Advanced AI for Integrated Financial Optimization of Wind Energy Assets

用于风能资产综合财务优化的先进人工智能

• 批准号: 77838
• 金额: $ 7.63万
• 依托单位: POWER ENABLE SOLUTIONS LIMITED
• 依托单位国家: 英国
• 项目类别: Small Business Research Initiative
• 财政年份: 2020
• 资助国家: 英国
• 起止时间: 2020 至 无数据
• 项目状态: 已结题
• 来源: https://gtr.ukri.org/projects?ref=77838
• 关键词: Advanced; AI; Integrated; Financial; Optimization

REOptimize Systems (REOS), was formed to exploit research developed at University of Edinburgh, which has implemented a unique approach to increasing the efficiency of wind turbines. Through advanced modelling and the application of novel machine learning techniques the algorithms minimise the end-to-end losses in the system. This technique has patents pending and is the result of 7 years of research at The University of Edinburgh. The success of the algorithms has been proven experimentally in small-scale wind turbines, and found to yield increases in energy capture of 6%. A 6% increase in energy capture can drive net profit increases for the operator on the order of 50%-100% depending on the specific turbine and location. If only half of UK turbines achieved a 6% gain, it would result in an additional 3000 GWhr of generation and a saving of 1.3 million tonnes of carbon in a single year. This is equivalent to removing around 290,000 petrol passenger cars from the streets. However, this 6% gain has been proven only in medium-scale wind at approximately 100 kW ratings. It is expected that larger turbines will start from a position of better control which will allow us to achieve gains on the order of 3%. REOS is currently preparing a pilot project to validate the technique on a MW scale Siemens 2.3-92 turbine, which is a workhorse of the UK onshore fleet. Now, through this new project, REOS will develop and integrate novel machine learning technologies into a single platform which will provide end-to-end financial optimization of wind power assets, with a truly holistic view of the entire wind system. The project will develop and integrate: * Continuous per-turbine settings optimization * Advanced detection of false alarms to increase in-service time * Advanced AI-based wind farm control This will create a step-change in the control and performance of wind energy assets with the aim of maintaining the gain of 6% increase in energy output in large modern wind farms. This will contribute to creating sustainable innovation and help deliver the transition to net-zero.

re优化系统公司（REOS）的成立是为了利用爱丁堡大学的研究成果，该研究采用了一种独特的方法来提高风力涡轮机的效率。通过先进的建模和新型机器学习技术的应用，算法最大限度地减少了系统中的端到端损失。这项技术正在申请专利，是爱丁堡大学7年研究的成果。该算法的成功已经在小型风力涡轮机上得到了实验证明，并发现能量捕获增加了6%。根据具体的涡轮机和位置，增加6%的能量捕获可以使运营商的净利润增加50%-100%。如果只有一半的英国涡轮机达到6%的增益，就会产生额外的3000吉瓦时的发电量，并在一年内节省130万吨碳。这相当于从街道上减少了大约29万辆汽油轿车。然而，这6%的增益仅在大约100千瓦的中等规模风力中得到证实。预计较大的涡轮机将从更好的控制位置开始，这将使我们能够获得约3%的收益。REOS目前正在准备一个试点项目，在兆瓦级西门子2.3-92涡轮机上验证该技术，该涡轮机是英国陆上舰队的主力。现在，通过这个新项目，REOS将开发并整合新的机器学习技术到一个单一的平台中，该平台将提供风电资产的端到端财务优化，并真正全面地了解整个风电系统。该项目将开发和整合：*连续的每台涡轮机设置优化*先进的假警报检测以增加服务时间*先进的基于人工智能的风电场控制这将在风能资产的控制和性能方面创造一个逐步变化，目的是在大型现代风电场中保持6%的能源输出增长。这将有助于创造可持续创新，并帮助实现向净零排放的过渡。