Beyond meter data to predictive intelligence


As the amount of data available to utilities via smart meters and other smart grid devices increases, the challenge is to turn this sea of data into intelligence. Engerati spoke with Joachim Gruber, senior manager, metering & EDM-services, EnBW AG; Rob Kopmeiners, research scientist, Alliander; and David Svensson, system manager, Göteborg Energi AB to get their perspectives.

According to Gruber, in the past utilities have not necessarily had the ability to optimise meter data and the amount of data they have or will have access to. In the future, however, he believes utilities will use this data more and more to predict the future. Kopmeiner agrees, saying that analytic maturity is quite country specific (in the Netherlands they are in the process of setting up and formulating the plans for a smart metering system). “I think the next step is where we will see the value proposition which will mean we can leverage smart metering data across the grid.”

Says Svensson: “We have had smart metering for some time, and I think it’s crucial to get more value out of having that much data. It takes time and knowledge to be able to use it in a way that can profit: for example, grid operations. But we have lowered or increased the voltage levels in a substation based on information from the smart meters across approximately 15 areas. So that’s one of the new areas that we’ve been working on.”

Taking a closer look at the eco-system, the key players are the ones who are collecting and then distributing the data – generally, the metering operator. However, because of the strict separation between suppliers and grid operators in many European countries, there can be a misalignment of the perceived benefits derived from smart meters and the resulting data. Regulators may want to drive energy efficiency benefits, but grid operators may well want to derive more benefits. As a result, the quality of the metering operators’ service is crucial in order to effectively utilise the data when it comes to direct interaction with the market if, for instance, the utility is wanting to institute variable tariffs or flexible grid tariffs. The key challenge is how to deliver benefit from this data and bring the various data streams together to do so.

Kopmeiner is of the opinion that it’s ultimately the customer that needs to profit from this data – by bringing the cost down.

To centralise or decentralise?

It’s crucial to have data that is collected and distributed to all the market participants in a non-discriminatory way. The number of players in the market is going to continue to grow, with Gruber of the opinion that the future market make up will include a DSO, a TSO, the retailer and the customer – but with other market players such as aggregators and prosumers who are trading energy on the wholesale market. He also believes that the metering systems and the meter data that is collected today will be used very differently in the future.

This could include a model where data is handled in a decentralised fashion by the DSO. This model has different benefits and conveniences from more centralised approaches. If you have a centralized data hub to distribute, collect and distribute the data, this is also a single point of failure, something which is less likely in a more decentralized energy system. However, a pro to a centralised approach is that the market participants don’t have to enter into agreements for data access with a lot of other parties active in the market – they only have to go to the central hub to get the data where, hopefully, this data has already been validated.

In a decentralized approach, the benefit of it is that those who are collecting the data are closer to the source of that data. This is particularly a challenge where there are complex metering points, where there are several meters and the metering operator has to add or subtract all the meter readings to get a real meter value.

The disadvantage of a decentralised approach, on the other hand, is the inconvenience for market participants who have to interconnect and communicate with more players, potentially creating dis-synergies if it comes to structure and infrastructure.

“The DSL collects data in a centralized way and distributes them through a portal to the parties that would require this information,” says Gruber.

Kopmeier reports that in the Netherlands: “Privacy regulations do not allow us to store this data and this is a fundamental issue with collecting data from a centralized point of view. We do not have a solution for that yet. It is possible to acquire this data from the metering system but to store and safeguard this for later usage is quite a difficult task at the moment.”

“There is something to be said for decentralised data. With the metering system that we have in the Netherlands we do have a consumer interface and we see this as the active port for all kinds of new parties – retailers, commercial parties – to get their data directly from the consumer. They can place a box next to the meter in the homes of the end users and collect the data. There is an issue, and I think Joachim already addressed it: if you create a vendor lock-in and that the data is only available to one commercial party, it can be an annoyance to the end user. So I think it’s still in question how to do this. From our DSO perspective we prefer the centralized data model.”

Svensson reports that some of the meters used by Göteborg Energi have ports that allow for data extraction directly from the meter, but he questions if this is the right way to go because of the need for extra equipment, and there is always the question of who will pick up the cost for this. Additionally, customers may not want to, or be interested in, checking their energy consumption as it is time consuming and requires some knowledge in order for the consumer to benefit from it.

Gruber believes that standardisation across Europe is key when it comes to data. He says that there are already discussions to institute a similar system to the Green Button initiative used in the US. This system has the advantage in that it provides clear guidance on the form the data is accessible in and ways in which the data can be accessed; additionally, the customer must approve who may/may not have access to this data. This provides a level of transparency for the customer around who has access to their data and allows the customer to manage access afterwards.

The importance of insights

Gruber says access to data is particularly important, as in many cases in Germany, across low and medium level voltage networks, the DSO is often unaware of the state of the system because of a lack of sensors. Sensors provide access to the state of the grid, and the value that is driven from that is not only operational efficiency and service level efficiency, but also valuable input for grid planning. In the future, it also makes managing the network intelligently a possibility – for instance requesting service flexibility or services from the market based on the data derived from the meter data.

He believes that technical information gathered around grid operations and grid efficiency belongs to the grid operator and clearly distinguishes the difference between consumer data and grid data. He adds that customer buy-in is vital to get the initial installation approved, but after that, it is equally important to assure the customer that nothing will happen to their data, that their privacy is respected and protected.

Yet in the Netherlands, the usage of data is bound by the purpose it will be used for, so if you are not able to provide the exact purpose for the data, it could be difficult to get access to this data. This is in addition to the difficulty of gathering and storing data. Kopmeier believes, however, that over time, the use of the data will also evolve, as will our perceptions thereof.

Svensson says that Göteborg Energi has used this information to good effect already. “From our 265,000 smart meters they found voltage values that were not physically possible. So I decided to investigate meters with voltage outside what’s expected, which means 230 Volts ± 10%. When we visited the first meter, we measured the voltage and current in the cable and compared it with what the meter was reporting. And phase one on that meter did not register the correct voltage or current or power on that phase. So we changed the meter – and the new meter registered twice as much consumption as the old one.

“The more research I’ve done the more interesting things I have discovered. All three of our meter brands had small but different kind of problems that I could identify by voltage and current values. Not only did I find meter related problems, but also meters not installed correctly; for example screws not tightened enough. I have realized that the meter has its own reality. If something affects the hardware for example, or components, the meter’s value changes, even if the physical reality stays the same.

“If the meter thinks it’s 300V in the cable on one phase instead of the actual 230V, it will register 0.3 kilowatts instead of 0.23 kilowatts at 1 Ampere load. My conclusion is that meter errors or installation errors cause patterns in voltage and current data. When I have identified a problem I have tried to define the pattern. One bad reading is needed to identify the specific problem and we get momentary voltage and current twice a day from each smart meter in our metering system.

“Today we need to do a manual search and analyse the data and create work orders for the meter, and we must check the installation. But my goal is to display those meters that need to be changed on their line or webpage in the future.”

The Göteborg Energi set-up provides two readings per day that gives the momentary value right then and there, along with voltage and current readings. This is then stored in a database and normal voltage and current is extracted. Then the remaining information is analysed. “I look for patterns, and I create work orders and get feedback from the technician out in the field as to what the cause was. We always compare the cable with what we get in the metering system and then you know if there’s something wrong with the meter or with the cables.”

If there is an issue with a meter, the utility is able to correct the consumer’s bill. In most cases, because there is a loss of the usage on just one phase, for example, the consumer is paying for two thirds of what they are consuming. In that case, the utility will often simply provide the customer with a new meter and ensure they are billed correctly from there.

“I think right now we have about 7,000 meters in the database, but that’s not because they are all problematic. We also check, for example, if there are meters with high loads because customers have change their fuse; and because some of the customers don’t inform us when they’d done that, we have to know what fuse they have, because some of the meters only allow 35 Amps. If they change the fuse size to 50 Amps it’s not that good. So we need to check that too.”

How can data be used beyond the utility field of activities?

There are quite a lot of opportunities, but concerns around data privacy and customer privacy may be a hindrance. Gruber however thinks that there may be opportunities for insurance or security companies or for any other kinds of living assistance services. For instance, a customer could inform their insurance or security company that he is away on holiday, but if there is power usage in that house that is contrary to that information, they could investigate to ensure there is nothing wrong. These are opportunities that could be based on meter data.

Big data and open data will allow commercial parties more scope as they are much less limited by those privacy rules. “I mean we all know Facebook, we know social media, where everything is exposed through the internet to everyone,” Kopmeier explains. “And in that sense I think it will be an opportunity if you can open up metering data, and I think we can do this by a local port completely controlled by the consumer and he can connect his own application directly to his smart meter if needed or to any service that can be thought of. As an example – I just heard it yesterday, although I can’t verify if it’s true – but there appears to be an application for your mobile phone that can identify your electronic equipment based on the patterns: unique very high resolution patterns, which we cannot provide through a centralized data system. But you can get it through this local port. Nowadays, you can get one second data resolution and based on that you can really analyse devices and what’s going on; and based on the app, you can get ‘commercial’ advertisements, but also really good advice such as ‘This refrigerator is taking too much energy. You should think about buying a new one.’”

Combined services

In the Netherlands gas and electricity are combined as a service. Yet in other countries it might be completely separate organizations that collect this data. Kopmeiner believes “we really should strive for an open data market where data can be leveraged by the consumer on a level playing field. I think that would be very important to do that. Specifically, for gas, as there’s less interest in general from this sector. In the Netherlands, up to 60% of our energy usage is gas, with only 30% electricity.”

In Germany the new Act on smart metering will require that gas meters be connected to the same gateway as the electricity meter. This will mean that household consumption, no matter if it is gas or electricity, can be connected via a single gateway, providing combined data to a single point. This gives a lot of transparency to the customer and a lot of insights for those who might propose additional services.

“It’s a really interesting subject – to get a better understanding of the customer,” concludes Gruber. MI