Derwent Hydro began in 1988 with a passion for water power. Growing steadily ever since, many hydro electric generators have been deployed nationwide, but with more generators in remote locations around the UK, it became more and more expensive to monitor their efficiency and detect potential problems before they affected power generation.
- In the know - to know how the company’s generators are performing requires making phone calls or journeys across the UK to the physical control panels at each turbine. When the turbines are not easily reachable, data gets stale, and not having the most up-to-date data can become costly if there. We needed to interface with the turbines’ control panels to extract the data in a meaningful manner.
- Mission control - live data means better decisions, and just as each turbine has its own control panel, the data needs to be presented to the team in an efficient system to allow making better informed decisions.
- Remote access - the number-one bottleneck in day-to-day business is having to collate data in order to enhance production and prevent failure. The team needs the ability to access the mission control data while on the road, removing the need to travel around the country for low priority jobs.
- Added value - deploying a hydroelectric turbine is a valuable asset in itself, but being able to supply customers with their own online data logging system is something that can set Derwent Hydro apart from the competition.
The open web, now with more hydropower.
The web is in our blood, but this kind of project is not your typical web build. We needed a strong plan of action and clear direction to tackle this project successfully. Before we could start, we needed to tackle the problem of data being distributed around the country on difficult-to-access hardware.
To keep costs to a minimum we used the Raspberry Pi micro-computer to interface with each generator around the country. Tiny web servers were used to output raw data over an HTTP stream from electronic sensors on the turbines and in the water. Once we had the bare bones in place, the data was abstracted enough to treat the rest of the project like any other web deployment.
We built an API to abstract the data from the sensors. It isn’t very often when an API endpoint is an actual river.
Little and often deployments.
Once we were back in the comfort zone of web languages, we used our Continuous Integration system to develop the application in an agile way, responding to customer needs as they came in.
We always split our projects into iterations. The first few iterations have highly specific needs, while the future iterations are more abstract and allow the project to grow in the real world without constraining its use. As the software is used out on sites, needs arise that weren’t in the original specification, but adhering to the Agile Manifesto throughout the project allows us to work these needs into future iterations, building a product that solves real issues and adds real value to the business.
Individuals and interactions over processes and tools
Working software over comprehensive documentation
Customer collaboration over contract negotiation
Responding to change over following a planThe Agile Manifesto, Kent Beck
One app for all.
One feature that wasn’t anticipated originally was the usefulness of having the data accessible on mobile devices. While we apply a mobile-first approach to most web projects, an application based heavily around graphs and figures was difficult to apply on a small screen, but the web being what it is allowed us to deploy new visual designs without any noticeable downtime.
The software was packaged as an added value to the turbine installation to the end-client, allowing the site owners themselves to log in from their computers and phones to view their power output and historical data.
One major feature added in a subsequent iteration was the ability to set alarms on particular remote sensors. For example, a site owner or maintainer could be sent an email when a river level became too low or a turbine begins creating a lot less power than it should. With the added insight, not only can operators of the turbines prevent any downtime, they can now actively work to enhance power level output over time.
Now Derwent Hydro have grown nationwide, with sites in South Wales, Yorkshire, Norfolk, Reading and Glasgow to name but a few, the savings in travel alone is immeasurable. Faults with turbines can now be identified remotely, but they can also be prevented through the automatic alarm triggers and analysis of historical data.
What does the future hold for this project? Now we’ve been running this system for a few years, any problems along the way have been noted, actioned and improved. We are discussing possibilities of creating our own affordable off-the-shelf monitoring products for other electricity production services using the Raspberry Pi device.
This project is growing year on year, with more hydro sites and more requirements being added all the time, putting the Raspberry Pi to the test but mainly showing off the value bespoke software can add to your business.