双向全桥DC-DC变换器故障智能检测系统设计

Design of fault intelligent detection system for bidirectional full-bridge DC-DC converter

传统DC-DC变换器故障检测系统只对变换器中某重要元器件进行故障检测,导致检测结果不准确。为此,设计了新的双向全桥DC-DC变换器故障智能检测系统。通过闭环霍尔电流传感器对电流信号进行采集,给出信号采集电路,通过分析双向全桥DC-DC变换器元件衰减过程提取故障数据,对不同故障状态样本进行采集,构建HMM模型,求出HMM模型初始参数,迭代参数直至其收敛到既定区间,对采集的电压、电流信号进行归一化处理,产生观察值序列,把观察值序列输入到训练完成的不同HMM模型中,求出相似或然率和相似或然率最大值相应的HMM模型,即为双向全桥DC-DC变换器故障检测结果。实验结果表明所设计系统检测准确性高。

The traditional fault detection system of the bidirectional full-bridge DC-DC converter can only detect the fault of a certain important component in the converter,and its detection result is inaccurate. Therefore,a new intelligent fault detection system for bidirectional full-bridge DC-DC converter was designed. The current signal is acquired through closed-loop Hall current sensor. The signal acquisition circuit is given. The component attenuation process of bidirectional full-bridge DC-DC converter is analyzed to extract the fault data. The samples with different fault conditions are acquired. The hidden Markov model(HMM)is constructed,and its initial parameter is determined. The parameter is iterated until it is converged to the established section.The acquired voltage and current signals are normalized to generate the observing values sequence. The observing values sequence is input into different HMMs after training to get the similarity probability and the HMM corresponding to the maximum similarity probability,whose output is the fault detection result of a bidirectional full-bridge DC-DC converter. The experimental results show that the designed system has high detection accuracy.