The Assessment Of Intrinsically Safe Parameters Based On Artificial Intelligence

With the arising from the theory of explosion protection for electrical equipment, this paper discussed the significance of the explosion protection for electrical equipment used in modern industry. Since a variety of advantages apply on the intrinsically safe technology such as high reliability and safety, smart dimension, can be repaired during power on, easy manufacturing, etc, the intrinsically safe technology plays a very important role in the scope of explosion protection for electrical equipment. At present, the most common used method to test and evaluate the intrinsically safe equipment among domestic and foreigner test labs are dependent on the minimum ignition curve and spark test apparatus. With the specification for the basic principle of intrinsically safe technology, this paper analyzed the limitations of the traditional test methods mentioned above, exploring the principle of the ignition due to the electrical spark, listing all main factors that will cause the spark ignition and illustrating the non-linearity and non-determinacy among these factors.Artificial intelligence is a new scientific and technological undertaking, which can solve the complicated problems by simulating the mankind's thinking and behavior of biologic. Artificial neural network and generic algorithm are two important research fields of Artificial intelligence. By simulating the structure of human brain and the behavior of evolution, they can be used to find the solution to the nonlinear optimization problems those cannot be described by traditional mathematical models.An assessment model for intrinsically safe parameters is built by applying the back propagation neural network. Set the important parameter and functions of the BP network. In order to benefit the performance of the network, generic algorithm was introduced to optimize the initial weight and threshold value. The raw test data were collected by conducting the spark test. The intrinsically safe parameter assessment model was realized by applying the artificial neural network and generic algorithm toolbox embedded in Matlab. Finally, the performance of the model was verified via spark ignition test.