Li Taifu

A Method of Synthetic Optimizing Adjustment Based on Fuzzy Genetic Arithmetic for Control of Electric Heater

The temperature control system of electric heater is a complex system. It has several characters such as nonlinear, parameter time variation, lots of control variables, controlled variables delay, etc. There exists interference caused by many uncertain factors. The traditional fuzzy controller can hardly ensure that the system controlling effect kept in ideal state all the time. Aimed at existent problem, a method of synthetic optimizing adjustment has been proposed by the research of fuzzy control method. Specifically, it includes membership function adjustment, parameter correction of analytic expression of fuzzy control rule, changeable regulation of the output membership function position, and the automatic optimized solving them with global optimization genetic arithmetic (GA) to make sure the system in the optimized state throughout the entire operation process. Not only the system reaches controlling quality but also has better controlling quality.

Research on liquid level detection based on acoustic field characteristic of the oil wells

In this paper, a detection method of oil well dynamic fluid level based on the sound field model of the column is presented. In this method, the typical string structure can be analyzed with the pipeline acoustic theory. Moreover, the relationship between the well dynamic fluid level depth and the resonance frequency of the tube sound field is established, according to the sound field characteristic of the column and resonance theory of tube air column, measuring the resonance frequency of the sound field in tube by Fast Fourier Transform (FFT), then liquid level depth of tube will be obtained. The experimental results show that the proposed protocol is practicable, providing a new way for the accurate detection of the well dynamic fluid level depth.

Remodeling of fuzzy PID controller based on BP neural network

The combination of widely-used PID controller with fuzzy system generates fuzzy PID controller which possesses excellent control quality. However, the fuzzy PID controller has some problems of computation complexity and real-time performance. To solve these problems, the paper expounds the process of training BP neural network through its universal function approximating ability and PID controllerpsilas input and output couple. Simulation researches reveal that the fuzzy PID controller can be effectively replaced by the trained BP neural network. Therefore, the method can simplify the computation complexity, enhance the real-time performance of fuzzy PID controller, and promote its implementation by hardware.