HIGH PROFILE
The AI monitored data is automatically and continuously analysed by our intelligent algorithms which can instantly recognise an anomaly and advise of the appropriate response .
Smart Pump Assistant ( SPA ) is available to users in an app on both Apple iOS and Android
unplanned downtime altogether . Internally we also realised , based on progress with data analytics , that we could create something really significant for our customers . So , I would say it is a combination – we have been on this journey together with them – but we also knew that we had to take the lead from an innovation point of view based on our experience and knowhow . We are the pump manufacturer so it is up to us to
develop the right tools to place at our mining customers ’ disposal . We saw the potential to offer them a new kind of service that would really help them reduce cost . When we saw the first field data coming in almost four years ago now , it really helped bring clarity to our sensor concept and give us confidence to move forward . not a unique thing – the skill is in using the minimum amount of them in the right places to get the data that you need . What really is unique
Q Is the Smart Pump Assistant already out there in the market and what success have you seen so far ? JP : SPA is already well established in the market all around the world , including in Latin America , Africa , Australia , India , US and Europe . We wanted to start SPA ’ s development with field data from as many pumps as possible rather than just relying on one or two prototypes , so that is how we approached it . The training input for the algorithms represents over 150,000 hours of pump operation and they are still ‘ self learning ’ today . We wanted SPA to be globally available and scalable and we have done that . We have also
Q What types of sensors are you using and did you develop your own algorithms to underpin the AI in Smart Pump Assistant ? JP : We use standard sensors , some of the best available , and of course which are designed for very harsh environments . The sensor concept is
Abel HMQ pumps in situ at a minesite is the AI that is able to interpret the data and our algorithms were developed completely in-house – there is no ChatGPT for piston diaphragm pumps !
established a process between ourselves and our customers in how we send over the advisories to them . To simplify this our SPA is also available to them in an app on both Apple iOS and Android . They don ’ t want the data , they want to know what action to take , and using the app is much more convenient for them than just using email . Different app icons are used to show different health areas of the pump – where a wear problem has been identified for example – it will also indicate the estimated remaining run time before failure . The algorithms do the work , at ABEL we don ’ t have to have our own experts constantly looking at the data – though of course our team helped design and develop the system in the first place to allow for this . For obvious reasons we started out with SPA on our largest and highest capex pumps as these offer the largest potential efficiency savings – and mining is one of the largest markets for these . But we can also install the technology on smaller pumps as well , such as those used with filter presses as an example . IM
20 International Mining | JUNE 2024