**Identifying Water Network Anomalies Using Multi Parameters Random Walk: Theory and Practice Parameters Random Walk: Theory and Practice**

**Identifying Water Network Anomalies Using Multi** 

DOI: 10.5772/intechopen.71566

Eyal Brill and Barak Brill Eyal Brill and Barak Brill Additional information is available at the end of the chapter

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/intechopen.71566

#### **Abstract**

A noise pattern analysis is used to demonstrate how water quality events can be classified. The algorithm presented mimics a random walk process in order to measure the level and type of noise in the water quality data. The resulting curve is analyzed and four different cases are identified. i.e. sensor problem, water source change, operational change and contamination. For each problem, the algorithm identifies a different pattern. This pattern can be used later to reduce the level of false alarms in the monitoring system.

**Keywords:** water network, abnormality detection, multi parameters, clustering, unsupervised learning
