Northern China has rich wind power and photovoltaic renewable resources. Combined Heat and Power (CHP) Units to meet the load demand and limit its peaking capacity in winter, to a certain extent, it results in structu...Northern China has rich wind power and photovoltaic renewable resources. Combined Heat and Power (CHP) Units to meet the load demand and limit its peaking capacity in winter, to a certain extent, it results in structural problems of wind-solar power and thermoelectric. To solve these problems, this paper proposes a plurality of units together to ensure supply of heat load on the premise, by building a thermoelectric power peaking considering thermal load unit group dynamic scheduling model, to achieve the potential of different thermoelectric properties peaking units of the excavation. Simulation examples show, if the unit group exists obvious relationship thermoelectric individual differences, the thermal load dynamic scheduling can be more significantly improved overall performance peaking unit group, effectively increase clean energy consumptive.展开更多
针对分布式电源和新型负荷容量累积造成负荷影响因素多元化和不确定性特性增强的问题,文中提出一种采用记忆神经网络和曲线形状修正的负荷预测方法。在负荷峰值预测中,采用最大信息系数计算负荷峰值与影响因素的非线性相关性,实现对输...针对分布式电源和新型负荷容量累积造成负荷影响因素多元化和不确定性特性增强的问题,文中提出一种采用记忆神经网络和曲线形状修正的负荷预测方法。在负荷峰值预测中,采用最大信息系数计算负荷峰值与影响因素的非线性相关性,实现对输入特征的筛选;综合考虑负荷峰值序列的长短期自相关性和输入特征与负荷峰值的不同程度相关性,结合Attention机制和双向长短时记忆(bidirectional long short-term memory,BiLSTM)神经网络建立负荷峰值预测模型。在负荷标幺曲线预测中,通过误差倒数法组合相似日和相邻日,建立负荷标幺曲线预测模型;针对预测偏差的非平稳特征,利用自适应噪声的完全集成经验模态分解和BiLSTM网络建立误差预测模型,对曲线形状进行修正。应用中国北方某城市的区域电网负荷数据为算例,验证了所提模型的有效性。展开更多
介绍了多直流馈入背景下上海电网的基本情况,分析了电网运行中的低谷调峰困难,从4个方面讨论了上海电网最小开机方式研究的约束条件。并在综合考虑电网调峰需求和电压稳定要求的基础上,对上海电网不同负荷水平下的开机方式进行了分析研...介绍了多直流馈入背景下上海电网的基本情况,分析了电网运行中的低谷调峰困难,从4个方面讨论了上海电网最小开机方式研究的约束条件。并在综合考虑电网调峰需求和电压稳定要求的基础上,对上海电网不同负荷水平下的开机方式进行了分析研究,结果表明:上海电网负荷水平3 100万k W,最小开机方式为500 k V机组开6台,220 k V机组开11台;负荷水平2 700万k W,500 k V机组开4台,220 k V机组开8台;负荷水平1 900万k W,500 k V机组开3台,220 k V机组开3台。展开更多
文摘Northern China has rich wind power and photovoltaic renewable resources. Combined Heat and Power (CHP) Units to meet the load demand and limit its peaking capacity in winter, to a certain extent, it results in structural problems of wind-solar power and thermoelectric. To solve these problems, this paper proposes a plurality of units together to ensure supply of heat load on the premise, by building a thermoelectric power peaking considering thermal load unit group dynamic scheduling model, to achieve the potential of different thermoelectric properties peaking units of the excavation. Simulation examples show, if the unit group exists obvious relationship thermoelectric individual differences, the thermal load dynamic scheduling can be more significantly improved overall performance peaking unit group, effectively increase clean energy consumptive.
文摘针对分布式电源和新型负荷容量累积造成负荷影响因素多元化和不确定性特性增强的问题,文中提出一种采用记忆神经网络和曲线形状修正的负荷预测方法。在负荷峰值预测中,采用最大信息系数计算负荷峰值与影响因素的非线性相关性,实现对输入特征的筛选;综合考虑负荷峰值序列的长短期自相关性和输入特征与负荷峰值的不同程度相关性,结合Attention机制和双向长短时记忆(bidirectional long short-term memory,BiLSTM)神经网络建立负荷峰值预测模型。在负荷标幺曲线预测中,通过误差倒数法组合相似日和相邻日,建立负荷标幺曲线预测模型;针对预测偏差的非平稳特征,利用自适应噪声的完全集成经验模态分解和BiLSTM网络建立误差预测模型,对曲线形状进行修正。应用中国北方某城市的区域电网负荷数据为算例,验证了所提模型的有效性。
文摘介绍了多直流馈入背景下上海电网的基本情况,分析了电网运行中的低谷调峰困难,从4个方面讨论了上海电网最小开机方式研究的约束条件。并在综合考虑电网调峰需求和电压稳定要求的基础上,对上海电网不同负荷水平下的开机方式进行了分析研究,结果表明:上海电网负荷水平3 100万k W,最小开机方式为500 k V机组开6台,220 k V机组开11台;负荷水平2 700万k W,500 k V机组开4台,220 k V机组开8台;负荷水平1 900万k W,500 k V机组开3台,220 k V机组开3台。
