Research On The Detecting Algorithm For The Lead-acid Batteries' Capacity

One of the focusing issues concerning the lead-acid batteries is to estimate their state of charge (SOC) and furthermore, to support the users on the optimized decisions, for the residual capacity of the lead-acid batteries, an essential of the batteries'performance, limits the energy outputs of the batteries.By studying the electrochemistry theory, it is found that the sulfuric acid concentration, which is one the most remarkable factors influencing on the SOC of the lead-acid batteries, is difficult for accurate measurement in some specific conditions. It has been proved that there is a mathematic relationship between the electrode potential and the electrolyte concentration, which also transfers the information referred to the polarization of the battery plates partly. According to this, a new algorithm, based on the theory of soft- sensor, is developed to monitor and measure the SOC in this thesis.Based on other researchers'work, this thesis makes the sensors composed of a pair of battery electrodes and a silver-silver sulfate reference electrode, and maps out the experiment scheme to validate the algorithm.In fact, every parameter is dynamically changed in the charge or discharging process. But the classical methods can not identify this function precisely. That's why the BP net, a kind of Artificial Neural Network, is introduced in this research. A method to initialize the net's weights is introduced in order to improve the BP algorithm's efficiency and decrease the training cost, which has been approved by the simulation and the experiments.Although the algorithm has been validated exercisable both in Matlab's simulation and in the laboratory experiments, more experiments and data are still required to validate its universal applicability.As a reference, the detecting algorithm mentioned in this thesis is beneficial to the further research on the SOC of the lead-acid batteries, although it just makes an attempt of applying intelligent control theory in electrochemistry. The expectation and advice on the development of the detecting algorithm to measure the SOC of the lead-acid batteries are given at the end of this thesis.