Offshore Component Condition Monitoring via Machine Learning Enabled Smart Sensors and Low-bandwidth Data Transmission

海上风能的运营和维护目前基于海事部门产生的旧模式，定期进行例行的定期维护。行业洞察表明，基于对海上风电资产中各个部件的准确状况监测，过渡到预测性维护框架的愿望很强烈。这一需求的一个关键原因是离岸维护要求的高昂成本。目前，海上风电场总支出的20%-30%是基于运营和维护，随着海上风电场数量的增加和变得更加偏远，资源(人力专业知识、闭路电视和SOV)将变得更加有限，因此成本可能会进一步增加。该领域的最新技术提供了用于监测机械系统的各个方面的传感器，如振动和声发射。这一领域的一个关键挑战是将数据传回控制中心所需的通信。FLICQ开发了基于工业物联网技术的智能传感器技术，不仅可以从资产中收集数据，还可以处理和分析传感器数据，以提取有意义的信息，并最大限度地减少需要传输回云或传统系统的数据量。为了执行此数据分析，FLICQ技术使用机器学习算法对此数据进行预处理以进行传输。通过准确建模资产状况的变化，预测模型可用于根据风险、预期停机时间和预计收入损失来规划最佳干预时间。该项目旨在通过Ore Catapult旗下的莱文茅斯海上示范涡轮机的实际应用和产品开发，展示FLICQ智能系统支持海上风电行业预测性维护策略开发的潜力。虽然FLICQ传感器技术已经应用于一系列其他领域，但该技术尚未在具有挑战性的海上风力环境中实施。由FLICQ提出的这项创新技术使用了收集数据的直观机器学习算法。这意味着传输到陆上控制中心所需的数据连接是低带宽和低数据使用率。这使得传感器可以保持低功率，并且在任何组件上都有很长的使用寿命。FLICQ在其他市场取得的成功代表着一个机会，可以使用以使能技术为动力的颠覆性智能传感器来颠覆海上风能的维护策略。

Operations and Maintenance in Offshore Wind is based currently on old models generated in the maritime sector, where routine, scheduled maintenance is carried out periodically. Industry insight has indicated a strong desire to transition to a predictive maintenance framework, based on accurate condition monitoring of individual components within an offshore wind asset. A key reason for this need is the prohibitive cost that offshore maintenance commands. Currently 20-30% of the overall expenditure for an offshore wind farm is based on Operations and Maintenance, and as the number of offshore wind farms increases and becomes even more remote, the resources (human expertise, CTV's and SOV's) will become more limited, and costs could therefore increase further.The current state of the art in the field offer sensors for monitoring aspects of mechanical systems such as vibration and acoustic emissions. A key challenge in this area is the communication required for transfer of data back to the control centre. FLICQ have developed Smart Sensor technology based on Industrial Internet of Things technology that not only collects data from the assets, but can also process and analyse sensor data to extract the meaningful information and minimise the amount of data required to be transferred back onshore to the cloud or to legacy systems. To carry out this data analysis, the FLICQ technology uses machine learning algorithms to pre-process this data for transmission. By accurately modelling how asset condition is changing, predictive models can be used to plan the best intervention time based on risk, anticipated downtime, and projected revenue loss.This project aims to demonstrate the potential of FLICQ smart systems to support the development of predictive maintenance strategies in the offshore wind sector through a practical application and product development at Levenmouth offshore demonstration turbine, owned by ORE Catapult. While FLICQ sensor technologies have been employed in a range of other sectors, the technology is yet to be implemented in the challenging offshore wind environment. The innovative technology proposed by FLICQ uses intuitive machine learning algorithms where the data is collected. This means that the data connection required for transmission to onshore control centres is low-bandwidth and low data usage. This allows the sensor to remain low-powered and have a long life in position on any component. The success FLICQ have experienced in other markets represents an opportunity to disrupt maintenance strategies in offshore wind using disruptive, Smart Sensors, powered by enabling technology.