Control Projects

Structural Design and Experiment with Fuzzy-PID Control

for the Hydraulic Pressure Control System

Background

This research is based on the industrial project aiming to design a on-land experimental device which can monitor the pressure environment in the deep ocean. Its target is to accurately control the pressure in a closed chamber device full of water. However, the mathematical model of this chamber device is difficult to construct. Since we use hydraulic-power system to control it, the strategy is critical and difficult to determine because of fluid's nonlinear property. Thus a experimental based method is proposed to solve this problem.

The following picture shows the typical structure of water-control system.

Traditionally, the similar project often used PID as its control strategy. Nevertheless, there exist disadvantages such as high following errors in the dynamic state and high overshoot in the stable state. Meanwhile, it is highly difficult to achieve a balance in both states with the same PID parameters based on its rigid controlling requirement. Thus, a controlling strategy with the adjustable parameters in PID based on Fuzzy is proposed.

Methods

Control Model

Simulation

• Based on the above comparison, Fuzzy-PID control strategy can decrease the overshoot, be faster to be in stable condition and obtain better dynamic control quality.

Implementation

Experimental Validation

Prototype

Design

Mechanical Design

The design of pressure-transfer device is the difficulty of the mechanical device. The challenge of using the hydraulic system(oil) to control the pressure is to solve the leakage problem under high pressure and its low viscosity. With the rubber capsule to transfer the pressure from oil to water, this device is designed to give a good solution.

Electric Design

The Electrical Schematic(sensor section)

Software Design

[Controller Design]

[GUI Design]

Experimental Device

Experimental Validation

Conclusion

• Control system using Fuzzy-PID has better quality in the stable condition, whose error is limited to 0.4 bar, and the system can follow the control target curve line without distortion and its error is limited to 2.5% FS in the dynamic condition(10-100bar).

Contribution

Research project is designed and tested on my own, including the component selection, mechanical design, software design, electrical design, and on-site testing experiment which lasted about 3 month in Yunnan Province of China.

As this research is related to an actual project, its target is to control the pressure of a closed container full of water, and its error should be limited in 2.5% of full scale according to the linear input.

As the research and development leader, I have completed the following critical things:

1. Built/ Simulated different algorithms including Adaptive PID based on Fuzzy and normal PID to verify their effectiveness based on initial model;
2. Designed the electrical schematic, built xPC real time communication with NI DAQ devices, and independently completed the test bed;
3. Designed the human-computer interface software with C# on host PC and built the target PC program with Simulink;
4. Verified the performances of algorithms on the test bed and obtain the conclusion.