Monitoring at sea
Structural condition monitoring will be essential to keep up with offshore wind developments, as Kevin Magee and Nick Stringer explain.
For those living around the coastline of the UK, the growth of offshore wind power and energy can hardly have escaped their notice with many sites deployed within a short distance of land and a constant stream of installation and support vessels sailing to-and-fro. Among these is London Array, the worlds largest offshore wind farm, with a 630MW capacity. With global offshore wind expenditure forecast to reach £210bln over the next 10 years, monitoring systems will provide operators with the ability to plan when inspections and repairs are necessary, lowering the frequency of offshore failures and costs.
The latest expansion of wind farms will bring even bigger arrays and the East Anglia Array, located around 30 miles off the east coast, in the North Sea, UK, to be developed in partnership by ScottishPower Renewables and Vattenfall has a potential output of 7.2GW, with offshore construction to begin this year and finalised in 2020. This will be more than the 4GW output capability of Drax, which is currently the largest power station in the UK. Individual offshore wind turbines are now capable of generating up to 10MW and the technology has seen step changes in size with turbine blades in the order of 180m in diameter and towers in excess of 220m high. The wingspan of an Airbus A380, the largest passenger jet, is less than the blade length for one of these offshore giants at 79.45m.
This technology does not come cheap and offshore wind farms are expensive and valuable assets. The latest developments cost billions of pounds for construction and commissioning. Of the total cost, approximately 70–80% of the overall cost of a wind farm is capital expenditure (CAPEX), with the other 20–30% being operational expenditure (OPEX). Operators need to generate electricity that can be sold to repay the initial investment and ongoing operating costs. The failure or compromised operation of these assets can have serious financial consequences and even a small reduction in downtime or an extension to expected lifetime can have a rapid payback. Long-term success of these wind farms will be dependent on their cost reduction-monitoring framework.
One way to reduce lost time is to add intelligence to asset management by using condition monitoring. The use of condition monitoring systems is not a new idea and most offshore wind turbines are bristling with sensors, predominately fitted on the mechanical components. Unfortunately, many systems only lead to the generation of vast amounts of data with little or no meaning. Extracting real value from condition monitoring equipment requires a systems-level approach to give meaning to the measurement. A collaboration between UK-based companies, Proeon Systems and Aquaterra Energy is implementing this approach by combining expertise in advanced control and monitoring with those in structural mechanics and analysis to collect the appropriate data and provide relevant interpretation. As such, this will enable operators with accurate, informed, and real-time condition analysis data to enhance maintenance, commissioning and decommissioning programmes.
Providing this information allows operators to build a comprehensive picture of the condition of individual assets and their life expectancy. Careful choice of the data to be measured and its subsequent processing can be used to reduce offshore failures by assessing the current condition and predicting the future states of installed equipment. Alarms can be monitored and acted upon appropriately, on-site inspections can be minimised, and maintenance can be planned for maximum effectiveness. In addition, monitoring can greatly reduce the need, risk and cost associated with inspection crews travelling to and from offshore wind farms. This technology has already been deployed on structures in the North Sea and has enabled the operator to gain greater understanding of the structural performance in order to manage their operations and risk accordingly. It has also enabled the operator to have greater operating uptime than expected, as the monitoring proved the conservatism of some of the predictive design methods used.
Acquiring and using data
A typical installation for a system of sensors has all data acquired, processed and transmitted to a dedicated shore side storage facility, either through the operator’s existing network and supervisory control and data acquisition or through dedicated systems. Data can be made available to other operational sites if required. Instrumentation is typically connected to a Data Acquisition Unit (DAQ) located on each wind turbine, providing local storage and ability to forward the data in cases of communications failure. An ethernet data connection supports high-speed data acquisition.
The number, type and position of sensors to be installed are designed for each application, depending on the details of the structure and wind turbine equipment. Typically, the data to be collected includes environmental, structural, mechanical and electrical systems. The sensors for environmental data acquisition includes moisture and humidity, temperature, wind speed direction, wave height and direction, water current speed and direction and levels of scour. Typical structural sensors include strain gauges and accelerometers. Sensors for the mechanical systems measure vibration of equipment and particulates in lubricants, with electrical systems also monitored. Additional data can be added including metrology and date-time information from satellite communications.
The key to extracting real value from the monitoring system is to integrate the acquired data with mathematical models. For example, consider the structures supporting the wind turbine. Potential problems that might occur are grout slippage at tower-foundation interfaces, degradation of the foundation, and change of support to the foundations from scour. Sensors for monitoring the structure can acquire data on strains within the tower (strain gauges) at multiple locations. The inertial sensors can gather data on the acceleration and position of the tower as it moves. The real value of monitoring this information comes when they are compared to the predicted behaviour of the structure using methods such as ambient vibration monitoring (AMV).
AMV uses the measured response of the structural system to environmental or other ambient loads. The natural frequency and damping can be calculated from acceleration data and the current condition of the structure determined by comparison to models. If there are sufficient sensors then the location of potential problems can be determined. The appropriate position, number and type of sensor are determined in advance by structural modelling. Specialist knowledge of structural mechanics is required to design and build the appropriate mathematical models.
Proeon and Aquaterra have developed a support system for such inertial sensing that combines acceleration with angular rate and inclination to be used in the North Sea. The offshore wind energy monitoring system will improve inspection and repair planning, and lower the frequency of offshore failures. The angular rate and the inclination data can be used to elaborate the picture and together the companies have developed novel algorithms for determining structural behaviour from these additional data. Determining the state of any sub-system only solves part of the puzzle and the next question is which avenue of investigation to follow. There are three avenues to explore using the knowledge gained – alarms, root cause analysis (RCA), and predictions of future condition.
Solving the puzzle
If the alarm is used, the measured behaviour of the system deviates from the expected behaviour and then an alarm is triggered. The alarm may require a quick response or it might be flagged for long-term review. An alarm for the electrical or mechanical systems may require on-site inspection and intervention or for equipment to be temporarily taken out of commission until intervention is possible. Conversely, an alarm for the structure might indicate that the system behaviour is degrading from the expected behaviour but the change might be occurring sufficiently slowly that no immediate intervention is required.
The use of real-time data with the resulting alarms and events is a key window into the state of the structure, but is only one part of a comprehensive condition monitoring system. It is likely that alarms and events created by the disparate systems for a wind farm may have causal links that can result in either alarm floods or what can appear as unrelated alarms and events. Such alarms and event floods can be confusing to operators who may be dealing with many tens of thousands of points of data.
One solution to overcome this is to use root cause analyse (RCA) techniques to drill down into data derived either from the condition monitoring system of other systems on the wind farm system such as the SCADA or meteorological systems. The RCA identifies a single source of failure known as the root cause and generate an appropriate actionable alarm. Real-time RCA techniques may include a variety of methods, which are employed to solve various problems and conditions. In all cases, it will interpret a set of symptoms and events and pin
point the source that is causing those occurrences.
While conceptually simple, this can be surprisingly difficult to achieve in the real world with real-time feeds of data from a number of sources. The RCA system is required to understand the relationship between information within the infrastructure together with the systems and users that rely on it. By identifying equipment that has failed with significant reductions in fault finding time and resources is crucial to reduce costs of offshore wind farms.
Real added value
Output information from the real-time analysis can be used as an input to further studies to identify long-term integrity issues. An alarm and its root cause can indicate a particular problem but investigation over time can reveal valuable insights and predict future states. Gradual changes over time can be investigated to predict future effects before an alarm occurs. For example, combining structural data with the metrological systems data might show changes in structural behaviour caused by changes to the seabed. Further examination of data over time and modelling of the evolution of the structural system might reveal that the structure will continue to perform satisfactorily for the design lifetime of the offshore towers. It can also reveal that the structure will not perform satisfactorily after a certain time.
Monitoring the evolution of system variables helps to predict future states of the systems. Intervention can be planned to maximise effectiveness and minimise cost, such as during a campaign in the most benign periods of the year. The structural system used as an example is one of several systems in a wind farm. The mechanical and electrical systems can and should also be monitored and similar investigations performed and appropriate interventions planned.
Wind farms are changing and evolving technologically to meet the demands of capacity, location and reductions in electricity generation costs. The latest developments are substantial in size and are expensive assets that represent considerable investments in time, effort and money. Reducing operating and maintenance cost, as well as managing the risk of disruption to supply, is a challenge for the operators of these assets. Real value can be added and costs reduced by adopting real-time monitoring and data acquisition with initial system modelling and subsequent data processing and interpretation. The UK is a market leader in offshore wind energy and using collaborative, UK-manufactured technology to implement monitoring solutions from the outset will enable wind farm operators to safely and cost-effectively improve the integrity and operational efficiency of their assets.
Kevin Magee is Managing Director of Proeon Systems, a chartered electrical engineer and CFSP with more than 25 years' experience in detailed design, project engineering, project management, commissioning and engineering management.
Nick Stringer is General Manager, Technical at Aquaterra Energy, a chartered mechanical engineer with 15 years' experience in analysis, who is responsible for the technical engineering teams including riser analysis, structural and design engineering.