我国东北寒冷地区中长期电力负荷预测分析-analysis of medium and long term power load forecasting in cold regions of northeast china.docx

哈尔滨工业大学管理学博士学位论文 哈尔滨工业大学管理学博士学位论文 Abs Abstract - - PAGE XIV - - - III - 于机理研究的离子群优化支持向量机模型，并以此模型来进行寒冷地区中长 期电力负荷的预测。 在构建基于机理研究的离子群优化支持向量机模型的基础上，针对以往 电力负荷预测模型评价算法的不足，通过引入预测精确性指标、预测平稳性 指标、训练样本数指标、预测方法复杂程度指标、预测方法通用性指标和输 入变量个数指标共六个指标，构建了电力负荷预测方法的综合评价体系，以 此为依据来判断预测模型的优良性。以综合评价体系为依据，利用东北三省 的数据对寒冷地区中长期电力负荷进行了实证研究，检验了论文中提出的理 论模型及相关内容，研究表明基于机理研究的粒子群优化支持向量机模型可 以有效地进行寒冷地区中长期电力负荷预测。 论文在最后根据前述研究结果，应用基于机理研究的粒子群优化支持向 量机模型对黑龙江省“十二五规划”期间的电力负荷进行了预测。同时结合本 研究的相关结论，为我国寒冷地区中长期电力负荷预测提出了相关的战略对 策，同时在结论部分也指出了关于寒冷地区中长期电力负荷预测有待进一步 深入探讨的问题。 关键词：电力负荷预测；节能减排；支持向量机；粒子群算法；综合评价体系 Abstract The electrical industry is not only one of China’s most basic energy industry, but also the lifeline of our countrys economy. Our economy development, state security, social stability and the quality of peoples life are all dependent on the supply of secure power. Electric power should be the first is our country’s consistent power policy which is also the precondition of economic development. With the rapid development of our economy, the electrical industry has developed rapidly and the reform of power system also has been constantly deepening. However, the development of the electric power industry failed to keep up with the speed of Chinas economic development, mainly affected by the lax power construction ideas that didn’t develop the thermal power in 3 years since 1998.Power projects in 1998 is 10.21 million kilowatts ，in 1999 and 2000, is only six million kilowatts. But after2000, China has entered a new round of economic construction the peak, and GDP growth to a rate of about 10% per annum..In recent years, power shortage has been more and more severe which sharply affects the development of our economy. Medium and long term forecast of electric load is one of good solution to the issue. Many foreign and domestic scholars have explored deeply into the medium and long term prediction of electric load for a long time. Meanwhile, they also form a series of effective forecast