Predictive artificial intelligence (AI) is on everybody’s lips right now. And for good reason. In the past couple of years, we have seen the emergence of many new and exciting applications for predictive AI within the energy industry to better maintain and optimise energy assets. For example, we see the use of predictive AI to inform when an asset is at a higher risk of sustaining damage and in need of preventative maintenance. We also see it combined with weather and traffic data to dispatch engineers to a site optimally. This is proving to be vital to increasing the reliability of the entire energy system. However, whilst advances in the technology have been nothing short of rapid, the challenge is to ensure that the ‘right’ data is being supplied to make them effective.
The challenges of net zero
Whilst the ongoing transition towards net-zero has to be applauded, it does disrupt both the supply side and the demand side of the energy ecosystem. Electric vehicles (EVs), residential solar, and electric heating are changing demand patterns past anything we have seen before. At the same time, an increase in renewables on the grid causes greater fluctuations in supply capacity. After all, a wind farm without wind, and a solar farm without sun is not particularly useful.
Luckily, predictive AI is excellent at learning these new patterns and deploying models into use rapidly to support demand flexibility. Yet, matching demand to available supply is the inverse of the traditional energy system.
This is key. By better predicting when the energy system will experience an imbalance in supply and demand means that the charging of EVs, for example, can be scheduled better to ensure the balancing of the grid. The reward is cheaper electricity for all. Additionally, if the charging can coincide with when there is a renewable energy supply, then the CO2 associated with that demand can also be reduced so it is a win-win.
Negating energy imbalance
Supply and demand energy imbalance is what the energy sector also seeks to avoid as it can lead to blackouts and crippling payouts. The ability to accurately forecast, therefore, is imperative to being able to negate such an imbalance happening.
Extreme weather not only impacts supply and demand profiles but can damage power lines and prevent power plants from operating properly. Thankfully, there are already certain innovative projects, such as one by Scottish Power, that are aiming to better predict when extreme weather events could lead to power outages and where these are most likely to occur.
The need for localised predictability
Balancing the energy system has always relied on being able to accurately predict customer behaviour. But this was always at the aggregate level when suppliers could turn up and down energy supply at will. Now, though, there is a greater need for localised predictability as distribution grids become more active with two-way power flows caused by distributed energy resources.
Thankfully, with predictive AI, it is now not only possible to learn customer demand patterns at the individual consumer level but even at the appliance level. Although not widely utilised yet, in the future predictive AI will be used increasingly to support demand side flexibility. Particularly with things like electric heating and EVs – often the largest loads in a house or building. If a building has an energy storage system, that too is more likely to come today with optimisation algorithms informed by predictive AI that can learn usage patterns to schedule battery import and export.
Predictive AI will play a major role in the future
According to a recent GlobalData report, predictive AI is already driving measurable improvements in renewable energy forecasting, grid operations and optimisation, the coordination of distributed energy assets, plus with demand-side management within the energy industry. There seems no doubt that the technology will play a major role in enhancing asset optimisation and customer segmentation in the years to come.
Predictive AI is changing the energy sector for the better, whether it be detecting and repairing faults, better predicting weather patterns, or providing more accurate usage monitoring. Whilst the future is exciting, it is still in the emerging technologies phase, though, so there will be some challenges often seen when scaling up that will have to be overcome.
To truly become successful, there will need to be more rigorous governance procedures added to ensure the quality of data used to train the new predictive models are up to scratch. It will be important to confirm the integrity of all training data through detailed logging, auditing trails, verification frameworks, and oversight procedures. And then to continuously evaluate the datasets for emerging issues. In my view, that is where a lot of digitalisation of the energy sector will focus over the coming years.
