A Method of Load Prediction in District-Heating System Based on Data Mining

Abstract:

According to the internal mechanism of the formation of heat load, the formation of heat load consists of two parts, the systemic heat load, which is determined by the building envelope and outdoor environmental parameters and random load caused by the users randomness of events and solar radiation etc. Toward systemic heat load, this paper considered the influence of environmental parameters before the prediction time and used the method of stepwise trials and MSE to obtain the optimal solution. Toward random load, it is considered that the day of the same type have the same variation pattern. On this basis, this paper introduced a correction coefficient to obtain random load eventually. This paper selected DeST, the widely used energy simulation software in China, to analysis the case. The result shows that the prediction method is feasible and 50% of the predicted loads have the relative error of less than 5%.

• Periodical: Advanced Materials Research (Volume 918)
• Pages: 154-159
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.918.154
• Online since: April 2014
• Authors: Lan Bin Liu*, Ai Juan Zou, Yu Fei Ma
• Keywords: Data Mining (DM), District-Heating System, Load Prediction

References

[1] Dotzauer, Erik. Simple model for prediction of loads in district-heating systems. Applied Energy 73. 3 (2002): 277-284.

DOI: 10.1016/s0306-2619(02)00078-8

[2] Sen Zou. Adaptive short-term load forecasting based on time series. Journal of Shandong Polytechnic University 18. 2(1988): 43-49. (In Chinese).

[3] Gadd, Henrik, and Sven Werner. Daily heat load variations in Swedish district heating systems. Applied Energy 106 (2013): 47-55.

DOI: 10.1016/j.apenergy.2013.01.030

[4] Catalina, Tiberiu, Vlad Iordache, and Bogdan Caracaleanu. Multiple regression model for fast prediction of the heating energy demand. Energy and Buildings (2012).

DOI: 10.1016/j.enbuild.2012.11.010