Performance Improvement of Electric Vehicle with Integrated Composite Power Systems and Power Estimation

The propose of this research is to improve the composite power system for electric vehicle. The proposed architecture is used lead-acid battery, lithium battery, and supercapacitor as a composite power supply pack as well as design a power source decision Algorithm based on electric power estimation method to improve the performance of composite power system. An electric forklift is used to test the proposed method, and the performance is compared in last step. The driving model is first established using pre-recorded electrical driving data such as current, and the discharge current estimation model is established using the MLP method. The discharge current estimation model established by the multilayer perceptron (MLP) method can predict the current trend and has only a slight delay. Using the regression method to estimate the accuracy of the estimated model, the regression equation between the true value and the estimated value has a slope of up to 0.9695 and an intercept of only 0.1751. The second step is the practical test. The performance test at this stage is divided into the original design, the composite power system without estimation function and the composite power system with estimation function. The last step is performance comparison, the weight of the forklift was reduced by 25% from 315 to 267 kg. With the appropriate control of the composite power pack combined with the estimation model, the maximum discharge current can be increased by 30%, and the power can be saved by 25% under the same test conditions.