Sequential Pattern Mining In Enterprise Energy Consumption Warning

With the international energy shortages and energy price rises, energy managementhas become a hot research topic in recent years. As industrial enterprises become the mainbody of our national’s energy consumption， energy consumption supervision of industrialenterprises has become more and more important. Due to the dispersion of the energymanagement and the complexity of industrial processes, energy consumption data of industryshow timing, level and the regional characteristics of time, especially in process industry.Against the energy consumption characteristics of modern industrial engineering,mining potential energy anomalies law with advanced computer technology plays an importantrole in energy regulation.Due to the energy consumption characteristics of modern industrial process, the paperimproved data mining techniques to dig the warning sequence rules during industrial energyconsumption data and achieved good practical results.The main work and achievements are as follows:1. Due to the timing and level of industrial energy data, this paper studiedthe sequence mode of energy data, dug out early warning sequence of abnormal events, andhelped enterprise predict the occurrence of the energy anomalies. This paper studiedmulti-layer excavation strategy of ordinary set, extended multi-layer excavation strategy fromthe ordinary set mining to sequential pattern mining, compared several sequential pattern miningalgorithms, selected an appropriate sequential pattern mining algorithm and improved it enablewhich to be well adapted to multi-storey sequential patterns mining of industrial data. This paperalso took layer filter search strategies to avoid the excessive sequence collection.2. Due to characteristics for the time zone of the industrial energy consumption data, thispaper studied the way of time dimension division inspired by improved the ideas of Apriori algorithm,applied the idea of data region to sequential pattern mining, dug out sequentialpatterns for different time zone by industrial database division, and improved the efficiency ofmining sequential patterns.3. Based on the research and design on energy early warning system, this paper built basicframework of the energy early warning system, described basic functions of the energy earlywarning system briefly, studied the design of the monitoring points, the selection of earlywarning indicators and the design of the energy early warning methods, so that the energy early warning system could multi-level monitor energy consumption. Besides, this paper applied theimproved sequential patterns algorithm to the early warning system to obtain the energy earlywarning sequence of unusual event.4. This paper summarized the work done, pointed out the inadequacies of the research work,and made outlook on sequential pattern mining prospects for industrial applications.