Research On Electrothermal Overload Current Measurement System Of Small Circuit Breaker Based On Neural Network

In recent years,with the development of the smart grid,the intelligentization of circuit breakers has become an inevitable trend of development.In the field of low voltage,the intelligent products of air circuit breakers and molded case circuit breakers have a large market share,their intelligent technology is relatively mature,but the most widely used small circuit breakers rarely see their intelligence related reports.Most intelligent circuit breakers use electronic trip devices.The core technology is to use a current transformer to measure the current.However,the current transformers on the market are relatively large and the internal space of small circuit breakers is small.Therefore,the method of using electronic current transformers to design electronic release devices is not suitable for small circuit breakers.At present,the intelligent research of small circuit breakers is restricted by the miniaturization and other factors.Therefore,for small circuit breakers,how to use new ideas to develop current detection methods suitable for smaller spaces has realistic research significance and broad development prospects.This thesis uses the thermal effect of current to propose an electrothermal current detection method,which not only solves the problem of intelligence of small circuit breakers,but also ensures the miniaturization of the volume.This thesis applies artificial neural network to the research of current detection.First,this thesis analyzes the influencing factors of the temperature rise of the miniature circuit breaker,excludes the interference of other factors,and determines that the current is the main influencing factor of the temperature rise of the miniature circuit breaker.This thesis uses DZ47-60 miniature circuit breaker as the measurement prototype.In order to prevent the small metal circuit breaker from tripping due to the heating and bending of the bimetal when passing the current,the bimetal is cut off before the experiment,and the rest of the internal structure is retained.In this way,the heating and cooling conditions inside the small circuit breaker are closer to actual work Situation.Set up a LabVIEW experimental measurement platform.This platform consists of a voltage regulator,a voltage regulator source,a three-phase dry-type transformer,a temperature transmitter,and a current transformer.It collects temperature data corresponding to different currents for modeling.Then,based on the single hidden layer BP neural network model,a double hidden layer BP neural network model was created,and the approximate structure of the network was determined using empirical formulas.The final current detection model based on BP neural network was determined by optimizing the network parameters one by one.Finally,in order to further reduce the current prediction error,the established model was optimized by using genetic algorithm,and a current detection model based on GA-BP neural network was established on the basis of finding the optimal weight-threshold combination.The current detection model is verified by using the data collected in the experiment.It can be seen from the verification results that the current detection model created in this thesis has high prediction accuracy and can meet the intelligent design requirements of small circuit breakers.