The Analysis And Implementation Of Heating Boiler’s Optimization Method Based On Time Series Data Mining

In recent years, a wide range of fog and haze weather appeared in China’s centraland eastern regions, the increasing emissions of pollutant from coal-fired heating causedby low temperature aggravate environmental pollution. However, there are a lot ofproblems exist in the combustion equipment of our heating system: equipment obsoletedue to low update speech, low penetration rate of digital control system and the manualcontrol can hardly meet the need of precision require, which contribute to lowcombustion efficiency of coal-fired and causing serious environmental problems.Therefore, it’s very important to help the managers to make control decisions which canimprove the combustion efficiency of the boiler system and achieve the purpose ofenergy saving.The heating boilers typical multiple input and output, strong non-linear,non-self-balanced complex system, the fluctuation of the boiler combustion parameterssuch as drums or lead air flow, negative pressure in furnace, furnace temperature, theamount of coal and other parameters have directly or in direct impact to boilercombustion system, which is very inconvenient to create accurate mathematical modelto simulate the system, thus, there still no satisfactory solution to the boiler optimizationproblems. Moreover, due to the large heating area and long heating pipeline, there is asignificant delay of the heating effect of the parameter (return water temperature)adjustment, thus, is not easy to make timely adjustment of the control for the currentheating needs. In addition, different need for the boiler heating effect and differentstandards for of working parameters of the boiler system also need to tack into accountwhen design the boiler optimization system.Based on the carefully analyze and study to the boiler system as well as domesticand foreign study, applied data mining techniques to the optimization of heating boilersby using the characteristic of time series of heating boilers parameters, this papercompleted the following studies:(1)Analyzing the concrete need of heating boiler system, extract and study thereal-time monitoring date and artificial records, and then establishing the time-series database of the heating boiler optimization system.(2)Studying the time-series date mining theory, using support vector regressiontime-series forecasting model to forecasting and simulate the trend of the boilerparameters.(3) Designing and implementing the functions inlcluding the work conditionsforecasting, the backwater temperature delay decision-making and the periodicoptimization of heating system, according to the simulative results.This subject analysis the boiler data of a Co-generation Center in Shandongprovince, combined with the many years of work experience of the Center’s experts,analysis the time-series data using date mining method, established the special timeseries featuring sub-function systems. The final system can provide effect trend forecastand recommended decisions for the Heating Center stall and then fulfilled the controloptimization and energy saving purpose.