为确保发电场正常供应电力,设计短时强降雨天气风电场发电功率预测模型,提升发电功率预测效果。通过欧式距离与角度原则扩充短时强降雨天气小样本;利用改进深度可分离卷积算法,在正常天气样本内,提取气象-功率时空特征,并输入长短期记...为确保发电场正常供应电力,设计短时强降雨天气风电场发电功率预测模型,提升发电功率预测效果。通过欧式距离与角度原则扩充短时强降雨天气小样本;利用改进深度可分离卷积算法,在正常天气样本内,提取气象-功率时空特征,并输入长短期记忆网络内,建立正常天气风电场发电功率基准值预测模型,得到发电功率基准值;在Transformer算法内,输入扩充样本,建立短时强降雨天气下发电功率损失值预测模型;利用基于注意力机制的Sequence to Sequence网络,结合扩样本,构造发电功率损失时间点预判模型,结合损失值预测模型,得到最终发电功率损失值;利用基准值减去损失值,得到短时强降雨天气下风电场发电功率预测结果。实验证明:该模型可有效扩充短时强降雨天气小样本;该方法可精准预判发电功率损失时间点,得到发电功率损失值,完成发电功率预测;不同风速下,该模型发电功率预测的关键失误指数与偏移程度均较低,即发电功率预测精度较高。展开更多
Nowadays manufacturers are facing fierce challenge.Apart from the products,providing customers with multiple maintenance options in the service contract becomes more popular,since it can help to improve customer satis...Nowadays manufacturers are facing fierce challenge.Apart from the products,providing customers with multiple maintenance options in the service contract becomes more popular,since it can help to improve customer satisfaction,and ultimately promote sales and maximize profit for the manufacturer.By considering the combinations of corrective maintenance and preventive maintenance,totally three types of maintenance service contracts are designed.Moreover,attractive incentive and penalty mechanisms are adopted in the contracts.On this basis,Nash non-cooperative game is applied to analyze the revenue for both the manufacturer and customers,and so as to optimize the pricing mechanism of maintenance service contract and achieve a win-win situation.Numerical experiments are conducted.The results show that by taking into account the incentive and penalty mechanisms,the revenue can be improved for both the customers and manufacturer.Moreover,with the increase of repair rate and improvement factor in the preventive maintenance,the revenue will increase gradually for both the parties.展开更多
文摘为确保发电场正常供应电力,设计短时强降雨天气风电场发电功率预测模型,提升发电功率预测效果。通过欧式距离与角度原则扩充短时强降雨天气小样本;利用改进深度可分离卷积算法,在正常天气样本内,提取气象-功率时空特征,并输入长短期记忆网络内,建立正常天气风电场发电功率基准值预测模型,得到发电功率基准值;在Transformer算法内,输入扩充样本,建立短时强降雨天气下发电功率损失值预测模型;利用基于注意力机制的Sequence to Sequence网络,结合扩样本,构造发电功率损失时间点预判模型,结合损失值预测模型,得到最终发电功率损失值;利用基准值减去损失值,得到短时强降雨天气下风电场发电功率预测结果。实验证明:该模型可有效扩充短时强降雨天气小样本;该方法可精准预判发电功率损失时间点,得到发电功率损失值,完成发电功率预测;不同风速下,该模型发电功率预测的关键失误指数与偏移程度均较低,即发电功率预测精度较高。
基金supported by the National Natural Science Foundation of China(71671035)。
文摘Nowadays manufacturers are facing fierce challenge.Apart from the products,providing customers with multiple maintenance options in the service contract becomes more popular,since it can help to improve customer satisfaction,and ultimately promote sales and maximize profit for the manufacturer.By considering the combinations of corrective maintenance and preventive maintenance,totally three types of maintenance service contracts are designed.Moreover,attractive incentive and penalty mechanisms are adopted in the contracts.On this basis,Nash non-cooperative game is applied to analyze the revenue for both the manufacturer and customers,and so as to optimize the pricing mechanism of maintenance service contract and achieve a win-win situation.Numerical experiments are conducted.The results show that by taking into account the incentive and penalty mechanisms,the revenue can be improved for both the customers and manufacturer.Moreover,with the increase of repair rate and improvement factor in the preventive maintenance,the revenue will increase gradually for both the parties.
