Multi-objective regression modeling for natural gas prediction with ridge regression and CMARS

Residential customers are the main users generally need a great quantity of natural gas in distribution systems, especially, in the wintry weather season since it is particularly consumed for cooking and space heating. Hence, it ought to be non-interruptible. Since distribution systems have a restricted ability for supply, reasonable planning and prediction through the whole year, especially in winter seasons, have emerged as vital. The Ridge Regression (RR) is formulated mainly to decrease collinearity results through shrinking the regression coefficients and reducing the impact in the model of variables. Conic multivariate adaptive regression splines ((C)MARS) model is constructed as an effective choice for MARS by using inverse problems, statistical learning, and multi-objective optimization theories. In this approach, the model complexity is penalized in the structure of RR and it is constructed a relaxation by utilizing continuous optimization, called Conic Quadratic Programming (CQP). In this study, CMARS and RR are applied to obtain forecasts of residential natural gas demand for local distribution companies (LDCs) that require short-term forecasts, and the model performances are compared by using some criteria. Here, our analysis shows that CMARS models outperform RR models. For one-day-ahead forecasts, CMARS yields a MAPE of about 4.8%, while the same value under RR reaches 8.5%. As the forecast horizon increases, it can be seen that the performance of the methods becomes worse, and for a forecast one week ahead, the MAPE values for CMARS and RR are 9.9% and 18.3%, respectively.

Download Full-text

Distribution System Fault Recovery with Undispatchable Distributed Generations

Applied Mechanics and Materials ◽

10.4028/www.scientific.net/amm.529.455 ◽

2014 ◽

Vol 529 ◽

pp. 455-459

Author(s):

Nan Xu ◽

Shan Shan Li ◽

Hao Ming Liu

Keyword(s):

Wind Power ◽

Distribution System ◽

Distribution Systems ◽

Solar Power ◽

Equivalent Model ◽

Point Estimate ◽

Fault Recovery ◽

Recovery Method ◽

Multi Objective ◽

The Impact

Considering the probabilistic of the wind power and the solar power, a fault recovery method for distribution systems with the wind power and the solar power is presented in this paper. For the wind power, a simplified steady-state equivalent model of an asynchronous wind generator is added into the Jacobian matrix to consider the impact of the wind power on systems. For the solar power, its output is considered as an injected power which is related with solar irradiance. Three-point estimate is employed to solve the probabilistic power flow of distribution systems with the wind power and the solar power. The restoration is described as a multi-objective problem with the mean of the system loss and the number of switch operations. Fast elitist non-dominated sorting partheno-genetic algorithm is used to solve this multi-objective problem. IEEE 33-bus system is used as an example and the results show that the models and algorithms in this paper are efficient.

Download Full-text

Multi-objective optimal operation of integrated thermal-natural gas-electrical energy distribution systems

Applied Thermal Engineering ◽

10.1016/j.applthermaleng.2020.115951 ◽

2020 ◽

Vol 181 ◽

pp. 115951

Author(s):

Dariush Keihan Asl ◽

Ali Reza Seifi ◽

Mohammad Rastegar ◽

Mohammad Mohammadi

Keyword(s):

Natural Gas ◽

Energy Distribution ◽

Distribution Systems ◽

Electrical Energy ◽

Optimal Operation ◽

Multi Objective

Download Full-text

Study on the multi-objective optimization of reliability and operating cost for natural gas pipeline network

Oil & Gas Science and Technology – Revue d’IFP Energies nouvelles ◽

10.2516/ogst/2021020 ◽

2021 ◽

Vol 76 ◽

pp. 42

Author(s):

Kaituo Jiao ◽

Peng Wang ◽

Yi Wang ◽

Bo Yu ◽

Bofeng Bai ◽

...

Keyword(s):

Natural Gas ◽

Optimization Model ◽

Network Reliability ◽

Operating Cost ◽

Gas Pipeline ◽

Multi Objective Optimization ◽

Pipeline Network ◽

Multi Objective ◽

Natural Gas Pipeline ◽

The Impact

The development of natural gas pipeline network towards larger scale and throughput has urged better reliability of the pipeline network to satisfy transportation requirement. Previously, studies of optimizing natural gas pipeline network have been mainly focused on reducing operating cost, with little concern on the reliability of pipeline network. For a natural gas pipeline network with a variety of components and complicated topology, a multi-objective optimization model of both reliability and operating cost is proposed in this study. Failure of each component and the state of pipeline network under failure conditions are taken into account, and minimum cut set method is employed to calculate the reliability of the pipeline network. The variables to be determined for the optimization objectives are the rotating speed of compressors and the opening of valves. Then the solving procedure of the proposed model is presented based on Decoupled Implicit Method for Efficient Network Simulation (DIMENS) method and NS-saDE algorithm. The validity of the optimization model is ascertained by its application on a complicated pipeline network. The results illustrate that the optimization model can depict the relative relationship between reliability and operating cost for different throughput, by which the operation scheme with both satisfying reliability and operating cost can be obtained. In addition, the customer reliability and the impact of the failure of each pipeline on the whole network can be evaluated quantitatively to identify the consumers and pipelines of maintenance priority. The pipeline network reliability can be improved through proper monitoring and maintenance of these consumers and pipelines.

Download Full-text

Natural Gas Consumption Forecasting Based on the Variability of External Meteorological Factors Using Machine Learning Algorithms

Energies ◽

10.3390/en15010348 ◽

2022 ◽

Vol 15 (1) ◽

pp. 348

Author(s):

Wojciech Panek ◽

Tomasz Włodek

Keyword(s):

Machine Learning ◽

Random Forest ◽

Natural Gas ◽

Learning Algorithms ◽

Weather Conditions ◽

Machine Learning Algorithms ◽

Forecast Horizon ◽

Gas Consumption ◽

Natural Gas Consumption ◽

The Impact

Natural gas consumption depends on many factors. Some of them, such as weather conditions or historical demand, can be accurately measured. The authors, based on the collected data, performed the modeling of temporary and future natural gas consumption by municipal consumers in one of the medium-sized cities in Poland. For this purpose, the machine learning algorithms, neural networks and two regression algorithms, MLR and Random Forest were used. Several variants of forecasting the demand for natural gas, with different lengths of the forecast horizon are presented and compared in this research. The results obtained using the MLR, Random Forest, and DNN algorithms show that for the tested input data, the best algorithm for predicting the demand for natural gas is RF. The differences in accuracy of prediction between algorithms were not significant. The research shows the differences in the impact of factors that create the demand for natural gas, as well as the accuracy of the prediction for each algorithm used, for each time horizon.

Download Full-text