| Metering pump as a branch of reciprocating pump is widely used in the industries of petroleum, chemical, water treatment, food, pharmaceutical, environmental, medical device, and has become the heart an engine of the process industry.With the expansion of application of metering pumps, higher requirements of precise control and remote control of metering pumps has been put forward. In recent years, the development of intelligent control provides the theory basis of the precise control of metering pump. In this paper, the neural network control technology is applied to the metering pump controlling.The metering pump system is droved by a three-phase asynchronous motor, this paper build the mathematical model of motor, design a single neuron PID controller, use the MATLAB/Simulink program to build the simulation model. The simulation results show that the metering pump frequency control system with the single neuron PID controller has a good acceleration and speed performance. The paper also put the intelligent control cross-application into the metering pump controlling, design a neuron-fuzzy network controller. The simulation results show that the neural network is able to adjust the membership function of the fuzzy controller by self-learning, which improves the performance of motor speed control system.The system hardware is designed to two boards, the control board which is a weak electricity board and the driver board which is a force electricity part. Chip disPIC30F6010A is selected to be the processor. It is a chip designed to control the motor. Using this chip and other circuits, the system achieves the human-computer interaction function, network communication function, speed acquisition function and so on.The system software design contains the key module, LCD module, PWM module, CAN module and so on. In order to meet the industrial filed for the metering pump group network operations, the CAN communication module is enhanced designed. All the key module function can be achieved by the CAN communication module. |
