Self-adaption electronic speed governing method for PID diesel engine

Abstract

The invention provides a self-adaption electronic speed governing method for a PID diesel engine based on an improved MAXQ algorithm, and relates to a diesel control method. The self-adaption electronic speed governing method for the PID diesel engine comprises the steps of setting a target rotary speed of a diesel based on an autonomous optimization S-MAXQ algorithm of the target rotary speed, using the target rotary speed of the diesel as the input of a self-organizing map neural network SOM, solving a PID parameter set corresponding to the target rotary speed required to be achieved at the current moment, using a Q value calculated by each subtask as the output of the SOM, acquiring a target position of a rack of the diesel through a rotary speed controller, calling the S-MAXQ to calculate a value function of a current speed governing strategy, and when the value function reaches to a maximum value, acting the value function to a rack position controller so as to adjust the rotary speed of the diesel. The diesel rotary speed is levelled off to a setting rotary speed according to an optimal speed governing index; the electronic speed governing adaptability of the diesel can be improved, and a diesel operating index is improved.

Description

technical field [0001] The invention relates to the technical field of industrial control, in particular to an autonomous control method of a diesel engine and a PID diesel engine self-adaptive speed regulation method based on the improved layered reinforcement learning MAXQ algorithm. Background technique [0002] The governor can detect the engine speed of the diesel engine, and generate the necessary force to adjust the fuel injection quantity according to the difference between the target speed and the actual speed. Therefore, the role of the governor is actually to control the speed. Due to the fluctuation and change of the speed of the diesel engine during the working process. Therefore, the governor must adopt closed-loop control to control the diesel engine speed. Closed-loop control first requires the accurate establishment of a mathematical model of the system. Because the structure and working process of the diesel engine are complex, there are many factors aff...

AI Technical Summary

Problems solved by technology

However, although the traditional PID control is convenient to use, easy to implement, and has no static error in the steady state, it has the following fatal flaws: First, the control parameters cannot be adjusted online after setting, so when encountering strong disturbances, recovery time is inevitable Phenomena such as extension, overshoot increase, etc., which restrict the improvement of the dynamic and static performance of the system; the second is that there is an integral saturation phenomenon when starting or large dynamic adjustment

It is applied to the digital electronic speed control system of diesel engine, and the simulation results show that the control algorithm not only has excellent dynamic performance, but also has good adaptability and robustness, which provides a basis for studying the intelligent control of digital electronic speed control system of diesel engine. A new way", its shortcoming is to use supervised learning for parameter optimization, and the teacher signal of supervised learning is difficult to obtain, in addition, this control method has no online learning ability, so its adaptability is poor

The patent with the publication number CN108008627 discloses an adaptive PID control method for parallel optimization and reinforcement learning. This method uses matlab to discretize the transfer function through the zero-order retainer method, and then uses the critic network for three-layer variable training. The method It solves the problem that the traditional PID controller is not easy to adjust parameters online in real time, but the three-layer training feedback of the Critic network also needs to calculate the final value function of reinforcement learning, which increases the complexity of the traditional PID algorithm

The patent with the publication number CN201510492758 discloses an adaptive PID control method for the actuator. The control method combines the expert PID controller and the fuzzy PID controller respectively connected with the actuator. The actuator selects the expert PID according to the current state information and expected information. controller or fuzzy PID controller, although this kind of controller can reduce the overshoot and has the characteristics of high control precision, but this kind of controller still needs a lot of prior knowledge of professionals to decide the use of the controller

Images