Hydro Tasmania to tap into predictive analytics for asset maintenance, operations
- 09 March, 2015 16:47
Hydro Tasmania is looking to further improve its asset maintenance and plant operations by using predictive analytics.
Ensuring assets are maintained properly and last in the long term is crucial for Hydro Tasmania. It has 55 main dams, 30 hydropower stations and some of the country’s largest wind farms.
Hydro Tasmania is a renewable energy business so weather forecasting is key to being able to operate plants efficiently and ensure there’s enough supply that meets demand.
Field workers constantly collect data on the condition and performance of assets based on their measurements and observations. This can include unstructured data such as photographs showing an abnormality in a dam, for example. SCADA (supervisory control and data acquisition) systems also capture data in real time.
“It’s being able to better predict failure times to find a way to increase the lifespan of an asset and also find a better way to ensure the spend we will do to maintain an asset is optimised, therefore potentially [making an asset] last over a longer period of time,” said Hydro Tasmania’s analytics and information manager, Scott Delaney.
“At a certain point on our radar is the need or the desire to move from predictive analytics to prescriptive analytics where we say, ‘Given this event is likely to occur in these conditions, what is the best course of action to actually now take to ensure that the life of the asset is prolonged and the spend on an asset is optimised’.”
David Shields, maintenance manager at Hydro Tasmania, said the business is using systems that are a “few generations old” to operate its plants. He is looking into SAP predictive analysis software and others for optimising operations by being able forecast weather and market conditions much faster and more accurately.
“Our inputs vary, so it’s about being able to trying to match our [energy] generation against that and get the best output for whatever rain might be falling and where the price in the market is. We’re just trying to get the best value we can from our resources that we have available to us,” he said.
Hydro Tasmania also wants to use predictive analytics to better tailor energy packages to specific customer types. The aim is to understand a class of customers' likely behaviour and what would likely drive or motivate them in their decisions.
“We want to provide the customer with information for them to make intelligent choices, which is the really important thing for us going forward. Them being able to pick the right plan to be on or whatever it might be, that’s an important differentiator for us moving forward,” said Shields.
Mobility keeping field workers in the field
For Hydro Tasmania, it’s important field workers who work directly with the assets stay out in the field and less time in the office as they are mostly the ones making sure the business runs smoothly.
This month last year, the Hydro Tasmania equipped 110 of its field workers with iPad Minis with SAP Work Manager to enter inspection forms electronically at the source instead of having to travel back to the office and manually key data into a database.
Shields said workers doing heavy-duty routine work, where they take down quite a lot of readings, are saving at least two hours a day. That time saved is then spent on paying more attention to detail or inspecting something further.
“If you’ve got an extra five minutes you can actually go and cut off a sapling or something of that nature which might seem really small. But if you don’t have the time to do those things that are growing on the back of a dam, it turns into a tree and it becomes a major exercise to remove it. So it enables us to save a lot of money into the future,” he said.
The data captured is also validated on site, which helps prevent the likelihood of human error or mistakes from occurring.
“We used to have a lot of false positives, I suppose would be the right way to put it, where somebody would [accidentally] put the wrong reading in, which would result in a whole heap of people trying to chase each other to find out it wasn’t actually a problem in the first place,” Shields said.
“We expect it to pay back within two years and it feels like that’s about right,” he added.