光伏指数保险是一种基于太阳辐射总量的天气指数保险产品,旨在为光伏电站提供发电量不足的风险保障。然而,光伏指数保险中存在着因缺乏客观数据与评价标准而导致的保险欺诈问题。本文针对这一问题,提出了一种基于深度学习的模型,结合LST...光伏指数保险是一种基于太阳辐射总量的天气指数保险产品,旨在为光伏电站提供发电量不足的风险保障。然而,光伏指数保险中存在着因缺乏客观数据与评价标准而导致的保险欺诈问题。本文针对这一问题,提出了一种基于深度学习的模型,结合LSTM、注意力机制和残差连接,该模型使用太阳辐射量对光电企业发电量进行预测。为验证模型的有效性,本文使用2019年新疆光电光伏数据集进行测试并与传统的最小二乘法的回归预测进行比较。测试结果表明,本文所提出的模型可以有效使用太阳辐射量对光电企业发电量进行预测,进而帮助保险企业针对光伏企业的发电量有关的保险欺诈行为进行识别,为光伏保险业务的规范化和发展提供了一种有效的技术手段。Photovoltaic (PV) index insurance is a weather index insurance product based on the total amount of solar radiation, designed to provide risk protection against insufficient power generation for PV power plants. However, PV index insurance faces issues of insurance fraud due to the lack of objective data and evaluation standards. To address this problem, this paper proposes a deep learning-based model that combines LSTM, attention mechanism, and residual connections. This model uses solar radiation data to predict the power generation of photovoltaic enterprises. To validate the effectiveness of the model, the 2019 Xinjiang photovoltaic dataset was used for testing and compared with traditional least squares regression predictions. The test results show that the proposed model can effectively use solar radiation data to predict the power generation of photovoltaic enterprises, thereby helping insurance companies identify fraudulent activities related to power generation in PV enterprises. This provides an effective technical means for the standardization and development of PV insurance business.展开更多
文摘光伏指数保险是一种基于太阳辐射总量的天气指数保险产品,旨在为光伏电站提供发电量不足的风险保障。然而,光伏指数保险中存在着因缺乏客观数据与评价标准而导致的保险欺诈问题。本文针对这一问题,提出了一种基于深度学习的模型,结合LSTM、注意力机制和残差连接,该模型使用太阳辐射量对光电企业发电量进行预测。为验证模型的有效性,本文使用2019年新疆光电光伏数据集进行测试并与传统的最小二乘法的回归预测进行比较。测试结果表明,本文所提出的模型可以有效使用太阳辐射量对光电企业发电量进行预测,进而帮助保险企业针对光伏企业的发电量有关的保险欺诈行为进行识别,为光伏保险业务的规范化和发展提供了一种有效的技术手段。Photovoltaic (PV) index insurance is a weather index insurance product based on the total amount of solar radiation, designed to provide risk protection against insufficient power generation for PV power plants. However, PV index insurance faces issues of insurance fraud due to the lack of objective data and evaluation standards. To address this problem, this paper proposes a deep learning-based model that combines LSTM, attention mechanism, and residual connections. This model uses solar radiation data to predict the power generation of photovoltaic enterprises. To validate the effectiveness of the model, the 2019 Xinjiang photovoltaic dataset was used for testing and compared with traditional least squares regression predictions. The test results show that the proposed model can effectively use solar radiation data to predict the power generation of photovoltaic enterprises, thereby helping insurance companies identify fraudulent activities related to power generation in PV enterprises. This provides an effective technical means for the standardization and development of PV insurance business.
