Automated Weather and Environmental Risk Alerts for Operators of Physical Energy Assets
The energy industry deploys substantial and expensive physical assets such as oil rigs, power stations, pylons, pipes, pumping stations, liquification plants, and substations. These are essential to ensure the supply of energy to end-consumers. Traditionally, these assets have been designed to withstand prevailing weather and environmental conditions based on long-run historic data, but climate change modelling suggests an increasing number of extreme events that will fall outside asset tolerance parameters, threatening serious and costly unscheduled outages. Asset operators are becoming increasingly vulnerable to climate change in terms of extreme weather events, and more gradual, insidious factors such as ground erosion, flooding, encroaching vegetation, and even human habitation. This fact is not lost on their financiers and shareholders, who are applying increasing pressure on physical asset operators to improve their monitoring of environmental risk factors. Indeed, most new physical asset projects cannot secure funding without demonstrating adequate environmental risk monitoring.
Many physical asset operators currently rely on a largely manual process to handle environmental risk notifications. This is often managed as a team effort between control room and supporting engineering and environmental staff. Weather forecasts arrive, and are interpreted in terms of local understanding the geographical region, and specific knowledge of the assets impacted. The team then rates the threat posed in the weather forecast, and plans its response accordingly, aiming to pre-empt an otherwise potentially disastrous unscheduled outage.
Weather and environmental risk alerting based upon a manual process is a time-consuming and resource-intensive approach. It can be easily overwhelmed by a large number of weather warnings covering a broad geographical area over a short time period. Furthermore, an undue focus on short-term, extreme weather events may lead to the failure to recognise insidious environmental risks that may take months, even years, to become manifest.
For physical network assets, the emphasis needs to shift to an automated approach that issued risk notifications when a forecast weather event threatened the performance of a network asset to the point where there was the risk of an unscheduled outage, applying user-defined criteria.
Each network asset would have its own parameter-driven AI-model, such that the user defines which weather factor(s) the asset is sensitive to. The model would collect the weather data, apply adjustments based on proximity to the neighbouring Weather Stations, and update the risk level for that asset. If the risk level was above a certain threshold, a notification would be circulated to users and the Risk-Rank-Table would get updated.
The models incorporate outage risks attributable to both ‘instant’ events, such as storm-force winds or heavy snow, and to ‘cumulative’ events, such as ground saturation caused by prolonged rainfall. On setup of the appropriate data feeds, longer-term or ‘insidious’ environmental factors can be added to the models, so users are given advance notice of an approach to criticality in factors like erosion, water-table, and vegetation.
Where required, the risk-level would be linked to a database list of commensurate remedial actions, so once a threshold of criticality was passed, the system could recommend a course of mitigation based on past experience (general or site-specific), learning from that experience. Being based on AI, the models are continuously retrained based on new experience, and/or the identification of additional risk factors.
Matrica’s Nominator software platform already offers a comprehensive range of AI and statistical forecasting models for energy physical assets that incorporate weather factors, alerting users to extreme weather events that adversely impact their performance. This range is being extended to cover physical asset operations across different sectors, including water and telecoms, and incorporating additional environmental factors.
Nominator also provides a full range of automated (and manual) data capture, query, display and export options via its UI. Users can schedule the display of the current risk level of their assets as a heat-map, RAG, ‘worry-list’ report, or Risk-Rank Table.
If you are physical asset operator concerned by the need to more effectively monitor weather and environmental risk factors, or would like to exchange views , please contact Dr Mark Earthey or Prudence Mauthoor.