Causal tracing method of return meter fault based on multisource data and machine learning

As the key equipment of electric energy metering, the installation scale of smart meters is growing rapidly, and the management mode of 8-year rotation will also produce a large number of returned meters. Fine fault detection and recycling of the return meter can produce huge economic benefits and environmental protection value. At present, the fault detection of the return meter mainly depends on manual work, which has low detection efficiency and insufficient accuracy. This paper proposes a causal tracing method for returned electric energy meters based on multi-source data and machine learning, which can intelligently identify the faulty components of returned electric energy meters without opening the cover, and can provide key technologies for the automatic recovery and disposal production line of returned electric energy meters. The method first performs a number of performance tests and power tests on the electric energy meter, and scans the broadband impedance of all exposed ports, and then trains the extreme random tree model to classify and judge the collected multi-source data, so as to determine the fault components. The experimental results show that this method has higher detection accuracy and reliability than the traditional manual detection method.