Tanzania woman points at meter

Project Details

Thematic Project
Catherine Wolfram
East Africa
West Africa
University of California, Berkeley
Catherine Wolfram profile picture

Catherine Wolfram

Principal Investigator


Measuring electricity reliability with GridWatch sensors 


Background, challenges and context

Many studies on the microeconomics of electrification in developing countries focus primarily on estimating the impacts of providing rural households and communities with access to electricity for the first time. In urban areas, however, many households and businesses are already connected to the grid. In these settings, the primary issue is reliability rather than lack of access – with power outages and voltage fluctuations being widespread and frequent. Studies on the macroeconomic impacts of outages suggest the costs may be large – in the range of 2-3% reduction in long-run GDP per capita.

Anecdotal evidence suggests that frequent outages and voltage fluctuations constrain the economic wellbeing of households and small businesses by discouraging investments in welfare-improving and income-generating appliances (such as fans, refrigerators, or production machinery), and reducing the output from existing appliances.

In surveys carried out with businesses in Ghana on the costs of voltage quality between 2018 and 2022, 26% of respondents reported that voltage fluctuations had damaged at least one appliance in the past year, and they had paid on average USD50 to repair or replace damaged appliances.

Households and firms might also need to devote considerable resources to alternatives to reliable grid power, including backup generators and voltage regulators, meaning they can’t spend money on other more profitable investments.

The scope of these problems is not well understood, and many countries have only limited data about reliability issues. Many utilities must therefore make important operational decisions without knowing the frequency, duration, and geographic extent of power outages and voltage fluctuations. And governments must make important investment decisions without knowing the extent to which the local economy is being affected, or how to improve grid reliability. 


Research overview and objectives  

To help fill these knowledge gaps, a team of engineers and economists has developed, tested, deployed, and operated a suite of low-cost, remote sensing devices, collectively called GridWatch. Unlike existing technologies, GridWatch automatically measures and centrally collects data on grid reliability, voltage quality, and power outages, publicly and independently of utility reports. 

To detect the presence and absence of grid power, the team developed a fixed-point sensor called PowerWatch that is plugged in at a home or business, with a battery to allow for continuous reporting throughout a power outage. PowerWatch senses power outages and power restorations timestamped to the millisecond, as well as GPS-based location, voltage, and grid frequency.

In Phase 1 of the project, the GridWatch technology recorded power outages, voltage fluctuations, and frequency instabilities in Accra, Ghana. It generated detailed reliability data for sites across the city on an ongoing basis, at a temporally and geographically higher frequency than existing utility outage recording and logging systems (many of which rely on manual detection and processing).

During Phase 2, the team aimed to learn more about the socioeconomic impacts of improved reliability and voltage quality, and to estimate the causal effect on outcomes such as wellbeing, productivity, and health for the residents of Accra.

To increase understanding of reliability in rural settings as well as urban ones, in Phase 2 the team also expanded the deployment of GridWatch to rural Kenya. The technology was used to collect data on the Last Mile Connectivity Project (LMCP) grid extension construction process. New infrastructure is being funded by both the World Bank and the African Development Bank; the team investigated whether the turn-key contracting method employed by the African Development Bank led to different power quality and reliability outcomes than the World Bank’s segregated contracting methodology.

Deploying the technology in Kenya also enabled the team’s engineers to adapt and test the GridWatch power sensors in a new environment, providing important information on their performance.


Research methodology

During the pilot stage (funded by the Millennium Challenge Corporation) and the EEG-funded Phase 1, the team installed over 400 PowerWatch sensors in households and firms in Accra, Ghana.

The team deployed the devices across the urban districts of Achimota, Dansoman, and Kaneshie in the Accra Metropolitan Area. The full sample included a set of sites being targeted for improvement by the Ghana Power Compact, a USD316 million investment designed to improve the grid generation, transmission, and distribution systems in Ghana, as well as a quasi-randomly selected set of control sites.

In Phase 2, the team investigated the link between connection quality (in terms of outages and reliability) and economic outcomes such as the purchasing and use of appliances, economic activity and earnings, and wellbeing. Socioeconomic surveys were conducted with 1,000 households and businesses located at treatment sites and 1,000 households and businesses located at control sites. Since these households and businesses were identical in expectation, and differed only in terms of the quality and reliability of power they received, this allowed the team to causally identify the socioeconomic impacts of poor reliability and voltage fluctuations. The analyses of causal impacts centred around the effect of improved reliability and voltage for households (such as decisions about which types of appliances to buy) and firms (such as whether new businesses are more likely to emerge, whether investments in income-generating capital increase, and whether profits grow).

In Kenya, 100 sensors were deployed in households and firms in 25 villages (four sensors per village) for two months at a time, covering 100 villages in total (rotating villages allowed the researchers to increase the breadth of their sample without having to deploy and manage more sensors). The study covered a large geographic rural area that included the counties of Kakamega, Kericho, Kisumu, Nandi, and Vihiga. The sensors maintained functionality in the new rural context and the quality of cellular connectivity was good enough to receive real-time data.


Research results, key messages, and recommendations

The geographic and temporal density of outage detection is a significant improvement on the data currently used to quantify the quality of the grid in Accra. The team was able to identify spatially and temporally related power state change events in Achimota, Dansoman, and Kaneshie, classify them into distinct power outages, and identify specific components of the electricity grid that are likely to be driving the outages.

GridWatch’s improved information could help utilities to assess the extent of an outage quickly and accurately, so they can dispatch repair crews where they’re most needed to better prioritise limited resources. With more analysis, the data may also help utilities to carry out better preventative maintenance.

The rich analytics and real-time monitoring of the grid will enable the team to measure the economic costs of unreliable electricity systems and answer important research questions. In the team’s ongoing work, combining the data with geospatial demographic and socioeconomic information will allow the researchers to investigate whether heterogeneity in voltage quality and power outages is correlated with socioeconomic status.

In Kenya, deploying GridWatch technology across a subset of the study sites allowed the team to estimate directly whether differences in donor contracting procedures impacted electricity system construction quality. Electricity distribution systems built under the World Bank’s segregated contracting procedures improved construction quality by 0.6 standard deviations relative to the turn-key methodology used by the African Development Bank. This may have important implications in the long-term longevity of the network, but did not improve power quality or voltage in the short term.

Preliminary findings suggest that both Ghana and Kenya experience periods of poor voltage quality. Voltage quality is generally worse in urban Ghana than in rural Kenya, which could be due to differences in infrastructure quality or load.

In both locations, voltage quality is lowest during the early evening, which is probably because this is when load peaks. The data also revealed significant seasonal variation in voltage quality.

GridWatch has the potential to help governments and utilities make more informed investment decisions. The research team will work with regional policy makers to understand how the information can best be used.

In addition, the team has been selected by Millennium Development Authority to be the sole-source provider of a reliability monitoring system that will aid the monitoring and evaluation of the efficacy of its power sector investments across Accra.


Local partners

Phase 1: Government of Ghana and local utilities

Phase 2: University of Ghana Institute of Statistical, Social and Economic Research (ISSER), Kenya Power



External Journal Articles
Can Cell Phones Help Improve Electricity Reliability?
Hardware, apps, and surveys at scale: insights from measuring grid reliability in Accra, Ghana