Interview: Catherine Wolfram, Principal Investigator
What impact would you like the GridWatch project to have?
Generally, we’re looking to prove the suite of technologies that the engineers on our team have developed to report and track power outages and other aspects of grid operations. We think the technology has broad applicability, so we’re excited to demonstrate its performance at scale.
We will also use the data in several ways – for example, to understand how reliability impacts economic growth and to measure how households and firms adjust their decisions in the face of poor reliability.
What impact do you think energy reliability can have on households and businesses?
Between 2012 and 2015, persistent power failures in Ghana negatively affected its economy and gave rise to the term ‘Dumsor’, meaning ‘lights off-on’ in the local Akan language.
Anecdotally, we know that frequent outages constrain the economic wellbeing of households and small businesses. And, according to the most recent World Bank Enterprise Surveys (2013), 61.2 per cent of firms in Ghana see electricity reliability as a major constraint, with firms reporting an average of over 700 hours of outages annually, compared to 1.5 hours for firms in the US.
Customers primarily consume electricity through electrical appliances, such as refrigerators or televisions, and machinery. Improved power reliability might allow them to improve their electricity usage through these channels, for example by operating their equipment for more hours, by purchasing better quality equipment, or by spending less money on replacements for reliable grid power, such as backup generators or voltage stabilisers. This could lead to higher productivity and revenues for firms and improve wellbeing for households.
What led you to carry the research out in Ghana?
We partnered with the Millennium Challenge Corporation (MCC) several years ago and they have a massive $500million+ compact to improve reliability in Ghana, so it seemed like the right place and the right time.
Phase 1 has recently been completed – what key things were achieved?
Since June 2018, we have successfully executed two large-scale deployments of GridWatch technologies, through pilot funding from MCC and Phase 1 funding from EEG.
In addition to providing outage, voltage, and frequency data on the grid at minute-level timescales, our deployments enabled us to refine and improve the designs of both our hardware and software, and we also built out technology to monitor and manage the deployments and analyse grid behaviour. Socioeconomic surveys were performed to provide data and support the deployment and development of the technology.
Deploying the GridWatch suite of technologies has allowed us to identify spatially and temporally related power state change events in three districts of Accra (Achimota, Dansoman, and Kaneshie) and classify these into distinct power outages. We were able to identify specific components of the electricity grid that are likely to be driving these power outages.
The geographic and temporal density of outage detection by the GridWatch technologies is a significant improvement on the data currently employed to quantify the quality of the grid in Accra.
We were able to demonstrate that GridWatch is fully capable of supporting an impact evaluation of infrastructure investments and improvements in an electric grid. By matching power outage events to specific elements of the grid, we can confidently identify heterogeneous power quality for a wide set of respondents. We are well-positioned to rigorously evaluate the impact of investments in the electricity sector on grid reliability and socioeconomic outcomes.
Additionally, we have been selected by the Millennium Development Authority (MiDA), a Ghanaian government agency, to be the sole-source provider of a reliability monitoring system that will aid them in monitoring and evaluating the efficacy of their power sector investments across Accra.
What challenges did you face?
There were a number of challenges, ranging from the engineering challenges of creating a new technology all the way to those related to contracting and invoicing in an unfamiliar country.
The engineering challenges were anticipated, and many were overcome with repeated pilot deployments and iterations based on lessons learned. The biggest challenges we experienced in engineering related to interactions with our participants, whether it was making sure they kept either PowerWatch or DumsorWatch installed or making sure that we sent them their participation incentives on time.
The problems related to conducting business were more surprising. These were overcome primarily through in-person discussions with our team, but led to some significant delays.
Now Phase 1 is complete, what’s next?
Phase 1 showed us a lot about how to take direct measurements of energy reliability. Moving forward we are imagining a two-pronged approach, the first being scaling these measurements both in Ghana and into other countries. We have designed our sensors to be easily deployable globally, and have built supporting systems to allow the deployment to scale from hundreds of sensors to many thousands without significant additional human effort. In fact, a big part of the engineering work was to make sure that as the number of sensors increased, our fairly small team could continue running the sensor network successfully.
Our second prong is really diving much more deeply into the data we have been collecting. For example, can the data teach us different planning and operational methods for grids in various social and economic contexts? From these measurements can we learn the root cause of failures so they can more quickly and directly be addressed? Or, somewhat simply, can we use our data to better determine exactly where repair trucks should be sent?
Why is energy reliability so hard to measure?
There is a lack of accurate data related to outages and restorations, and the reason for this is primarily that it costs money to instrument the grid with sensors or smart meters. Many utilities do not have the funds to invest in these technologies, particularly the large investment required to provide widespread coverage at the distribution (customer) end of their grids.
As a result, what we have seen in Ghana and Kenya, and believe to be true in many other countries, is that the most common estimates of energy reliability available are those calculated from customer reports of outages and subsequent manual classifications by utility workers of outage scope and duration. With humans so involved in the measurement, there are numerous opportunities for error to be introduced, whether from under-reporting (when outages are common occurrences how many people actually still report them?), misclassification (was the reporting customer location recorded accurately? Do different reports really belong to the same outage or are there different outages simultaneously occurring?), or misaligned incentives (are there utility target performance metrics that may influence the manual classification process?).
In the GridWatch project, smartphone apps and sensors have been deployed – what impact is technology having on research?
The technology the engineers have developed has been instrumental to giving us unique insights on the grid. For example, we will be able to pinpoint which sections are the most prone to power outages, and whether power outages are primarily driven by failures in the low-voltage system or the medium- and high-voltage systems. On the social science side, this level of granularity will give us a lot of leverage to understand how people who are experiencing different patterns of outages are adjusting their economic lives in response to blackouts and brownouts.
What interests you most about the energy sector?
Energy is a fundamental input to so many aspects of modern life – we use electricity at home to cook, clean, do laundry, and watch TV, and it’s also essential to power our hospitals, schools, factories, police stations, etc. At the same time, the energy sector is the main contributor to climate change and other environmental issues, so it’s really crucial that we make the best decisions about how energy is priced, produced, regulated, and used.