用电价分布概率预测的发电商报价策略模型

电力市场环境下发电企业的报价策略直接影响着该企业的收益。以市场边际电价报价既可成功竞价上网又能使发电利润达到最大,是最理想的报价方案。波动剧烈的实际市场电价很难精确预测,但可预测出其可能的分布概率。为此建立了一个基于电价分布概率预测的发电企业报价策略模型,并提出了采用二维实数编码策略的改进遗传算法优化求解。针对市场竞价中申报数据要求为一组符合递增规律的“价格—电量”数据点的要求,特别设计了独特的遗传操作算子以得到合理有效的报价方案。算例仿真验证了该模型与算法的可行性。

Bidding Model Based on Forecasting the Probability of Electricity Price Distribution for Power Generation Enterprises

In a power market, the profit of the power supplier is influenced directly by the bidder's bidding strategies. The most perfect bidding plan is to bid according to the market marginal price, which can both ensure the power supplier to succeed in bid competition and to maximize the total revenue. As we all know, the electricity price is strongly fluctuant since it is dependent on the uncertainty conditions such as climate, system load, availability of generation units and so on, so it is difficult to forecast accurately. However, forecasting the distribution probability of the future electricity price is doable. Accordingly, the paper presents a bidding model based on forecasting the probability of the electricity price distribution. An improved genetic algorithm adopting the two- dimensional code strategy is applied to solve the problem. To meet the particular requirements of the "electric price - electric quantity" data which should be monotone increased in the bidder declaring , the special genetic operator is designed. By this method, the two decision variables of bidding price and output can be simultaneously optimized. At the end, the computation and analysis results of a numerical example are illustrated in detail, verifying the effectiveness and practicability of the proposed model and optimization algorithm. Results show that adopting the competitive model may be helpful for the power supplier in making bidding decision.