Macro Forum: The Street View
Actis in practice: Big data driving performance
Atlas - Quilapilun Solar Park
The Atlas Virtual Manager (AVM) is a big data tool, developed by Atlas Renewable Energy (Atlas), to track Solar PV project’s performance. Atlas is a 100% Actis- owned renewable energy platform, targeting markets across Latin America, and focused primarily on solar-powered generation.
Using plant data obtained from the distributed control system, the AVM allows for real time monitoring of equipment operating conditions, at the different plant levels - strings, combiner boxes, inverters, etc. Thus, the AVM can detect issues from a single tracker not moving as expected, to an entire area of the plant affected by abnormal soiling. Having this level of granularity, allows maintenance teams and operators to quickly identify and triage problems and schedule the best course of actions based on the expected impact to generation.
Developed in-house by Atlas’ engineers, the AVM incorporates performance algorithms that use machine learning and artificial intelligence. These algorithms determine the expected performance of the equipment, considering ongoing environmental conditions. If such performance is below estimations, the AVM triggers an alert informing the site crew of the location, description of the issue and most likely cause of the underperformance.
AVM is constantly being improved with the introduction of new prediction algorithms. For example, by using information obtained from inverters, the AVM can currently predict when a DC cable will fail several days in advance. Given that the AVM can issue alarms to signal these predictions, plant operators are able to optimise maintenance plans and avoid the associated production impact or catastrophic equipment failures. The ability to predict events also helps optimise the spare parts inventory and budgeting of future O&M activities.
As an extension of the AVM, the Atlas team is currently developing a series of tools that will allow an automatic forecast of required labour and spare parts and the integration of these into the enterprise resource planning system (ERP). These tools will help to automatically schedule work orders and provide plant personnel with more time to focus on key issues, such as improving long-term planning.
At Atlas, we decided to combine AVM with other state-of-the-art maintenance and diagnostic tools, such as a drone survey with infrared cameras, IV string tests (current-voltage) or robotic panel cleaning to achieve the highest level of performance and generation efficiency in our plants. As a result of these strategies, Atlas was able to achieve a stellar 99.6% availability across its portfolio during 2020.