An Online Early Warning Model Of Abnormal Energy Consumption In High Energy-consuming Equipment

There are many problems in the operation of high energy-consuming equipment,such as large energy consumption,low efficiency,extensive management and serious waste.The existence of these problems seriously consumes existing resources of our country.High energyconsuming equipment has great potential for energy saving.With the rapid development of Internet of Things,many high energy-consuming equipment have realized real-time collection of energy consumption data and accumulated a large amount of data,which is an important basis for assessing and analyzing equipment's energy consumption.However,for lacking of technology and management,there are many problem data and it is difficult to ensure the authenticity and accuracy of data.At the same time,a large amount of data is continuously accumulating in database,normal energy consumption is constantly changing,abnormal energy consumption is difficult to identify and the boundaries between normal data and abnormal data are extremely unclear.It is hard for managers to discover abnormal energy consumption timely and ultimately making it difficult for energy consumption data to exert its due value.Therefore,how to obtain useful information from a large amount of energy consumption data quickly and efficiently,find loopholes in energy using process of equipment timely and make equipment's energy consumption more reasonable and efficient become a key issue to be solved urgently in energy-saving work of high energy-consuming equipment.First step in reducing energy consumption of high energy-consuming equipment is to prevent energy waste due to inefficient operation of equipment or unreasonable energy use behaviors.Online early warning of abnormal energy consumption can detect abnormal energy consumption in using process of equipment.It is of great significance for timely detecting unreasonable use of equipment and management,effectively eliminating the occurrence of uncontrollable factors such as unplanned downtime caused by abnormalities,advancing energysaving emission reduction and promoting safe production.Therefore,this paper studies online early warning problem of abnormal energy consumption in high energy-consuming equipment.The main research work of this paper is as follows:(1)Classification method of energy consumption data and analysis of research problem: This paper deeply analyzes the characteristics of energy consumption data,the source of abnormal data,the typical anomaly contexts of energy consumption,the problems in the process of data collection?energy use and management and the complexity and solution idea of research problem.Based on this,we propose a classification method for energy consumption data.(2)An online early warning model of abnormal energy consumption for high energyconsuming equipment: During the operation of high energy-consuming equipment,energy consumption patterns change frequently,types of abnormal contexts are diverse and the characteristics of energy consumption data is nonlinear,multifactor,time-varying,high overlap and strong noise.To solve this difficult problem effectively,based on the combination of rules and data mining method,an online early warning model of abnormal energy consumption for high energy-consuming equipment with the objective of improving real-time,rationality and reliability of early warning is built.This model synthetically considers the impact of equipment operation,energy-use behavior and time-variation on energy consumption of equipment and it can identify different types of abnormal energy consumption data(3)Real-time monitoring method of abnormal energy consumption for high energyconsuming equipment: Different methods are designed to identify each type of abnormal energy consumption data.In order to reduce the size of search space and increase the speed of algorithm calculation,unsupervised learning and supervised learning are organically integrated.First,analyzing the implicit energy law in energy consumption data with unsupervised method.Then,learning and grasping the energy law with supervised method.Finally,analyzing energy consumption data with same energy consumption rules to determine the degree of abnormality.(4)Application research: An application research is conducted with energy consumption data of a laboratory high energy-consuming equipment.The result shows that our model can dynamically adapt to the change of energy consumption patterns,effectively sort,identify and replenish different types of abnormal data and report the abnormal condition real-timely.This paper provides a scientific and effective model for solving online early warning problem of abnormal energy consumption for high energy-consuming equipment under the environment of Internet of Things,which is good to improve real-time,rationality and reliability of online early warning.In practice,this model can be embedded in energy management system to help managers timely and effectively monitor the status of equipment energy consumption,discover irrationalities in equipment use and management and provide decision support for effective energy-saving measures and fault diagnosis.This model has important theoretical and practical significance.