Application of Paraconsistent Artificial Neural Network in Statistical Process Control acting on voltage level monitoring in Electrical Power Systems

In this paper, we present an application of the Paraconsistent Artificial Neural Network (PANnet) in a Statistical Process Control to trigger alarms in an electrical power System. The PANnet is based on Paraconsistent Annotated Logic (PAL) which is a non-classical logic with properties of accepting contradictions in their fundamentals. In this work, we make a joint application technique of Paraconsistent Annotated logic, in the form of PANnet, with the concepts of Statistical Process Control (SPC). The Statistical Process Control Paraconsistent (SCP-PAL) is used in this paper to make a dynamic monitoring of the conditions of the electrical voltage in an electrical power system. To test and validate the functioning of SPC-PAL Controller we use data obtained by random computational processes and we also use a database with actual values generated by a power electrical System of a company installed in Brazil. The SPC-PAL Controller, due to the use of the algorithms of PAL in its construction, has some features that does afford optimized electrical voltage monitoring. The result of the analysis made with the SPC-PAL Controller provides three types of alarms: 1) rapid Variation in the voltage; 2) Imbalance of Mean in the Voltage data; 3) High dispersion index in the voltage data. Based on these three types of alarms we can obtain a better understanding about the operating state of the electrical power system. In the various tests performed, the SPC-PAL Controller presented a good performance and detected changes in Mean and in the variation of tension, as well as sudden changes in voltage data.