The Design and Development of Micro Grid Electrical Power Supply for Seismo Sensor with An Artificial Perceptron Neural Network

Abstract

This research presented the application of the neural network theory in order to solve the problems of the 
electrical power supply system used in seismo sensor equipment with no continuity vibration and the 
system requiring the multiple power supply resources for the alternative power. Therefore, the designed 
system applied a perceptron mathematical model as a function to stimulate this system in making a 
decision to choose the micro grid power sources between solar power cells at 300 watts, 24 volts and the 
external power source at 720 watts, 24 volts. From testing its performance, it indicated that when the 
power supply source system did not use the artificial perception neural network, the system slowly 
responded to the decision - making time in an average time of 120 milliseconds while the ripple voltage 
was increased at peak to peak 0.5 volts. This could affect the power system to the seismo sensor which 
had no stability and continuity. However, it was applied to the system which was likely more effective and 
reliable due to having a decision-making from the data set, size 100. Moreover, the decrease of the delay 
time was over 0.8 which caused the ripple voltage at peak-to-peak 0.26 volt. As a result, it could conclude 
that the studies were able to help the electrical power supply system for a seismo sensor have more 
effectiveness and it could be adapted for the smart grid system in the future.