Study On The Modeling And Its Application Of Multidimensional Conditions Influence To Measurement Performance Of Electrical Energy Meters

Electrical energy meters have been widely instructed in grid as terminal metrology instrument, and their accuracy is associated with the equity of power trade. The national standard of electrical energy meters largely ensures the accuracy of electrical energy meters, but they just includes the reference conditions or single condition (the influence quantity test). The test conditions don’t consist of multi-dimension conditions in working mode. In order to further study the mea surement performance of electrical energy meters and ensure the accuracy of electrical energy meters under multi-dimension working conditions, this paper makes a series theoretical and experimental study and build an evaluation model of electrical energy meter that can apply to effect of multi-dimension conditions in working mode. This paper has done several work below:(1) Made a summary of study background on multidimensional conditions to the measurement performance of electrical energy meters, and analyzed the current study modeling methods of electrical energy meters. (2) Built a model based on back propagation neural network to synthetically response the effect of multidimensional conditions on the measurement performance of electrical energy meters. The preliminary inputs of model are voltage, current, power factor, temperature, humidity and pressure, while the output of model is measurement error. (3) Constructed multidimensional test platform, and made an investigation on the effect of humidity and pressure coupling with temperature to measurement performance of electrical energy meters. Test showed that humidity and pressure had little effect on the measurement performance of electrical energy meters, and the final inputs of model are voltage, current, power factor and temperature. (4) Used data of the basic error test, voltage and temperature influence quantity test points to complete the division of the model training samples and test samples, and built the model by the training samples. Finally made a performance test, and made practical application of the modeling method to other single phase electrical energy meters and three phase gateway electrical energy meters. (5) Through the modeling method of electrical energy meters, we provided a supplement for the calibration of electrical energy meters in working conditions to ensure the accuracy of verification work under working conditions.