There is increased adoption of EVs globally and this has been spurred by the need to cut down on greenhouse gas emissions, reduce air pollution and impact climate change. Coincidentally, this adoption is led by the countries with the largest carbon footprint. According to the intergovernmental panel on climate change, the transportation sector, specifically vehicles account for 23% of the total energy related carbon dioxide emissions. This may explain the importance of shifting from the internal combustion engine to the electric mobility systems, specifically, electric vehicles. A study conducted in 2021 by Eurelectric reveals that 70 million EVs could be integrated in the EU transport system by the year 2030. While there is no sufficient data for the global South, it suffices to say that this trend is expected to follow a similar trajectory, even though gradually.
It is thus important to consider the likely challenges that may arise due to the future scenarios with EV systems. In general, EVs like the other distributed energy resources affect the three facets of network planning including load forecasting, reliability planning and network design. Of these three, this discussion is centered on the last component because of its criticality in the power system objective: ensuring that the power delivered to the end user is safe, reliable and within the prescribed quality of supply limits (voltage, frequency etc.). From this perspective, aspects such as feeder conductor selection and reinforcement as well as distribution transformer sizing become crucial.Â
So, what has this got to do with EVs, one might ask. Everything. From a distribution network operator’s perspective, EVs just like other distributed energy resources like behind the meter solar photovoltaics (PV)are more likely to disrupt the operations of low voltage distribution feeders as they are usually connected there. Consider the scenario in Europe where 70 million EVs are expected to be on the road by 2030. A planner does not see the EVs. The planner sees an increase in the network load – a load whose characteristics introduce problems in the otherwise stable distribution system.Â
Closer home in Africa, a recent study conducted in South Africa observes that if 10% the vehicles in South Africa were to become electric, there would be an additional 6GW of demand added to the peak hour load. To underscore the difficulty that these may bring, and the likely deterioration in distribution network, one needs to first appreciate the difficulty that ESKOM already has in meeting the current national demand. To add about 30 to 50 GWh of energy requirement may not be manageable. It is not possible to imagine the impact that a load of that magnitude on the ageing and fragile KPLC network in Kenya. And even though this has got a lot to do with reliability planning, this discussion seeks to shed light on the very nature of EVs which make their requirement unique – and what we are currently working on.Â
Â
What are these unique challenges that planners must confront? How do we enhance grid preparedness for a future with EVs?Â
To answer this question, we must first characterize the EV system. To that end, there is a consensus within the technical expert community that a high penetration of variable and renewable distributed energy resources, results in difficulties in the operations of distribution networks given that most of them are connected to the low voltage network level. For EVs, several aspects require consideration and extensive modeling. Planners must understand the relationship between the EV state of charge (SoC) and the mobility factors, which include but are not limited to distance traveled, time of arrival as well as time of departure. EV-centric factors like the vehicle charge and discharge rate are also critical in determining the EV SoC as well as the time it takes for the EV to attain full charge. This list of complexities is not exhaustive. Further considerations must be given to the EV user time of charging preference, which usually determines whether the feeder will be congested. The key take away is to remember that these crucial variables are uncertain – and variable and as such cannot be modeled using the business as usual approaches.
Â
What therefore is being done? – The answer to this question, we would say, a lot. As it can be deduced from the above discussion, both the EV-user factors and the mobility dependent factors vary across demographics and even in the case of homogeneous power requirements (capacity), there is a great deal of variation arising the varying time of use preferences, different initial SoC levels across different EV users, and more importantly for the planner, the variation in the location of the EV loads on the network. Considering this along with the normal customer load, we have formulated an approach to simulate the impact of EVs on the network, for a range of penetration levels. This method has been tested on several feeders and the performance indicates that planners can obtain the granular voltage and loading performance of the network nodes and branches under varying EV penetration levels.Â
Â
Why is this important? – This work can help the planner to identify the sections of the network that are likely to be constrained under any planned scenario of EVs. In a lay man’s language, it is possible to identify areas of the network that could be easily overloaded or could experience extreme voltage drop in one’s design. Initially developed as a diagnostic tool for the existing networks, this tool can be extensively applied in the sizing and selection of cables for new electrification networks that incorporate EVs in their planning objectives. Lastly, for the existing distribution network’s this tool can evaluate the individual branch and node and node capacity limits, under a given penetration level while at the same time provide an extended application to size alternative cables for the constrained conductors where possible.Â
What is the way forward? – Considering that it would be a disastrous venture to allow uncoordinated charging of EVs on the network, the next phase revolves around testing the method under varying tariff and policy scenarios. This will help formulate appropriate response to policy makers from a technical perspective.