asset management in order to help decision-making. For example, in this scenario, would it be possible to purchase more equipment with this specific nominal voltage from Manufacturer 2?

These analyses can be extended for other voltage classes, such as 345 and 440 kV, as can be seen in **Figures 9** and **10**.

*Application of Expert Systems - Theoretical and Practical Aspects*

**Voltage (kV) Records** 13.8 1221 34.0 194 69.0 95 88.0 1497 138.0 2092 145.0 4 230.0 545 345.0 451 440.0 690

*Histogram (count percentage) of the transformers with respect to preventive and corrective maintenance, taking* 

that must be done by an efficient asset management system is to take into account the lifetime of the power transformer and the maintenance procedures held during

One can verify, in **Figure 4**, the relationship between the probability of maintenance as a function of the lifetime of the power transformers. In this figure, the cumulative distribution functions of preventive and corrective maintenance are

The curves in **Figure 4** raise a management question, and, therefore, explain the importance of such analysis: the more preventive maintenance, the less corrective

**88**

**Figure 3.**

**Table 2.**

*into account if the lifetime was achieved (1) or not (0).*

*Voltage attribute from the power transformers database.*

**3. Statistical distribution analysis**

its lifetime [3, 4]. Such analysis is shown in **Figure 3**.

shown and have, clearly, a bathtub shape (**Figure 5**) [5].

**Figure 6.** *Cumulative probability function for corrective maintenance.*

**Figure 7.**

*Maintenance rate with respect to the age of the assets for manufacturer A.*

**91**

**4. Risk analysis**

**Figure 10.**

**Figure 9.**

range or age, as shown in **Figure 13**.

*Efficient Asset Management Practices for Power Systems Using Expert Systems*

Defining a process for creating a condition index for an asset is very important since the index can be used for asset management. For example, managers could consider alternative manufacturer scenarios, replacement or new acquisitions. Thus, the condition index of equipment can be an important feedback tool for the electric utility regulator since good maintenance practices dictated by the regulator,

Also, risk assessment allows the analysis of failure probability, as shown in **Figures 11** and **12**, in which all power transformers in the database were analyzed. Alternatively, it can also be established that the reliability function R (t) is given by R (t) = 1 − F (t), where F (t) is the cumulative probability function. Function R (t) can be used to analyze the probability of failure of an asset above a particular

As a case study, one can then investigate what the failure rate would be for a particular transformer cluster. Now, suppose the cluster composed of all equipment

such as the frequencies for maintenance, can be updated [6, 7].

*Maintenance rate with respect to the age of the assets for voltage 440 kV.*

*DOI: http://dx.doi.org/10.5772/intechopen.89766*

*Maintenance rate with respect to the age of the assets for voltage 345 kV.*

**Figure 8.** *Maintenance rate with respect to the age of the assets for manufacturer B.*

*Efficient Asset Management Practices for Power Systems Using Expert Systems DOI: http://dx.doi.org/10.5772/intechopen.89766*

**Figure 9.**

*Application of Expert Systems - Theoretical and Practical Aspects*

*Maintenance rate with respect to the age of the assets for manufacturer A.*

*Cumulative probability function for corrective maintenance.*

*Maintenance rate with respect to the age of the assets for manufacturer B.*

**90**

**Figure 8.**

**Figure 7.**

**Figure 6.**

*Maintenance rate with respect to the age of the assets for voltage 345 kV.*

**Figure 10.** *Maintenance rate with respect to the age of the assets for voltage 440 kV.*

#### **4. Risk analysis**

Defining a process for creating a condition index for an asset is very important since the index can be used for asset management. For example, managers could consider alternative manufacturer scenarios, replacement or new acquisitions. Thus, the condition index of equipment can be an important feedback tool for the electric utility regulator since good maintenance practices dictated by the regulator, such as the frequencies for maintenance, can be updated [6, 7].

Also, risk assessment allows the analysis of failure probability, as shown in **Figures 11** and **12**, in which all power transformers in the database were analyzed.

Alternatively, it can also be established that the reliability function R (t) is given by R (t) = 1 − F (t), where F (t) is the cumulative probability function. Function R (t) can be used to analyze the probability of failure of an asset above a particular range or age, as shown in **Figure 13**.

As a case study, one can then investigate what the failure rate would be for a particular transformer cluster. Now, suppose the cluster composed of all equipment

#### **Figure 11.**

*Corrective maintenance rate against operating time.*

**Figure 12.** *Probability of failure of all transformers with respect to their age.*

**93**

**Figure 16.**

*Empirical distribution approximation.*

**Figure 14.**

**Figure 15.**

*Efficient Asset Management Practices for Power Systems Using Expert Systems*

*Manufacturer 1 maintenance rate (138 kV) with respect to operating time.*

*Failure rate of manufacturer 1 transformers (138 kV) with respect to their age.*

*DOI: http://dx.doi.org/10.5772/intechopen.89766*

**Figure 13.** *Reliability rate of all transformers with respect to their age.*

*Efficient Asset Management Practices for Power Systems Using Expert Systems DOI: http://dx.doi.org/10.5772/intechopen.89766*

#### **Figure 14.**

*Application of Expert Systems - Theoretical and Practical Aspects*

**92**

**Figure 13.**

**Figure 11.**

**Figure 12.**

*Corrective maintenance rate against operating time.*

*Probability of failure of all transformers with respect to their age.*

*Reliability rate of all transformers with respect to their age.*

*Manufacturer 1 maintenance rate (138 kV) with respect to operating time.*

**Figure 16.** *Empirical distribution approximation.*

#### **Figure 17.**

*Transformer ID 147511 (manufacturer 1) failure rate will increase by 7% between 48 and 49 years of age of the asset.*

from Manufacturer 1 and with nominal voltage of 138 kV. We then obtain the results shown in **Figures 14** and **15**. **Figure 16** shows the approximation of this distribution.

In this case, for a particular transformer, i.e., transformer ID 147511 (manufacturer 1 // voltage class 138 kV // age 48 years), one can observe that the failure rate of this asset will increase 7%, between 48 and 49 years of age. The results of such analysis can be seen from the examination of **Figure 17**.

#### **5. Critical state index estimation**

The so-called critical state index allows one to verify if the failure rate of the asset is coherent with the moment in its life cycle. Thus, it becomes possible, by considering the age of the asset and its history of failures, to point out the assets that require distinguished attention [7].

There are different methodologies for identifying and determining the type of defect or failure in equipment immersed in insulating mineral oil based on the results of chromatographic analysis. These methods, such as those listed as follows, are directives for estimating a state index for transformers:


These methods analyze gas concentration limits to estimate the state of the transformer, i.e., gas concentrations are used to diagnose the equipment. To make

**95**

*Efficient Asset Management Practices for Power Systems Using Expert Systems*

this diagnosis, that is, to quantify the probability of a transformer defect, the fol-

**Gas Limit L1 (ppm) Limit G2 (ppm per month)**

H2 100 50 CH4 75 38 C2H2 3 3 C2H4 75 38 C2H6 75 38 CO 700 350 CO2 7000 3500

• Condition 1 refers to a normal operating condition, that is, if the dissolved gases are at levels below those presented in **Table 3**, then the transformer is

• Condition 2 indicates a possible failure and many mineral oil samples should

• Condition 3 indicates a high level of cellulose and/or mineral oil decomposi-

• In condition 4 there is an indication of excessive cellulose and/or mineral oil

The joint assessment between the equipment critical state index and its risk analysis, which is based on the history of maintenance procedures, allows for a more assertive decision-making process. To illustrate how the decision-making process can be enriched with both information, consider a 440 kV power transformer that entered into operation in 1981. The graph shown by **Figure 18** illustrates the number of accumulated failures since the last total fault, which has required

**Figure 18** shows that a number of failures between 1.3 and 1.6 failures were expected since the last intervention. This analysis is made with a 90% confidence

In order to construct curves involving the operating state of the devices, more than 25,000 gas chromatography assays were evaluated. The tests date from 1977 to the present day, and each test was evaluated according to the standards: IEEE, IEC, Duval, Doernenburg, Roger, and Key Gas. The IEEE criterion was adopted to

The integrated Analysis, which is made by considering the mean behavior from

level by considering the asset history and the database of similar assets.

evaluate the operating condition, as represented by **Figures 19**–**22**.

decomposition. Continued operation under Condition 4 may result in a total fault.

To confirm the existence of a fault, one of the gases shown in **Table 3** must have a concentration equal to or higher than L1 and more significant than the value indicated by G2. Thus, this methodology has an appropriate way to evaluate the condition of a transformer based on the historical data about its chromatographic assays. When the historical data are not available, the IEEE method can be more

*DOI: http://dx.doi.org/10.5772/intechopen.89766*

lowing information in **Table 3** can be employed.

*Duval criteria for identifying transformer defects.*

**Table 3.**

considered to be operating correctly.

efficient, since it uses four conditions to identify failure states:

be taken to determine the tendency of gas growth.

tion, and there is probably a transformer failure.

intervention, as a function of the operating time.

all assays can be shown in **Figure 23**.



**Table 3.**

*Application of Expert Systems - Theoretical and Practical Aspects*

from Manufacturer 1 and with nominal voltage of 138 kV. We then obtain the results shown in **Figures 14** and **15**. **Figure 16** shows the approximation of this

analysis can be seen from the examination of **Figure 17**.

are directives for estimating a state index for transformers:

**5. Critical state index estimation**

that require distinguished attention [7].

In this case, for a particular transformer, i.e., transformer ID 147511 (manufacturer 1 // voltage class 138 kV // age 48 years), one can observe that the failure rate of this asset will increase 7%, between 48 and 49 years of age. The results of such

*Transformer ID 147511 (manufacturer 1) failure rate will increase by 7% between 48 and 49 years of age of the* 

The so-called critical state index allows one to verify if the failure rate of the asset is coherent with the moment in its life cycle. Thus, it becomes possible, by considering the age of the asset and its history of failures, to point out the assets

There are different methodologies for identifying and determining the type of defect or failure in equipment immersed in insulating mineral oil based on the results of chromatographic analysis. These methods, such as those listed as follows,

These methods analyze gas concentration limits to estimate the state of the transformer, i.e., gas concentrations are used to diagnose the equipment. To make

**94**

distribution.

**Figure 17.**

*asset.*

• Key Gas

• Doernenburg

• Duval triangle

• IEC ratio

• IEEE

• Roger

*Duval criteria for identifying transformer defects.*

this diagnosis, that is, to quantify the probability of a transformer defect, the following information in **Table 3** can be employed.

To confirm the existence of a fault, one of the gases shown in **Table 3** must have a concentration equal to or higher than L1 and more significant than the value indicated by G2. Thus, this methodology has an appropriate way to evaluate the condition of a transformer based on the historical data about its chromatographic assays.

When the historical data are not available, the IEEE method can be more efficient, since it uses four conditions to identify failure states:


The joint assessment between the equipment critical state index and its risk analysis, which is based on the history of maintenance procedures, allows for a more assertive decision-making process. To illustrate how the decision-making process can be enriched with both information, consider a 440 kV power transformer that entered into operation in 1981. The graph shown by **Figure 18** illustrates the number of accumulated failures since the last total fault, which has required intervention, as a function of the operating time.

**Figure 18** shows that a number of failures between 1.3 and 1.6 failures were expected since the last intervention. This analysis is made with a 90% confidence level by considering the asset history and the database of similar assets.

In order to construct curves involving the operating state of the devices, more than 25,000 gas chromatography assays were evaluated. The tests date from 1977 to the present day, and each test was evaluated according to the standards: IEEE, IEC, Duval, Doernenburg, Roger, and Key Gas. The IEEE criterion was adopted to evaluate the operating condition, as represented by **Figures 19**–**22**.

The integrated Analysis, which is made by considering the mean behavior from all assays can be shown in **Figure 23**.

**Figure 18.**

*Number of failures since the last intervention for a given transformer.*

#### **Figure 19.**

*Percentage of transformers in condition 1.*

**97**

**Figure 23.** *Integrated analysis.*

**Figure 21.**

**Figure 22.**

*Percentage of transformers in condition 3.*

*Percentage of transformers in condition 4.*

*Efficient Asset Management Practices for Power Systems Using Expert Systems*

*DOI: http://dx.doi.org/10.5772/intechopen.89766*

**Figure 20.** *Percentage of transformers in condition 2.*

*Efficient Asset Management Practices for Power Systems Using Expert Systems DOI: http://dx.doi.org/10.5772/intechopen.89766*

#### **Figure 21.**

*Application of Expert Systems - Theoretical and Practical Aspects*

*Number of failures since the last intervention for a given transformer.*

**96**

**Figure 20.**

**Figure 18.**

**Figure 19.**

*Percentage of transformers in condition 1.*

*Percentage of transformers in condition 2.*

*Percentage of transformers in condition 3.*

**Figure 22.** *Percentage of transformers in condition 4.*

**Figure 23.** *Integrated analysis.*

### **6. Conclusions**

Critical assets in the transmission and distribution industries need special care and attention, mainly regarding the aging of the equipment since their lifespan impacts profits as well as the reliability and safety of the electric system.

As the lifetime of these devices become longer, it is justifiable to develop methods to identify their health condition, taking into account not only historical data but also all available asset management tools that companies currently own.

As the vast majority of companies still struggle to learn from the abundant data acquired, such a method should have significant implications in helping managers evaluate and question in-company policies regarding manufacturers and preventive maintenance practices.

In this context, the methodology presented in this chapter can be applied to determine the lifespan of transmission system equipment, based on the determination of failure rates, by using statistical analysis.

#### **Acknowledgements**

The authors thank the ANEEL R & D Program, contract number PD-0068-0037/2016.

#### **Author details**

Danilo Spatti1 \*, Luisa H.B. Liboni2 , Marcel Araújo3 , Renato Bossolan4 and Bruno Vitti4

1 University of São Paulo, São Carlos, Brazil

2 Federal Institute of Education, Science, and Technology of São Paulo, Sertãozinho, Brazil

3 Federal Rural University of Pernambuco, Recife, Brazil

4 São Paulo State Electric Power Transmission Company, São Paulo, Brazil

\*Address all correspondence to: spatti@icmc.usp.br

© 2019 The Author(s). Licensee IntechOpen. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/ by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

**99**

*Efficient Asset Management Practices for Power Systems Using Expert Systems*

*DOI: http://dx.doi.org/10.5772/intechopen.89766*

[1] Khuntia SR, Rueda JL, Bouwmanb S, Meijden M. Classification, domains and risk assessment in asset management: A literature study. In: 2015 50th International Universities Power Engineering Conference (UPEC). 2015

[2] Nowlan FS, Heap HF. Reliability Centered Maintenance. National Technical Information Service, EUA, Report no. AD/A066-579; 1978

[3] Lapworth JA, Wilson A. The asset health review for managing reliability risks associated with ongoing use of ageing system power transformers. In: 2008 IEEE CMD–International Conference on Condition Monitoring and Diagnosis. 2008. pp. 605-608

[4] Jahromi A, Piercy R, Cress S, Service J, Fan W. An approach to power transformer asset management using health index. IEEE Electrical Insulation

Magazine. 2009;**25**(2):20-34

America. 2012. p. 8

2008. pp. 132-135

[6] Nemeth B, Benyo T, Jager A, Csepes G, Woynarovich G. Complex diagnostic methods for lifetime extension of power transformers. In: 2008 IEEE ISEI–International Symposium on Electrical Insulation.

[7] Zhang X, Gockenbach E. Assetmanagement of transformers based on condition monitoring and standard diagnosis. IEEE Electrical Insulation

Magazine. 2008;**24**(4):26-40

[5] Carneiro JC, Jardini JA, Brittes JLP. Substation power transformer risk management: Reflecting on reliability centered maintenance and monitoring. In: 2012 IEEE/PES T&D–LA–Sixth Transmission and Distribution Conference and Exposition Latin

**References**

*Efficient Asset Management Practices for Power Systems Using Expert Systems DOI: http://dx.doi.org/10.5772/intechopen.89766*

#### **References**

*Application of Expert Systems - Theoretical and Practical Aspects*

tion of failure rates, by using statistical analysis.

\*, Luisa H.B. Liboni2

3 Federal Rural University of Pernambuco, Recife, Brazil

\*Address all correspondence to: spatti@icmc.usp.br

provided the original work is properly cited.

1 University of São Paulo, São Carlos, Brazil

Critical assets in the transmission and distribution industries need special care and attention, mainly regarding the aging of the equipment since their lifespan

As the lifetime of these devices become longer, it is justifiable to develop methods to identify their health condition, taking into account not only historical data but also all available asset management tools that companies currently own.

As the vast majority of companies still struggle to learn from the abundant data acquired, such a method should have significant implications in helping managers evaluate and question in-company policies regarding manufacturers and preventive

In this context, the methodology presented in this chapter can be applied to determine the lifespan of transmission system equipment, based on the determina-

, Marcel Araújo3

2 Federal Institute of Education, Science, and Technology of São Paulo,

4 São Paulo State Electric Power Transmission Company, São Paulo, Brazil

© 2019 The Author(s). Licensee IntechOpen. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/ by/3.0), which permits unrestricted use, distribution, and reproduction in any medium,

, Renato Bossolan4

impacts profits as well as the reliability and safety of the electric system.

The authors thank the ANEEL R & D Program, contract number

**6. Conclusions**

maintenance practices.

**Acknowledgements**

PD-0068-0037/2016.

**Author details**

and Bruno Vitti4

Sertãozinho, Brazil

Danilo Spatti1

**98**

[1] Khuntia SR, Rueda JL, Bouwmanb S, Meijden M. Classification, domains and risk assessment in asset management: A literature study. In: 2015 50th International Universities Power Engineering Conference (UPEC). 2015

[2] Nowlan FS, Heap HF. Reliability Centered Maintenance. National Technical Information Service, EUA, Report no. AD/A066-579; 1978

[3] Lapworth JA, Wilson A. The asset health review for managing reliability risks associated with ongoing use of ageing system power transformers. In: 2008 IEEE CMD–International Conference on Condition Monitoring and Diagnosis. 2008. pp. 605-608

[4] Jahromi A, Piercy R, Cress S, Service J, Fan W. An approach to power transformer asset management using health index. IEEE Electrical Insulation Magazine. 2009;**25**(2):20-34

[5] Carneiro JC, Jardini JA, Brittes JLP. Substation power transformer risk management: Reflecting on reliability centered maintenance and monitoring. In: 2012 IEEE/PES T&D–LA–Sixth Transmission and Distribution Conference and Exposition Latin America. 2012. p. 8

[6] Nemeth B, Benyo T, Jager A, Csepes G, Woynarovich G. Complex diagnostic methods for lifetime extension of power transformers. In: 2008 IEEE ISEI–International Symposium on Electrical Insulation. 2008. pp. 132-135

[7] Zhang X, Gockenbach E. Assetmanagement of transformers based on condition monitoring and standard diagnosis. IEEE Electrical Insulation Magazine. 2008;**24**(4):26-40

**101**

Engine.

**Chapter 8**

**Abstract**

Transformers

Business Intelligence-based.

**1. Introduction**

*Renato Bossolan and Bruno Vitti*

Computational Intelligence to

Estimate Fault Rates in Power

*Danilo Spatti, Luisa H.B. Liboni, Marcel Araújo,* 

Asset management in power transmission systems is one of the significant practices carried out by power companies. With the aging of the devices, the development of optimized tools, capable of considering failure rates, regulatory scenarios, and operational parameters, is increasingly mandatory. The purpose of this work is to present a statistics-based tool for optimized asset management. For such an objective, we have developed a computational method based on database processing and statistical studies that can support decision-making on preventive maintenance in the equipment of the electric sector. The final system interface is

**Keywords:** computational intelligence, failure rates, power transformers

easily accessible to the users through a graphical interface [3].

In this chapter, we will describe a software, based on Business Intelligence, that has health data (test data, inspections, and operation) of assets as inputs, as well as their technical and constructive characteristics [1, 2]. The software has as outputs the possible categorization of assets into families, the calculation and analysis of failure rates, and the detection of current or incipient anomalies. Therefore, the software includes relational graphs, statistical analysis, and machine learning tools. In **Figure 1**, the central conception of the system, which is named as the ISA CTEEP Asset Management Support System (AMSS), is shown. Thus, data from critical assets, such as power transformers, will also be considered as inputs to the software. In addition to making trend and relational graphs, the output module for anomaly identification and failure rate calculation, by analyzing test and inspection data, indicates the probability of a particular asset to needing special care. The software architecture was developed so that both input and output modules are

The algorithms that perform the relational and critical analysis of the health of an asset and calculate its failure rate are hosted within the program called Analysis

Input data may be made available manually by the user or may be directly acquired at the ISA CTEEP database. The software should query the database

#### **Chapter 8**

## Computational Intelligence to Estimate Fault Rates in Power Transformers

*Danilo Spatti, Luisa H.B. Liboni, Marcel Araújo, Renato Bossolan and Bruno Vitti*

#### **Abstract**

Asset management in power transmission systems is one of the significant practices carried out by power companies. With the aging of the devices, the development of optimized tools, capable of considering failure rates, regulatory scenarios, and operational parameters, is increasingly mandatory. The purpose of this work is to present a statistics-based tool for optimized asset management. For such an objective, we have developed a computational method based on database processing and statistical studies that can support decision-making on preventive maintenance in the equipment of the electric sector. The final system interface is Business Intelligence-based.

**Keywords:** computational intelligence, failure rates, power transformers

#### **1. Introduction**

In this chapter, we will describe a software, based on Business Intelligence, that has health data (test data, inspections, and operation) of assets as inputs, as well as their technical and constructive characteristics [1, 2]. The software has as outputs the possible categorization of assets into families, the calculation and analysis of failure rates, and the detection of current or incipient anomalies. Therefore, the software includes relational graphs, statistical analysis, and machine learning tools.

In **Figure 1**, the central conception of the system, which is named as the ISA CTEEP Asset Management Support System (AMSS), is shown. Thus, data from critical assets, such as power transformers, will also be considered as inputs to the software. In addition to making trend and relational graphs, the output module for anomaly identification and failure rate calculation, by analyzing test and inspection data, indicates the probability of a particular asset to needing special care. The software architecture was developed so that both input and output modules are easily accessible to the users through a graphical interface [3].

The algorithms that perform the relational and critical analysis of the health of an asset and calculate its failure rate are hosted within the program called Analysis Engine.

Input data may be made available manually by the user or may be directly acquired at the ISA CTEEP database. The software should query the database

#### **Figure 1.**

*Structure of the AMSS system.*

through a database abstraction (interpreter) layer. These architectural details can be seen in **Figure 1**. **Figure 2** shows, in detail, the functional layers of the software.

**103**

**Figure 3.**

*Computational Intelligence to Estimate Fault Rates in Power Transformers*

Maintenance procedures are considered inputs to the fault analysis we shall present. In total, there are 5371 maintenance records available in the database from the end of 2008 to the present moment. Maintenance records are mined in order to find relevant information for the analysis. Thus, nonrelevant corrective maintenance records are excluded from the analysis. **Figure 3** shows the relationship between the number of maintenance records as a function of the year of occurrence and the age

*Summary of the number of records according to the year of occurrence and the age of the asset.*

*DOI: http://dx.doi.org/10.5772/intechopen.89768*

○ Liquid chromatography

c. Specialized tests:

• Insulation resistance

• Ohmic resistance

• Special tests

**2. Modeling fault as states**

• Etc.

of the asset [5].

• Transformation ratio

• Inspection of accessories

• Short circuit impedance

○ Transformer power factor test

The assessment of the health of major interest assets, such as transformers, autotransformers, and reactors, will be made considering three aspects [4]:

	- Insulating mineral oil tests:
		- Physicochemical tests
		- Gas chromatography

*Computational Intelligence to Estimate Fault Rates in Power Transformers DOI: http://dx.doi.org/10.5772/intechopen.89768*


*Application of Expert Systems - Theoretical and Practical Aspects*

through a database abstraction (interpreter) layer. These architectural details can be seen in **Figure 1**. **Figure 2** shows, in detail, the functional layers of the software. The assessment of the health of major interest assets, such as transformers, autotransformers, and reactors, will be made considering three aspects [4]:

a. History of operation and maintenance

• Insulating mineral oil tests:

○ Physicochemical tests

○ Gas chromatography

b.Routine tests:

*Layers of the AMSS system.*

**102**

**Figure 1.**

**Figure 2.**

*Structure of the AMSS system.*


#### **2. Modeling fault as states**

Maintenance procedures are considered inputs to the fault analysis we shall present. In total, there are 5371 maintenance records available in the database from the end of 2008 to the present moment. Maintenance records are mined in order to find relevant information for the analysis. Thus, nonrelevant corrective maintenance records are excluded from the analysis. **Figure 3** shows the relationship between the number of maintenance records as a function of the year of occurrence and the age of the asset [5].

**Figure 3.**

*Summary of the number of records according to the year of occurrence and the age of the asset.*

From **Figure 3** it is possible to observe the existence of sparse data, as better highlighted in **Figure 4**. The same sparsity can be perceived in **Figure 5**, where the relationship between the number of equipment and its age is shown.

Assets can be categorized into states. The ideal state in which the equipment must operate is the **NORMAL** state.

**Preventive Maintenance** processes can ensure that the asset remains in its **NORMAL** state.

However, even with preventive maintenance procedures, there is a transition from the **NORMAL** state to the **DEFECT** state, as shown in **Figure 6**. The state transition occurs given a defect rate.

The transition from the **DEFECT** state (or fault state) to the **NORMAL** state occurs when **Corrective maintenance** actions are made. The transition takes place through a Corrective Maintenance rate, as represented in **Figure 7**.

Still, there is the **FAILURE** state, which is characterized by the complete withdrawal of operation of the asset. The transition takes place through a failure rate, as in **Figure 8**.

**Figure 4.**

*Number of maintenance records with respect to equipment age and year of occurrence. Sparsity highlighted.*

**105**

transformers.

*Computational Intelligence to Estimate Fault Rates in Power Transformers*

In the case of transformers, the **FAILURE** state can be divided into two others,

The approximation of the defect rate function should take into account the quality of the available information. The quality of the information depends on the number of maintenance records available and the number of devices [6]. The

A similar approximation can be made for the failure rate, as depicted in **Figure 11**, which is modeled by means of a power type function. **Figure 12** also shows the failures involving peripheral elements and the active parts of

i.e., **Internal Failure** and **External failure**, as shown in **Figure 9**.

relative quality of the data can be seen in **Figure 10**.

*DOI: http://dx.doi.org/10.5772/intechopen.89768*

**Figure 6.**

**Figure 7.**

**Figure 8.**

*State transition functions (II).*

*State transition functions (III).*

*State transition functions (I).*

**Figure 5.**

*Number of equipment with respect to equipment age. Sparsity highlighted.*

*Computational Intelligence to Estimate Fault Rates in Power Transformers DOI: http://dx.doi.org/10.5772/intechopen.89768*

*Application of Expert Systems - Theoretical and Practical Aspects*

must operate is the **NORMAL** state.

transition occurs given a defect rate.

**NORMAL** state.

in **Figure 8**.

From **Figure 3** it is possible to observe the existence of sparse data, as better highlighted in **Figure 4**. The same sparsity can be perceived in **Figure 5**, where the

Assets can be categorized into states. The ideal state in which the equipment

**Preventive Maintenance** processes can ensure that the asset remains in its

However, even with preventive maintenance procedures, there is a transition from the **NORMAL** state to the **DEFECT** state, as shown in **Figure 6**. The state

The transition from the **DEFECT** state (or fault state) to the **NORMAL** state occurs when **Corrective maintenance** actions are made. The transition takes place

Still, there is the **FAILURE** state, which is characterized by the complete withdrawal of operation of the asset. The transition takes place through a failure rate, as

*Number of maintenance records with respect to equipment age and year of occurrence. Sparsity highlighted.*

relationship between the number of equipment and its age is shown.

through a Corrective Maintenance rate, as represented in **Figure 7**.

**104**

**Figure 5.**

**Figure 4.**

*Number of equipment with respect to equipment age. Sparsity highlighted.*

**Figure 8.** *State transition functions (III).*

In the case of transformers, the **FAILURE** state can be divided into two others, i.e., **Internal Failure** and **External failure**, as shown in **Figure 9**.

The approximation of the defect rate function should take into account the quality of the available information. The quality of the information depends on the number of maintenance records available and the number of devices [6]. The relative quality of the data can be seen in **Figure 10**.

A similar approximation can be made for the failure rate, as depicted in **Figure 11**, which is modeled by means of a power type function. **Figure 12** also shows the failures involving peripheral elements and the active parts of transformers.

**Figure 9.** *State transition functions (IV).*

**Figure 10.**

*Relative quality of the available data as a function of the asset age.*

**107**

**Figure 13.**

*Business intelligence Interface page 1.*

*Computational Intelligence to Estimate Fault Rates in Power Transformers*

*DOI: http://dx.doi.org/10.5772/intechopen.89768*

**3. Business intelligence interface**

maintenance records, as exemplified in **Figure 14**.

*Failures for peripheral (orange) elements and active (red) parts.*

page can be seen in **Figure 13**.

**Figure 12.**

All intelligence embedded in the AMSS systems was implemented using the concept of Business Intelligence, which uses graphical reports to represent data. The software was divided into nine pages, with several graphical analyses each. The first

In this screen, the user can choose the **Substation**, the **Equipment** type, and also the **Voltage** class, among **729** available equipment in **4843** available corrective

**Figure 11.** *Approximation of the failure rate.*

#### *Computational Intelligence to Estimate Fault Rates in Power Transformers DOI: http://dx.doi.org/10.5772/intechopen.89768*

*Application of Expert Systems - Theoretical and Practical Aspects*

**106**

**Figure 11.**

*Approximation of the failure rate.*

**Figure 9.**

**Figure 10.**

*Relative quality of the available data as a function of the asset age.*

*State transition functions (IV).*

**Figure 12.** *Failures for peripheral (orange) elements and active (red) parts.*

### **3. Business intelligence interface**

All intelligence embedded in the AMSS systems was implemented using the concept of Business Intelligence, which uses graphical reports to represent data. The software was divided into nine pages, with several graphical analyses each. The first page can be seen in **Figure 13**.

In this screen, the user can choose the **Substation**, the **Equipment** type, and also the **Voltage** class, among **729** available equipment in **4843** available corrective maintenance records, as exemplified in **Figure 14**.

#### **Figure 13.**

*Business intelligence Interface page 1.*

#### *Application of Expert Systems - Theoretical and Practical Aspects*


#### **Figure 14.**

*Selecting substations, equipment, and voltage class detail from page 1.*

**Figure 15.** *Equipment fabrication histogram by year (page 1).*

**109**

**Figure 18.**

*Business intelligence Interface page 2.*

*Computational Intelligence to Estimate Fault Rates in Power Transformers*

In **Figure 13**, it is still possible to visualize, in detail, the histogram of the manufacturing year of the e**quipment**, as well as the **Corrective maintenance records** 

Page 2 displays graphical reports involving all corrective maintenance records

In Page 3 of the AMSS system, **Figure 19**, it is possible to see the correlations of corrective maintenance records with the age of the assets, which are shown in

*DOI: http://dx.doi.org/10.5772/intechopen.89768*

**by age**, detailed in **Figures 15** and **16**, respectively. The **Maintenance priority** is detailed in **Figure 17**.

for all devices, as can be shown in **Figure 18**.

**Figures 20** and **21**.

**Figure 17.**

*Maintenance priority (page 1).*

**Figure 16.** *Corrective maintenance records by age (page 1).*

*Computational Intelligence to Estimate Fault Rates in Power Transformers DOI: http://dx.doi.org/10.5772/intechopen.89768*

In **Figure 13**, it is still possible to visualize, in detail, the histogram of the manufacturing year of the e**quipment**, as well as the **Corrective maintenance records by age**, detailed in **Figures 15** and **16**, respectively.

The **Maintenance priority** is detailed in **Figure 17**.

Page 2 displays graphical reports involving all corrective maintenance records for all devices, as can be shown in **Figure 18**.

In Page 3 of the AMSS system, **Figure 19**, it is possible to see the correlations of corrective maintenance records with the age of the assets, which are shown in **Figures 20** and **21**.

#### **Figure 17.** *Maintenance priority (page 1).*

*Application of Expert Systems - Theoretical and Practical Aspects*

*Selecting substations, equipment, and voltage class detail from page 1.*

**108**

**Figure 16.**

**Figure 14.**

**Figure 15.**

*Equipment fabrication histogram by year (page 1).*

*Corrective maintenance records by age (page 1).*


**Figure 18.** *Business intelligence Interface page 2.*

**Figure 19.** *Business intelligence Interface page 3.*

**Figure 20.** *Detailed number of corrective maintenance records by asset age from page 3.*

**111**

respectively.

**Figure 23.**

*Business intelligence Interface page 5.*

**Figure 22.**

*Business intelligence Interface page 4.*

*Computational Intelligence to Estimate Fault Rates in Power Transformers*

Page 4 presents the user with information from chromatographic assays for power transformers that can be individually selected, as can be seen in **Figure 22**. Pages 5 and 6 still address information from chromatography tests, showing the gas emission evolution over the years, as well as the correlation between the tests and the corrective maintenance history for each asset, as seen in **Figures 23** and **24**,

The emission evolution of H2, CO, CO2, total dissolved combustible gases, CH4,

Page 7 shows the behavior of the dissolved gases as a function of time. The user can select a particular asset or analyze the evolution globally, this is, for all assets.

C2H4, C2H6, and C2H2 can be seen in details in **Figures 25**–**32**.

*DOI: http://dx.doi.org/10.5772/intechopen.89768*

**Figure 21.** *Detailed age count by age from page 3.*

*Computational Intelligence to Estimate Fault Rates in Power Transformers DOI: http://dx.doi.org/10.5772/intechopen.89768*

**Figure 22.** *Business intelligence Interface page 4.*

*Application of Expert Systems - Theoretical and Practical Aspects*

*Detailed number of corrective maintenance records by asset age from page 3.*

**110**

**Figure 21.**

*Detailed age count by age from page 3.*

**Figure 19.**

**Figure 20.**

*Business intelligence Interface page 3.*

**Figure 23.** *Business intelligence Interface page 5.*

Page 4 presents the user with information from chromatographic assays for power transformers that can be individually selected, as can be seen in **Figure 22**.

Pages 5 and 6 still address information from chromatography tests, showing the gas emission evolution over the years, as well as the correlation between the tests and the corrective maintenance history for each asset, as seen in **Figures 23** and **24**, respectively.

The emission evolution of H2, CO, CO2, total dissolved combustible gases, CH4, C2H4, C2H6, and C2H2 can be seen in details in **Figures 25**–**32**.

Page 7 shows the behavior of the dissolved gases as a function of time. The user can select a particular asset or analyze the evolution globally, this is, for all assets.

#### *Application of Expert Systems - Theoretical and Practical Aspects*

**Figure 24.** *Business intelligence Interface page 6.*

**Figure 25.** *Detailed H2 emission evolution from page 5.*

**113**

**Figure 29.**

**Figure 27.**

**Figure 28.**

*Computational Intelligence to Estimate Fault Rates in Power Transformers*

*DOI: http://dx.doi.org/10.5772/intechopen.89768*

*Detailed CO2 emission evolution from page 5.*

*Detailed TDCG emission evolution from page 5.*

*Detailed CH4 emission evolution from page 6.*

**Figure 26.** *Detailed CO emission evolution from page 5.*

*Computational Intelligence to Estimate Fault Rates in Power Transformers DOI: http://dx.doi.org/10.5772/intechopen.89768*

**Figure 27.** *Detailed CO2 emission evolution from page 5.*

*Application of Expert Systems - Theoretical and Practical Aspects*

**112**

**Figure 26.**

**Figure 24.**

**Figure 25.**

*Business intelligence Interface page 6.*

*Detailed H2 emission evolution from page 5.*

*Detailed CO emission evolution from page 5.*

**Figure 28.** *Detailed TDCG emission evolution from page 5.*

**Figure 29.** *Detailed CH4 emission evolution from page 6.*

**Figure 30.** *Detailed C2H4 emission evolution from page 6.*

**Figure 31.** *Detailed C2H6 emission evolution from page 6.*

**115**

**Figure 34.**

**Figure 33.**

*Business intelligence Interface page 7.*

*Computational Intelligence to Estimate Fault Rates in Power Transformers*

This behavior is shown in **Figure 33**. The detailed evolution of the dissolved gases as

Finally, Page 9, which is shown in **Figure 37**, performs an integrated analysis of

It can be seen from **Figure 37** that the circles represent the transformers, and the diameter of each circle is related to the corrective maintenance rate per year to which this asset is subjected. The ordinate axis represents the current condition that this asset is in.

Page 8 presents the integrated analyses involving the relationship between corrective maintenance records and chromatographic tests, as exemplified in **Figure 35**. **Figure 36** shows the trend line of maintenance records per year as a

a function of the age of the assets can be seen in **Figure 34**.

the health of an asset, along with its maintenance history.

function of the age of the asset.

*Detailed dissolved gases evolution from page 7.*

*DOI: http://dx.doi.org/10.5772/intechopen.89768*

**Figure 32.** *Detailed C2H2 emission evolution from page 6.*

#### *Computational Intelligence to Estimate Fault Rates in Power Transformers DOI: http://dx.doi.org/10.5772/intechopen.89768*

**Figure 33.** *Business intelligence Interface page 7.*

*Application of Expert Systems - Theoretical and Practical Aspects*

**114**

**Figure 32.**

**Figure 30.**

**Figure 31.**

*Detailed C2H4 emission evolution from page 6.*

*Detailed C2H6 emission evolution from page 6.*

*Detailed C2H2 emission evolution from page 6.*

**Figure 34.** *Detailed dissolved gases evolution from page 7.*

This behavior is shown in **Figure 33**. The detailed evolution of the dissolved gases as a function of the age of the assets can be seen in **Figure 34**.

Page 8 presents the integrated analyses involving the relationship between corrective maintenance records and chromatographic tests, as exemplified in **Figure 35**. **Figure 36** shows the trend line of maintenance records per year as a function of the age of the asset.

Finally, Page 9, which is shown in **Figure 37**, performs an integrated analysis of the health of an asset, along with its maintenance history.

It can be seen from **Figure 37** that the circles represent the transformers, and the diameter of each circle is related to the corrective maintenance rate per year to which this asset is subjected. The ordinate axis represents the current condition that this asset is in.

#### *Application of Expert Systems - Theoretical and Practical Aspects*

**Figure 35.** *Business intelligence Interface page 8.*

**Figure 36.** *Detailed trend line for corrective maintenance from page 8.*

### **4. Conclusions**

Asset management in the power sector, especially in transmission systems, has been driving the development of increasingly efficient feature extraction tools.

This chapter has introduced an integrated method based on business intelligence that analyzes data and failure rates in order to assist decision-making. The computational system takes into account technical information from the assets, data from chromatographic tests, as well as standard information regarding the operative condition of the assets, especially power transformers.

**117**

**Author details**

**Acknowledgements**

*Business intelligence Interface page 9.*

**Figure 37.**

PD-0068-0037/2016.

and Bruno Vitti4

Sertãozinho, Brazil

\*, Luisa H.B. Liboni2

3 Federal Rural University of Pernambuco, Recife, Brazil

\*Address all correspondence to: spatti@icmc.usp.br

provided the original work is properly cited.

1 University of São Paulo, São Carlos, Brazil

, Marcel Araújo3

2 Federal Institute of Education, Science, and Technology of São Paulo,

The authors thank the ANEEL R & D Program contract number

4 São Paulo State Electric Power Transmission Company, São Paulo, Brazil

© 2019 The Author(s). Licensee IntechOpen. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/ by/3.0), which permits unrestricted use, distribution, and reproduction in any medium,

, Renato Bossolan4

Danilo Spatti1

*Computational Intelligence to Estimate Fault Rates in Power Transformers*

*DOI: http://dx.doi.org/10.5772/intechopen.89768*

*Computational Intelligence to Estimate Fault Rates in Power Transformers DOI: http://dx.doi.org/10.5772/intechopen.89768*

**Figure 37.** *Business intelligence Interface page 9.*

### **Acknowledgements**

*Application of Expert Systems - Theoretical and Practical Aspects*

**116**

tools.

**Figure 36.**

**Figure 35.**

*Business intelligence Interface page 8.*

**4. Conclusions**

Asset management in the power sector, especially in transmission systems, has been driving the development of increasingly efficient feature extraction

This chapter has introduced an integrated method based on business intelligence that analyzes data and failure rates in order to assist decision-making. The computational system takes into account technical information from the assets, data from chromatographic tests, as well as standard information regarding the operative

condition of the assets, especially power transformers.

*Detailed trend line for corrective maintenance from page 8.*

The authors thank the ANEEL R & D Program contract number PD-0068-0037/2016.

### **Author details**

Danilo Spatti1 \*, Luisa H.B. Liboni2 , Marcel Araújo3 , Renato Bossolan4 and Bruno Vitti4

1 University of São Paulo, São Carlos, Brazil

2 Federal Institute of Education, Science, and Technology of São Paulo, Sertãozinho, Brazil

3 Federal Rural University of Pernambuco, Recife, Brazil

4 São Paulo State Electric Power Transmission Company, São Paulo, Brazil

\*Address all correspondence to: spatti@icmc.usp.br

© 2019 The Author(s). Licensee IntechOpen. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/ by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

### **References**

[1] Jahromi A, Piercy R, Cress S, Service J, Fan W. An approach to power transformer asset management using health index. IEEE Electrical Insulation Magazine. 2009;**25**(2):20-34

[2] Nemeth B, Benyo T, Jager A, Csepes G, Woynarovich G. Complex diagnostic methods for lifetime extension of power transformers. In: 2008 IEEE ISEI—International Symposium on Electrical Insulation. 2008. pp. 132-135

[3] Zhang X, Gockenbach E. Assetmanagement of transformers based on condition monitoring and standard diagnosis. IEEE Electrical Insulation Magazine. 2008;**24**(4):26-40

[4] Coble JB. Merging data sources to predict remaining useful life' an automated method to identify prognostic parameters (University of Tennessee PhD dissertation). 2010

[5] Heo JH, Kim MK, Park GP, Yoon YT, Park JK, Lee SS, et al. A reliabilitycentered approach to an optimal maintenance strategy in transmission systems using a genetic algorithm. IEEE Transactions on Power Delivery. 2011;**26**(4):2171-2179

[6] Sarchiz D, Bica D, Georgescu O. Mathematical model of reliability centered, maintenance (RCM). Power transmission and distribution networks applications. In: 2009 IEEE PowerTech. 2009. 4 p

**118**

*Application of Expert Systems - Theoretical and Practical Aspects*

[1] Jahromi A, Piercy R, Cress S, Service J, Fan W. An approach to power transformer asset management using health index. IEEE Electrical Insulation

Magazine. 2009;**25**(2):20-34

2008. pp. 132-135

**References**

[2] Nemeth B, Benyo T, Jager A, Csepes G, Woynarovich G. Complex diagnostic methods for lifetime extension of power transformers. In: 2008 IEEE ISEI—International Symposium on Electrical Insulation.

[3] Zhang X, Gockenbach E. Assetmanagement of transformers based on condition monitoring and standard diagnosis. IEEE Electrical Insulation

[4] Coble JB. Merging data sources to predict remaining useful life' an automated method to identify prognostic parameters (University of Tennessee PhD dissertation). 2010

[5] Heo JH, Kim MK, Park GP, Yoon YT, Park JK, Lee SS, et al. A reliabilitycentered approach to an optimal maintenance strategy in transmission systems using a genetic algorithm. IEEE Transactions on Power Delivery.

[6] Sarchiz D, Bica D, Georgescu O. Mathematical model of reliability centered, maintenance (RCM). Power transmission and distribution networks applications. In: 2009 IEEE PowerTech.

Magazine. 2008;**24**(4):26-40

2011;**26**(4):2171-2179

2009. 4 p

*Edited by Ivan Nunes da Silva and Rogério Andrade Flauzino*

What are expert systems and what are their purposes? What are the impacts resulting from their implementations? This book aims to answer these questions and more. Written by experts in the field, chapters It explores different concepts of expert systems such as computational intelligence, signal processing, real time systems, systems optimization, electric power systems, fault diagnosis, asset management, and smart cityescities. This book will appeal to wide range of readers, including those interested in acquiring basic knowledge and those who are motivated to learn more about the technical elements and technological applications of expert systems.

Published in London, UK © 2020 IntechOpen © ktsimage / iStock

Application of Expert Systems - Theoretical and Practical Aspects

Application of Expert Systems

Theoretical and Practical Aspects

*Edited by Ivan Nunes da Silva* 

*and Rogério Andrade Flauzino*