Fuel cell adaptive control method and system based on power prediction

Abstract

The invention discloses a fuel cell adaptive control method based on power prediction. The method comprises the steps: carrying out the prediction according to historical data in the working process of a fuel cell system to obtain a power prediction value, and adjusting the air supply amount according to the power prediction value; when power prediction is carried out, acquiring a feature parameter vector from historical data by using a feature parameter extraction and correlation analysis method, and calculating a power prediction value by using an artificial intelligence algorithm; and if the predicted power value of the fuel cell system is larger than the real-time output power, adopting response control action in advance to improve the air supply amount. According to the method, the RBF neural network is adopted to improve the power prediction accuracy, the calculation efficiency is improved by reducing the feature parameter dimension, and the variable load response speed of the fuel cell system can be remarkably improved.

Description

technical field [0001] The invention relates to the technical field of fuel cell systems, in particular to a fuel cell adaptive control method and system based on power prediction. Background technique [0002] Fuel cells have the advantages of high power generation efficiency, zero emissions, and low noise, and have broad application prospects in the fields of automobiles, drones, and stationary power generation. The stack is the core component of the fuel cell system, and usually forms a complete power generation system with auxiliary components such as hydrogen storage system, hydrogen circulation system, air supply system, thermal management system, humidifier, and DC / DC. In practical applications, the response speed of fuel cell discharge power is low, especially in the field of fuel cell vehicles. When performing fast load changes, the response speed of the fuel cell system is much lower than the demand of the drive motor, and must be controlled by the power battery. ...

AI Technical Summary

Problems solved by technology

The first control strategy is simple and easy to implement, but the operating efficiency of the system is low and the requirements for the power battery are high

The second control strategy is more complex, the system has high operating efficiency, but the response speed is slow when the load changes

[0005] However, most of the existing patents based on power prediction are aimed at predicting the power demand of electric loads. When the power demand of electric loads changes, the transmission of power demand commands to the fuel cell system still has the problem of slow response to load changes. question

Moreover, the existing power prediction model algorithm has high complexity, and in order to ensure real-time performance, high-performance hardware conditions are required, which is not conducive to reducing power consumption and production costs.

Images