Research On Non-intrusive Load Monitoring Algorithm Based On Steady State Information

With the continuous development of technology,the proportion of electricity in end-use energy is growing.As global warming and energy shortage have become the two major issues in modern society.Several researches have indicated that human production activities are their primary driver.For a long time,governments and public institutions have been vigorously promoting energy conservation and emission reduction.Thus,effective use of electricity will bring significant energy savings.Indeed,some studies have indicated that providing comprehensive information of appliance power consumption can facilitate potential energy saving by more than20%.Moreover,the availability of power consumption information provides significant benefits for power system operation and scheduling and power grid planning.Thus,it can enhance asset utilization and energy efficiency.An important method for obtaining and analyzing power consumption information is non-intrusive load monitoring(NILM),which allows to infer the power consumption of individual appliance based on the aggregated power consumption measured at a single point instead of installing several meters.Therefore,NILM is regarded as a novel and cost effective technique for monitoring the details of load power consumption,gets rapid implementation in recent years.In order to extend NILM to all households equipped with smart meters.This paper proposes an extensible and comprehensive model to solve the load disaggregation problem,which employs the additive factorial approximate maximum a posteriori(AFAMAP)based on iterative fuzzy c-means(IFCM).To make sure that the model is adaptive to other households,the hidden Markov models(HMMs)are applied to obtain the independent load model of each appliance,and the IFCM is used to determine the number of hidden states adaptively.Then,the load models of all appliances are integrated into a database to apply in other household.Finally,the AFAMAP is utilized to decompose the aggregated power consumption based on the independent load models built by HMM.Simulation studies are conducted on the open database of Almanac of Minutely Power dataset(AMPds),and the results have demonstrated that the proposed model is more accurate in comparison with the other models.This paper carried out research in the following four aspects(1)The paper combines with the theoretical research of non-intrusive load monitoring,a factorial hidden Markov model(FHMM)for load disaggregation is established.The model uses the k-means clustering method to determine the power consumption corresponding to the operating state by pre-setting the number of clustering clusters,then HMM is used to train the individual load model.Finally,the load disaggregation is performed based on FHMM to obtain the final result.Six commonly used appliances were selected as the research object,to analyze the effectiveness of the method.(2)In order to improve the extensibility of non-intrusive load monitoring,the paper proposes IFCM to determine the number of operating states and corresponding power adaptively.Six common appliances are also selected as research objects to analyze the effectiveness of the method.At the same time,in order to verify that the number of operating states selected by the IFCM is the optimal,the number of operating states of the six appliance is changed one by one,and the accuracy is analyzed in turn.(3)In order to further improve the accuracy and efficiency of non-intrusive load monitoring,this paper abandons the traditional FHMM to disaggregate the aggregated power consumption signal.The AFAMAP model with the advantage of addictive Factorial Hidden Markov Model(AFHMM)and difference hidden Markov model(DFHMM)is used for load disaggregation.In view of the tedious computational complexity of AFAMAP,this paper proposes a non-overlapping window technique with adaptive threshold to determine whether the state of the window to be decomposed changes,the load disaggregation are conducted on the window that contains state change,and finally simulation studies are carried out for comparison.(4)In order to apply non-intrusive load monitoring better in reality,in addition to establishing a load model for individual appliance,this paper considers the Row(Rest of world)model in the noise situation,and finally AFAMAP is used to disaggregate the actual total power consumption.