负荷预报中负荷规律性评价方法的研究

负荷预报是一个历久不衰的研究课题，以往的研究主要集中于预报方法，而缺乏对负荷自身变化规律的分析和评价。这不仅导致了难以公允地评价负荷预报方法，也使运行部门对负荷预报所提出的精度要求缺乏客观依据，该文阐述了评价负荷规律性的必要性，并建立了相应的评价方法，文中分析了负荷历史数据，负荷预报以及预报误差之间的关系，构建了一种基于统计分析的负荷规律性评价方法。在此基础上，建立了预估负荷预报误差极限的分析方法，运用所提出的方法对负荷变化的规律性进行评价，不仅可以客观地评价负荷预报方法的优劣，还可以根据不同负荷的规律性确定切实可行的负荷预报精度要求，文中给出的算例说明了方法的有效性。

STUDIES ON LOAD REGULARITY EVALUATING METHOD FOR LOAD FORECASTING

As an unflagging research project, load forecast research has be mainly concentrated on development of new forecast methods, but rarely on the effect of load regularity itself. It is evident that different loads in different region or even at different time period mean variant mode. Load varies with regularity, so it can be forecasted. Load is not exactly regular, and then the forecast error can not be entirely avoided. Disregarding the effect of load regularity results in not only the difficulty to fairly evaluate forecast methods but also the scarcities of objective evidence to determine the request of forecasting precision. In this paper, the necessity of load regularity evaluation has been clarified, and an evaluation method has been proposed. By using statistics theory, the evaluation of load regularity is based on the investigation of relationships among historical load data, load forecast mathematics model and forecast errors. The analysis of forecast error constitution can lead to estimation of the lower limit of load forecast error. By evaluating load regularity, load forecast methods can be objectively evaluated and a feasible request of forecasting precision can be determined according to individual load data. Some examples including loads in three different regions illustrate the effectiveness of the load regularity evaluation method.