摘要
在电力市场环境下,电价取决于众多因素的共同作用,它的演化过程呈十分复杂的不规则运动。为了揭示这种貌似随机的演化过程的内在规律,作者首先借助混沌理论,对电价的混沌特性进行了验证。在由电价单变量时间序列重构的相空间上,提取了吸引子的分形维数和Lyapunov指数,表明电价具有混沌特性;并且通过替代数据检验法进一步验证了电价的这种混沌行为,从而为借助混沌理论来进行电价的短期预测提供了依据。然后,采用电价及其相关因素构成的多变量时间序列重构了更为准确的相空间,通过跟踪相空间中相邻相点的演化趋势,建立起基于递归神经网络的全局和局域电价预测模型,并对NewEngland市场的电价进行了成功的预测。
In electricity market the price of electricity depends on the common effect of many factors and a very complicated random motion appears in its evolution process. To reveal the inherent law of such a seemingly random evolution process, firstly the chaotic feature of electricity price is verified by means of chaos theory, the Lyapunov exponents and the fractal dimensionalities of the attractors are extracted in the phase space reconstructed by the single variable time series of electricity price and therefrom it is indicated that the electricity price possesses chaotic characteristics, and such a chaotic behavior of electricity price is further verified by the verification method of substitutional data, thereby, the basis to perform the short-term forecasting of electricity price with the help of chaos theory. Then, a more accurate phase space is reconstructed by multi-variable time series constituted by electricity price and its correlated factors, i.e., the system load and available generating capacity time series. Through tracing the evolution trend of the adjacent phase points in the phase space, the global and local electricity price forecasting models based on the recursive neural network are established. With the established model the electricity prices in New England electricity market are successfully forecasted.
出处
《电网技术》
EI
CSCD
北大核心
2004年第3期59-64,共6页
Power System Technology
基金
教育部高等学校优秀青年教师教育与科研基金资助项目
关键词
电力系统
输电网
电价
混沌特性分析
预测模型
电力市场
Chaos theory
Costs
Electric load flow
Forecasting
Lyapunov methods
Marketing
Mathematical models
Neural networks
Time series analysis
