European countries are committed to increase their electricity generation from renewable resources to meet the European targets for 2030. Several strategies to meet this goal focus on increasing the installed renewable energy capacity by deploying new plants and repowering existing ones.
A complementary path, and probably the most effective from an ecological and economical point of view, consists on improving the actual performances and reducing the operational costs of the existing renewable power plants. One of the most reliable ways to succeed in this goal is the application of artificial intelligence to convert the immense amount of data that these plants generate into information that can be used to make the right operational decisions at the right moment.
The possibilities to optimize performance, increase availability and reduce operational cost through the application of data mining algorithms is widely demonstrated, provided that high quality, high granularity, consistent and robust data is processed within the framework of a proper system model.
This might look intimidating from a traditional approach to Asset Management, but the technology required to perform a professional and structured improvement to the performances of Solar PV and Wind assets is available, tested and validated.
The impact of this professionalized way of managing renewable portfolios goes beyond the immediate effect that the capacity of plants is fully exploited but it also allows that the overall energy industry profits increase (via a decrease in LCOE) and that operational risks are mitigated better and earlier. This will ultimately attract more investments on both renewable projects and grid infrastructure, generating a virtuous circle that will trigger new installed capacity deployment and higher penetration of renewables in the energy mix.
The power of data
Data mining is the process of retrieving vast amounts of data from one or more sources and relating them together with the aim of understanding ongoing anomalies and predicting devices future behaviours. Big data analytics can bring added value at any stage of asset management: analysis from observation of collected information to fault detection, fault diagnosis and finally optimisation through recommendations issued from an advanced performance monitoring system. Today different approaches are proposed. Whereas classic Artificial Intelligence (AI) proposes advanced diagnostic through knowledge-based models, unsupervised and supervised learning methods offer different possibilities (e.g. neural networks) using statistical approaches.
Thanks to current data mining techniques, asset operators can easily take decisions on the most effective way of performing their daily operations and maintenance activities, improving the performances of the portfolio and anticipating failures on the devices composing these complex systems. Monitoring and performance improvements platforms such as 3E’s SynaptiQ are capable today to combine an independent monitoring and data collection to the most reliable performance analysis algorithms in the PV industry.
In the following figure, the team of SynaptiQ ranked the PR of 640 utility and commercial scale plants and observed that – while in general plants are assumed to perform as expected, some of them are experiencing serious underperformance. In this analysis, the actual performances are compared to the expected ones established using the state of the art in terms of algorithms from the PV market and relying on on-site measurement.
These results are even more significative when considering the analysis on the reliability of the irradiation measurement conducted by iLab, 3E’s R&D department, comparing over 80 physical irradiation sensors with high-granularity and high-quality solar satellite irradiation data provided by 3E’s Data Services. As shown in the figure below, 50% of the sensors are providing an underestimate measurement for solar resources, possibly due to lack of calibration, incorrect alignment or lack of cleaning.
Reliable irradiation measurements from site are crucial for performance assessment and optimization. Today, tools such as 3E Solar Sensor Check® are able to deliver an advanced analysis of sensor quality of measurements.
By comparison with 3E’s satellite-based irradiance data, using the most advanced Cloud Physical Properties (CPP) models, the Solar Sensor Check® delivers precise indications on potential calibration issues and determining the impact on the irradiation quantification.
Reducing plant losses & optimizing plant performances
A carpet analysis of data available from the assets, associated with a well-known set of parameters of the site and consistent algorithms, is employed by advanced monitoring tools such as SynaptiQ and the powerful PV Health Scan, the 3E’s automated diagnostics tool. This tool applied by SynaptiQ allow a full losses root cause analysis based on the comparison of expected and actual losses in each conversion step of a system. In this way, performance optimization through targeted recommendations for immediate or mid terms actions can be achieved by a sensitive reduction of device downtime and underperformances. Additionally, a smoother planning of activities and a better device replacement scheduling aim at reducing devices and operational costs.
The following waterfall diagram is an example showing the PR degradation of one plant extracted from the PV Health Scan. Here the operator can easily focus his attention on the main event that create the higher PR losses (in this case the DC current) and then drill down into the data to investigate the problem in further detail, understand since when it’s happening and evaluate possible root causes.
Just by using the existing data this operator improved his knowledge leading to an immediate and effective decision that required nearly no time consumption on the identification, allowing him to focus on the solution.
The invaluable simplicity of SynaptiQ
The global effort to increase the share of renewables in the energy mix will not cease until the climate crisis humanity is going through is halted.
All efforts to this end are valuable. But the simplest are invaluable because they are more effective.
It’s much more effective to improve the actual performance and reduce the operational costs of existing renewable power plants, than to build new ones.
And 3E’s SynaptiQ does exactly that, by applying artificial intelligence to convert the immense amount of data that these plants generate into information that can be used to make the right operational decisions at the right moment.
Who could ask for more?
Martina Pianta (3E) Sales Manager
Vasilis Mertiris (3E) Sales Manager
Nikos Chalkiadakis (REDPro) Analyst