Sift AML (Accessible Machine Learning): Rapid and Robust Automated Analysis for Wind Farms

Sift AML（无障碍机器学习）：风电场的快速、稳健的自动化分析

• 批准号: 68658
• 金额: $ 21.45万
• 依托单位: BITBLOOM LTD
• 依托单位国家: 英国
• 项目类别: Study
• 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=68658
• 关键词: Sift; AML; Accessible; Machine; Learning

As the world turns away from fossil fuels, wind power will become an increasingly significant contributor to the global energy supply system in the coming decades. By harnessing the power of Machine Learning (ML) and Artificial Intelligence (AI), Sift AML transforms the mountains of data produced by wind farms to actionable knowledge that helps to increase their operating efficiency, reliability and maintenance. By doing so, they operate at a higher capacity for longer, saving us all money on our electricity bills.Wind farms create vast quantities of data. A modern offshore wind farm may comprise a hundred or more wind turbines. Each turbine produces hundreds of data signals that contain crucial information about the state of each sub-system and the condition of the thousands of individual components. There are also several support systems that also produce data, such as substations, array cables, service vessels and meteorological masts.In total, a modern offshore wind farm may generate in the order of 10 petabytes of 1Hz data and 15TB of 10-minute statistics annually. Hidden within these data is critical information about component wear and pending failures. When utilised correctly, these data can be transformed into knowledge that can help owners to improve efficiency and reliability, optimise maintenance regimes and reduce operating costs.Most wind farms are connected to central data warehousing systems with web portals where the owners and operators can analyse the information for monitoring purposes. Generally, however, the capabilities of these systems to provide in-depth technical analysis and prognostics are very limited.Sift AML is a simple to deploy, in-depth analytics and prognostics system that enables energy producers of any size to build advanced analytics on top of their existing data warehousing solutions. Using state-of-the-art developments in ML and AI, Sift AML provides access to these revolutionary technologies to everyone, not just data scientists and software programmers.By using Sift AML wind turbine operators can gain rapid access to advanced analytics that turn data produced by plant machinery into actionable insights, including guidance on pending component failures and opportunities for increasing power production. In power production, even small increases in efficiency can generate significant returns at scale for operators, lowering the cost of electricity and protecting the planet by accelerating the shift toward carbon-neutral energy production across the world.

随着世界逐渐远离化石燃料，风力发电将在未来几十年内成为全球能源供应系统的重要贡献者。通过利用机器学习（ML）和人工智能（AI）的力量，Sift AML将风电场产生的大量数据转化为可操作的知识，有助于提高其运营效率，可靠性和维护性。通过这样做，他们在更高的容量运行更长的时间，节省我们所有的钱在我们的电费。风力发电场产生大量的数据。现代海上风电场可以包括一百个或更多个风力涡轮机。每个涡轮机产生数百个数据信号，其中包含有关每个子系统状态和数千个单独部件状况的关键信息。此外，还有几个支持系统也会产生数据，如变电站、阵列电缆、服务船和气象桅杆。一个现代海上风电场每年可产生约10 PB的1Hz数据和15 TB的10分钟统计数据。隐藏在这些数据中的是关于部件磨损和待处理故障的关键信息。如果正确使用，这些数据可以转化为知识，帮助业主提高效率和可靠性，优化维护制度并降低运营成本。大多数风电场都连接到带有门户网站的中央数据仓库系统，业主和运营商可以分析信息以进行监控。然而，一般来说，这些系统提供深入技术分析和预测的能力非常有限。Sift AML是一种易于部署的深入分析和预测系统，使任何规模的能源生产商都能够在其基础上构建高级分析现有的数据仓库解决方案。Sift AML利用机器学习和人工智能的最新发展，让所有人都能使用这些革命性的技术，而不仅仅是数据科学家和软件程序员。通过使用Sift AML，风力涡轮机运营商可以快速访问高级分析，将工厂机械产生的数据转化为可操作的见解，包括对未决组件故障的指导和提高发电量的机会。在电力生产中，即使是效率的小幅提高也可以为运营商带来巨大的规模回报，降低电力成本，并通过加速全球向碳中和能源生产的转变来保护地球。