Reconstruction Design Of Oiling System In An Enterprise

The original transformer oil management system is a manual control system.It is easy to make a mistake,the liquid level is unknowable and the transformer oil flow is uncontrollable.This paper,based on the computer technology and industrial bus technology,designs a transformer oil transformation system.This system uses the SIEMENS S7-300PLC as the main station and uses the distributed I/O controller ET200M as the slave station.Using the fieldbus PEOFIBUS communicate between main station and the slave station.The host computer monitoring software is MCGS,it is combined with the ACCESS database to design a monitoring system which has the function of monitoring the oil tank level and recording the data of each oil injection.For different types of transformers,there are strict requirements for oil flow.There is no overshoot in the oil flow rate of the transformer.If the oil flow rate is too fast,it will case a little acetylene in transformer oil.It is hard for the quality of transformers.In the hardware this system uses the frequency converter to control the gear pump to control the speed of the motor to control the flow.The traditional PID control strategy can't achieve the required control precision because of the interference and hysteresis factors,and it also can't determine the appropriate proportion,integral and differential parameters.On the country,the single neuron PID algorithm with adaptive ability can adjust the parameter on line,so this paper uses single neuron PID control algorithm to control transformer oil flow,and further discusses single the control effect in different learning rules of single neuron PID algorithm in MATLAB.Compare the control effect of several single neuron PID control strategy and draw the conclusion:the supervised HEBB single neuron PID control effect is the best.At last use LAD language to achieve the supervised HEBB single neuron PID control algorithm in PLC,The practical application shows that the single neuron PID control algorithm can meet the requirements of the factory.