The Unmanned Power Plant—from Science Fiction to Industrial Reality
Sep 06, 2017
Azeez Mohammed

Driverless carsartificial intelligence, even unmanned ships—in every corner of life it would seem that what was once science fiction is rapidly becoming reality. The same is true in solar power generation—while the idea of unmanned solar plants may appear fanciful, the digitalization of the sector means that it isn’t simply a concept, but something that is soon to take over solar plants across the globe.

To get to a point where this becomes more widespread, we have to look at the ability of solar plants to maximize their productivity and cut back on waste and unnecessary costs. Whether we have a sunny day or a cloudy day is not something that we can control and so, being able to maximize that resource when it’s available is critical. That means making sure that the assets on fields generating solar power are working at peak efficiency and never suffer unscheduled downtime. This is precisely where the opportunity for digital lies.

The evolution of maintenance

Maintenance may not seem like the most compelling aspect of solar power generation—fundamental, certainly, but perhaps not exciting. Yet, it’s in the maintenance of existing assets that operators can reap significant rewards, and it is in this space that we are seeing accelerated innovation.

Traditionally, maintenance is run on a calendar basis, with operations and maintenance teams following manufacturers’ recommendations for checks and servicing. These checks fit into a routine rather than relating to the actual workload or performance of the equipment. The scheduling of maintenance itself is often done in a third-party system, completely separate from the tools that actually track asset performance. Crucially, this approach does not lend itself to being measured for effectiveness—each check is made because the calendar says so, not because the machinery requires it—so performance measures are rarely in place.

This inefficient approach is ripe for disruption. A new approach—one that collates and analyzes data and uses those insights to inform clear decision-making is needed. In this instance, it’s about using digital technology to optimize maintenance.

The digital disruption of solar plant maintenance

Before we can realize a future in which totally unmanned plants become an everyday reality, there are a few building blocks we need to put in place that completely disrupt the current status quo for plant management.

The first is about enabling predictive maintenance. We need to move from our calendar-based system to one that assesses the plant as it is, in real-time, and anticipates maintenance needs in such a way that critical human intervention is never required.

This comes in two parts. First, real-time access to sensor data across a plant estate. This data can immediately flag obviously anomalous against its normal operation readings—at GE, we call this digital replica of physical assets the Digital Twin—and prioritize resources for maintenance or replacement if and before they start to fail. Second, critically, by establishing a baseline virtual model for the plant and simulating what it should be doing under optimal working conditions, it becomes possible to fine tune parameters and enable assets to run at optimum levels.

The net output here? Not only do you reduce the unplanned downtime, but you can also boost efficiency by making optimization tweaks based on the insights drawn from Digital Twin data. It also can reduce maintenance costs by only replacing components that are showing the tell-tale signs of wear.

The predictive maintenance approach leads us to another critical area of an unmanned plant future, which lies around the concept of service resources management optimization. What we mean by this is that there is a need to have integrated tools that rely on accurate data to deliver predictive alerts therefore, accelerating service by ensuring as-needed dispatch of parts, engineers and so on.

Keeping a large stock of spare parts, plus a substantial engineering base, drives costs and complexity of plant management. Being able to tap into these at, or just ahead of, demand delivers savings and efficiency while improving overall services. If you can harness the data and automate these processes, as well as the process of choosing the best field service engineer resources—from skills, cost and location perspectives to support service operations—then a fully unmanned plant becomes a solid reality rather than future fantasy.

Delivering tomorrow, today

Being able to autonomously optimize maintenance at their solar plants saves operators time and money. It ensures that only necessary maintenance is conducted, rather than the less-efficient, calendar-driven model, which lacks real performance measurement and predictive ability.

GE’s own Asset Performance Management (APM) provides this level of insight. Powered by Predix, it gives operators access to critical data in order to build a picture of operational effectiveness or lack thereof. The cloud-based nature of Predix also means that data can be stored and shared easily, no matter where the plant is.

What’s more, the likes of APM are just the start of how digital could be deployed to optimize maintenance and deliver an unmanned power plant. For instance, soiling—a buildup of debris such as snow, dirt, dust, leaves, pollen and bird droppings on photovoltaic (PV) panels—continues to be a major contributor to maintenance costs. Deploying digital tools, which monitor array soiling build up, would allow for automatic maintenance scheduling once performance dropped below a certain threshold.

Another potential digital innovation is the remote monitoring of PV panel health. In large plants, being able to check on the physical appearance of panels is not practical. By deploying remote monitoring, such as drones linked to centralized systems conducting thermographic surveys, operators would be able to process imagery and other data, which could help pinpoint areas of concern for maintenance after sunset.

Maintenance is resource intensive. Being able to deploy digital solutions gives operators a way of accelerating the evolution of maintenance and to cut down on costs and time spent fixing problems. This ultimately leads to a more-efficient, cost-effective plant, which optimizes power generation. Digital solutions such as APM are a critical first step in the journey towards the fully autonomous solar power plant. Plant developers that invest now will not only see immediate improvements, but will also be future proofing themselves for long-term performance gains.



Azeez Mohammed

Azeez is the President & CEO of GE’ Power Conversion business that drives the electric and digital transformation of the world’s energy infrastructure across multiple industries. Before taking his role in Power Conversion, Azeez was the President & CEO, Middle East and Africa for Power Services, a $15 billion organization, offering customers total plant capabilities, local resources and expertise to help them be more successful. Azeez started his career with GE Research & Development in 1998, where he worked on advanced technologies for application in the Energy and Healthcare industries.