PowerX: “We thrive where there is uncertainty”

How implementing AI at the heart of day-to-day operations drives the next round of efficiencies in passive infrastructure

Read this article to learn:

  • The history of PowerX and their mission
  • Use cases for AI and ML to impact energy consumption
  • Case studies on how PowerX has used these technologies to drive efficiencies
  • Where else can AI and ML impact passive infrastructure.

Data capture and analysis is front of mind for all towercos looking to optimise the operational efficiency of the passive infrastructure they provide. The challenge that comes after collecting data from a portfolio of sites is knowing what to do with it and how to quickly extract and implement efficiencies at scale.

While monitoring systems can flag when something is going wrong, it can be incredibly difficult and time consuming for those responsible for analysing the data to spot where a site is operating sub-optimally. It is harder still to use the data to recommend the necessary adjustments and finally implement them, especially as portfolios are growing in size and complexity while the information collected from them is increasing alongside. Enter PowerX.

TowerXchange: Can you tell us a little more about PowerX?

Justin Head, CEO, PowerX: I founded PowerX in 2017 out of a determination to solve a problem that has been in the industry for years. During my time in telecoms, I’ve been struck to see there’s been so much investment in the active infrastructure to enable faster and faster connectivity – but the passive infrastructure has really lagged behind.

Passive infrastructure investment lacks the intelligence that its active counterpart has been afforded over the years. There really has been no data intelligence available for tower management teams who run complex tower infrastructures in the same way to help deliver efficiencies.

So I decided to change that. I set up PowerX and we developed the world’s first artificial intelligence platform, focused solely on boosting the performance inefficiencies of cell phone towers. We help businesses in the telecom space meet those massive scale and pace demands that are placed on towercos all the time.
By using AI and machine learning, we bring data science techniques that can harness huge amounts of data to spot how to run things more efficiently, how to reduce CO2, and we bring tools to the towerco or the operator so their teams are able to optimise their network at scale and in real-time.

TowerXchange: Towercos and MNOs are analysing and using the data they are collecting around energy use already. What makes PowerX unique?

Justin Head, CEO, PowerX: We give the towercos and the operators the tools to be able to do it at scale and pace. So yes, there is a lot of monitoring data and analysis of data. Some data is ignored because it’s just so vast. It’s very difficult for a human to get their head around the granularity of the data when they manage thousands of towers and when site-level data is in most cases not in the easiest of formats to read, let alone come up with insights and areas for optimisations to drive down cost and CO2 emissions at scale.

One of the other challenges of doing this manually is that networks are only going to increase in complexity and scale over time. We all agree to that, which means the amount of data is going to increase and the amount of analysis that is needed is going to increase and the importance of running an efficient network is going to increase.

In my opinion, the only way to support these networks is with machine learning, artificial intelligence, and some very clever data science techniques, which is what PowerX offers.

TowerXchange: Achieving tower efficiencies, be it energy, maintenance, CAPEX or OPEX, is bread and butter to towercos and operators. They have been doing it for years. What benefits can towercos and operators derive from using artificial intelligence solutions?

Justin Head, CEO, PowerX: The PowerX platform is agnostic to hardware and monitoring systems installed at the towers. That’s important because it allows us to sit as an intelligent layer above them and take data from all the different power systems, and all the different RMS systems. We run the data through our platform and we can spot anomalies, opportunities for efficiency improvements, and opportunities for CO2 and cost reduction.

Where the artificial intelligence layer we bring comes in is that it allows tower management teams to optimise towers individually, at scale, all at the same time. Each individual tower is unique and operates in a unique environment where they will have specific requirements. No two towers are identical.

By using artificial intelligence, we help optimise each individual tower’s performance, and automatically make adjustments at a specific site, as soon as the system spots that there is something that could be improved. We do this at scale, 24 hours a day and most importantly it’s automatic, with no manual interventions or visits to the sites. That’s a benefit of AI that you won’t get anywhere else.

TowerXchange: What are some examples of how your clients have used data collection and analysis to optimise energy use?

Justin Head, CEO, PowerX: I can give you three specific site-level examples of how we’ve been able to make a huge difference to the way that a site is run.

Firstly, a very simple and straightforward case – one of our customers had a diesel generator that would run every night to recharge the site batteries to 100% so every day they were guaranteed that if the solar solution at their hybrid site didn’t produce the kind of power that they needed, the site would continue to be powered.

We implemented our platform and by looking at the data, we realised that the batteries didn’t actually need to be recharged to 100% every single night because they never got discharged below a certain point. Rather than recharging every single night and effectively waste a lot of money and increase the CO2 footprint, our platform recharges the batteries as and when required to be recharged based on a prediction of the future load based on recent learnings and how much power there’s going to be delivered by the solar and the hybrid sites. This has made the site far more efficient in terms of both financials and carbon emissions. What is even more important, this types of efficiencies are spotted and rectified in real-time across thousands of sites.

Another quick example: We had a client whose site was running perfectly well. That is to say, everything was doing exactly what it should do, and all the lights were green on the monitoring equipment. However, PowerX AI spotted that when the site was receiving power from the solar implementation, which was enough power to run it, the diesel generator was also running for eight to nine hours during a day.

Looking into this problem with the customer, it was apparent that there was a piece of equipment in the cabin that was overheating.

What this meant was that both air conditioning units in the cabin had to run on full power to keep the cabin cool. In order to power the air conditioners, the diesel generator had to kick in, as the solar solution was unable to power the site and both the air conditioner units. Once the overheating piece of equipment was replaced, it ended up saving about 1000 litres of fuel a month as the generator was used far less often. No one knows how long the site had been running like this, but without PowerX it most likely still would be.

My third example is related to PowerX’s ability to detect where assets are underutilised. We had a customer who was generating solar power all day, but their battery capacity wasn’t sufficient enough to capture it all. So just after lunchtime, all the solar power that was being generated was wasted.

So that was the first problem. The second problem was that meant that at 4am the diesel generator from the site had to fire up because the battery capacity wasn’t enough to take it through to when the sun came up.

The platform was able to dynamically adjust the batteries’ depth of discharge at individual site-level to use more solar energy available. In addition, the information from our platform helped the customer see how much diesel they would be saving by making an adjustment to their battery capacity. For sites where solar has already been implemented, it was only a small additional investment to upgrade battery capacity – but with huge upsides.

I think all three of these examples show that just because nothing is technically wrong with a site, it doesn’t mean that everything is right. There can be inefficiencies hidden under the surface, and hidden in the data that it would be very difficult for a human to spot. By harnessing the power of AI, tower teams can deliver efficiencies at scale with reduced manual interventions and visit to the sites.

TowerXchange: Can you tell us more about the predictive modelling functionality? Lots of towercos are looking to make sure it is as easy for them as possible to accommodate new tenants and equipment, so how can PowerX help with that?

Justin Head, CEO, PowerX: When additional tenants are added and when towers need 4G and 5G upgrades we champion the use of the real data that has actually come from the tower as key inputs into the decision makings to de-risk for both the lessor and the tenant.

Specifically, we work with towercos to run “what if” scenarios to understand the implications on the power infrastructure that it is currently in place. We also benchmark between towers to understand how efficient different sites are and that really helps speed up decision making, and empowers tower cos and operators to make smarter decisions around their networks as well.

And of course, this still ties into our core philosophy of using real data running through our AI platform to ensure that the tower co has the absolute best tools at its disposal to make those decisions at scale, across their whole network, on an individual tower basis.

TowerXchange Which type of sites benefit the most from PowerX?

Justin Head, CEO, PowerX: We thrive where there is uncertainty. Whether that be grid uncertainty, solar uncertainty, battery uncertainty – it doesn’t matter. If there are volatile conditions, then we have the power to help paint a clearer picture of what is going on. If you think about hybrid sites, they obviously require complex power provision to ensure network availability and tower up time. They’re often in remote locations with increasingly growing populations and the demand for mobile connectivity is growing alongside.

So with that in mind, running an energy efficient operation is vital. But it’s not just rural, hybrid sites, its everywhere really – everywhere needs to be able to support greater co-location or greater coverage. We are also seeing the emergence of OpenRAN, and this again is going to require new power solutions.

TowerXchange: What else does the future hold for PowerX? Where else do you see the machine learning platform being able to create efficiencies?

Justin Head, CEO, PowerX: Looking beyond the immediate passive infrastructure, we see a lot of benefits of the platform. Leveraging data in this kind of way might be new to the tower world, but look at some of the largest tech companies in the world – Netflix, Amazon, Google, Uber – all of these use artificial intelligence and machine learning as a core business practice. And I believe its now time for our industry to embrace that, and take advantage of the huge amounts of data that will be generated as edge data centres increase in scale as 5G is rolled out for example.

Having processes, systems and tools in place to leverage that data is absolutely the only way you can ensure an optimal network at the lowest possible cost with the lowest possible carbon footprint. We’ve talked a lot about that today, but as more trust is built up in the analytics we provide, I’m certain PowerX will be able to help towercos design optimal towers and networks across the world.

To learn more about PowerX schedule a meeting at TowerXchange Meetup Africa (7-8 October, Online; and 12-13 October, Dubai)

Leave a Reply