*A Review of Alternative Marine Fuels DOI: http://dx.doi.org/10.5772/intechopen.97871*

**Figure 8.** *Decision Tree Results.*

ship movement by one complete rotation of the propeller or the propeller pitch is calculated by shipyard and written in the "ship's manual". Engine distance is calculated by multiplying the propeller pitch to propeller distances. For a certain time period the ship movement distance can be calculated by engine distance. But in reality, engine distance can vary due to weather conditions such as wind, current and swells directions, fouling on the ship's hull, etc. Therefore, the observed distance might be less or more than engine distance.

Slip is a rate of the difference between the engine distance and observed distance. The simple formulation showed below:

$$\text{Slip} = \text{100} \times \frac{\text{Engine Distance} - \text{Observed Distance}}{\text{Engine Distance}} \tag{1}$$

#### *A Review of Alternative Marine Fuels DOI: http://dx.doi.org/10.5772/intechopen.97871*

If this machine learning model can be fed by long term data, the engine performance under the same sea conditions can be predicted. In **Table 4**, the actual slip rates predicted by the model. The daily slip, from the noon reports which is daily given by ship captain, can be compared with the actual slip. In that way, by comparing the daily slip with actual slip, potential problems associated with ship performance could be spotted. Since there is not enough chief engineers who have experiences with alternative fuel powered vessels, this kind of machine learning algorithms is going to accelerate the experience accumulation in the technical department of the companies. Shipping market could be ready for an engine evolution, but the industry has not enough well-experienced engineers for this conversion.

If we can use the algorithms efficiently and feed the machine learning by real ship data, the developing models can be trained and after be used to give predictions and suggestion in a short time as well as well-experienced engineers working at ocean-going vessels. By intensive use of algorithms, the market can close the gap of the well-experienced engineers on alternative fuel powered engine.

**Table 6** was generated to demonstrate what kind of element can affect the bunkering operations. From the study visit to industry, some parameters were found. During the fuel transfer, there are many parameters which can affect the soundness of the operation;

	- Ship Tank
	- Bunker Tank
	- Line
	- Manifold
	- Weather
	- Ship Tank
	- Bunker Tank
	- Line
	- Manifold
	- Ship tank
	- Bunker tank (rest)


**Table 6.**

*ML Alternative*

 *Fuel Powered Vessels Data Table for ML* 

*Applications.*

## *Environmental Health*


These parameters are dynamic and thus frequently change due to inherent nature of the water. During the bunkering, assigned personnel observes the changes. Here, we should bear in mind the associated human errors. **Table 6** presents mentioned above main parameters that affect the bunkering operations.

By the use of ML algorithms, the shipping industry can learn about alternative fuels more and more in the future, and the **Table 6** is most likely to expand with new columns.
