Study On Forecasting Of City Gas Load And Peak-Shaving Plan Decision

Carbon emissions,environmental pollution and other issues caused by coal,oil and other solid fossil fuels become increasingly serious.Combustion of natural gas has some advantages,such as,high calorific value,clean,no waste water,no residue et al.And In the call of national energy-saving emission reduction policy,natural gas as a clean energy has been widespread concern.The construction and operation of "West-East Gas Pipeline","Sichuan-East Gas Transmission" and other major national natural gas pipeline network engineering change the city's original single natural gas pipeline network gradually into a national large natural gas pipeline network.Obviously,this change will raise a higher demand for the upstream production capacity,the middle of the pipeline network transport capacity,the downstream city pipe network with gas capacity,gas storage capacity,and city natural gas peaking and scheduling.At the same time,when natural gas using peak appears,due to the larger gas consumption,there has been "no gas available" problem.It can be from two perspectives to solve the problem of "gas shortage" in the city,One is from the perspective of "gas supply side" and the other from the perspective of "gas consumption side".In this study,from the "gas consumption side" point of view,that is,a city builds its peaking facilities to achieve the purposes of "cutting peak filling valley".To formulate and decide the peak regulation on the basis of determining the city natural gas load from the perspective of "gas consumption side " which is mainly from the city natural gas "load" perspective(also can be expressed as " gas requirement","gas consumption",his article is called "load").The specific research work and the main conclusions are as follows:(1)The purpose of this study is to formulate and decide the peak regulation on the basis of the city natural gas load forecasting.Therefore,to analyze firstly the characteristics of hour,daily,year load of a city and peaking mode used in a city at present.The results show that the hour load of city natural gas has obvious periodicity,the daily load has obvious nonlinearity and the year load has monotonically increasing;Using high-pressure pipeline gas storage and LNG(Liquefied Natural Gas)peaking is a development trend of city peaking and the gas storage peak is gradually replaced.On the basis of this trend,city natural gas load forecasting model and peak-shaving scheme decision are researched.(2)A combined forecasting model based on Wavelet theory and RBF(Radial Basis Function)-Elman was proposed for city natural gas load forecasting.Firstly,the time series of city natural gas hour load was decomposed according to the wavelet theory;Secondly,RBF was used to predict the high frequency components,and Elman is used to predict the low frequency components;Finally,the wavelet was reconstructed.The influence of different order of different wavelet basis functions on the prediction results and the influence of the different decomposition layers of different wavelet basis functions on the forecasting results at the same order are analyzed.And the different forecasting models are compared.The results show that this model has higher forecasting performance than using RBF alone,using Elman alone and using Wavelet theory-RBF model.And it is an important improvement to the wavelet theory-RBF prediction model,and provides a reference for city natural gas load forecasting.(3)To analyze the chaotic characteristics of city natural gas load forecasting and reconstruct phase space.The Volterra adaptive filter was used to forecast it.To compare the prediction performance of the different order the single-step and multi-step of the filter.The results show that the hour load of city natural gas has the chaotic characteristics and the third-order single-step of Volterra adaptive filter has higher forecasting performance.This is a new research direction for the hour load of city natural gas forecasting.(4)EMD and EEMD were used to decompose the daily load time series of city natural gas;Secondly,the least squares support vector machine(LSSVM)and the sequence minimum optimization algorithm support vector machine(SMOSVM)were used to predict the decomposed components;Finally,reconstruction.The prediction accuracy of EMD-LSSVM,EMD-SMOSVM,EEMD-LSSVM and EEMD-SMOSVM prediction models was compared and in order to further improve their prediction accuracy,two methods are proposed:(1)Use PSO to optimize the parameters in LSSVM;(2)Use analytic modal decomposition to take a noise reduction of urban natural gas daily load;The approximate entropy calculation was used to calculate the components decomposed by EEMD,and the high noise component is subjected to secondary noise reduction using FIR digital filter.Calculations proves that: EEMD-LSSVM has higher prediction accuracy than EMD-LSSVM,EMD-SMOSVM and EEMD-SMOVM prediction models,and the two methods further improve the prediction accuracy of EEMD-LSSVM prediction model.It is an important improvement for the forecasting model on the basis of changing Wavelet and SVM and it create a new thinking for daily load of city natural gas forecasting.(5)The index system of city natural gas year load forecasting was constructed from the perspective of external environment,internal environment and user consumption,and the correlation coefficient method is used to optimize the index,then the forecast model of urban natural gas year load based on improved GA-BP neural network was established.Taking the year load of natural gas in Beijing and Shanghai as an example,this model was compared with the multiple regression,GBP and GA-BP forecasting models.The results show that the city natural gas year load forecasting model based on improved GA-BP neural network has higher prediction accuracy than the multiple regression,the GBP and GA-BP prediction models,which provides new idea for city natural gas year load forecasting.(6)To consider the quantitative index and qualitative index at the same time and give the relative membership degree calculation method of quantitative and qualitative indexes then improve the entropy value method to establish the decision-making model of city natural gas peak-shaving scheme based on improved entropy value method and multi-objective fuzzy decision-making theory.On the basis of this model,the natural gas load of Nanchang city was forecasted.TGNET simulation software was used to calculate the peak-shaving scheme considering only high-pressure pipelines.At the same time,the peak-shaving scheme of high-pressure pipeline and LNG joint was calculated and several peak-shaving schemes were decided in these two cases.The results show that the proposed model has good applicability and the optimal peak adjustment scheme in Nanchang is calculated which can provide the basis for any other cities that have not been built or partially constructed peaking facilities.