ELECTRIC VEHICLE SALES AND INFRASTRUCTURE ANALYSIS

INTRODUCTION

In March of 2021, the Argonne National Laboratory released a report showcasing the continued popularity of electric vehicles (EVs) throughout the United States. Both battery and plug-in hybrid EVs had their strongest sales in that month since December 2018 and reached record high sales. The Biden administration is looking to have only half of new vehicle sales in 2030 be traditional combustion engines with half of new sales to be EVs. However, for PEVs to continue their growth, there must be charging infrastructure in place, which is often listed as a primary reason people are unwilling to make the jump to electric vehicles. Without proper infrastructure in place, reaching the 2030 goal will be extremely difficult (Ge, Simeone, Duvall, & Wood, 2021).

Louisiana currently lags other states in the adoption of EVs. It ranks 34th in the United States with only 27,452 total EVs sold (compared to nearly 2 million in California and almost 400,000 in Florida). When adjusted for population, California has eight times the EV sales of Louisiana and Florida three times the EV sales of Louisiana (Alliance for Automotive Innovation, 2021).

The primary location that most EV owners will have consistent access to charging infrastructure will be at the home. The first group of people who will be able to transition to at-home charging will be homeowners who live in single family homes. Immediately, this restricts the amount of people who have access, as apartment renters are unlikely to have broad access to a community charging station and renters in single family homes must follow the lead of the homeowner or landlord. These renters will be unable to make modifications to the home without the owner's permission and both the renter and homeowner are unlikely to pay for the expense of installing an at-home charger because the renter is not a permanent resident and the homeowner does not want to pay for a home upgrade that will not directly benefit them (Traut, Cherng, Hendrickson, & Michalek, 2013).

USING DIFFERENT FORECAST METHODS TO PREDICT ELECTRIC VEHICLE SALES IN LOUISIANA

Using the census data (American Community Survey (ACS), 2019), there are 1,741,072 total light weight vehicles registered in Louisiana. According to the Alliance for Automotive Innovation (2021), there were only 4,127 newly registered EVs in Louisiana as of June 2021, and 27,452 total EVs registered in Louisiana since 2011 (Figure 1). Even if every single EV sold in the last 10 years were still on the road, that would only represent 1.6% of all vehicles in Louisiana.

View Figure

• Download as Powerpoint
• Download Figure

In order to observe trends and predict sales forecasts for future sales of EVs, different forecasting models can be used to calculate the future sales in Louisiana.

The first forecasting model run was for moving averages. This model uses the previous sales numbers and finds the average over a set period. Using the equation in Equation 1, where \(\def\upalpha{\unicode[Times]{x3B1}}\)\(\def\upbeta{\unicode[Times]{x3B2}}\)\(\def\upgamma{\unicode[Times]{x3B3}}\)\(\def\updelta{\unicode[Times]{x3B4}}\)\(\def\upvarepsilon{\unicode[Times]{x3B5}}\)\(\def\upzeta{\unicode[Times]{x3B6}}\)\(\def\upeta{\unicode[Times]{x3B7}}\)\(\def\uptheta{\unicode[Times]{x3B8}}\)\(\def\upiota{\unicode[Times]{x3B9}}\)\(\def\upkappa{\unicode[Times]{x3BA}}\)\(\def\uplambda{\unicode[Times]{x3BB}}\)\(\def\upmu{\unicode[Times]{x3BC}}\)\(\def\upnu{\unicode[Times]{x3BD}}\)\(\def\upxi{\unicode[Times]{x3BE}}\)\(\def\upomicron{\unicode[Times]{x3BF}}\)\(\def\uppi{\unicode[Times]{x3C0}}\)\(\def\uprho{\unicode[Times]{x3C1}}\)\(\def\upsigma{\unicode[Times]{x3C3}}\)\(\def\uptau{\unicode[Times]{x3C4}}\)\(\def\upupsilon{\unicode[Times]{x3C5}}\)\(\def\upphi{\unicode[Times]{x3C6}}\)\(\def\upchi{\unicode[Times]{x3C7}}\)\(\def\uppsy{\unicode[Times]{x3C8}}\)\(\def\upomega{\unicode[Times]{x3C9}}\)\(\def\bialpha{\boldsymbol{\alpha}}\)\(\def\bibeta{\boldsymbol{\beta}}\)\(\def\bigamma{\boldsymbol{\gamma}}\)\(\def\bidelta{\boldsymbol{\delta}}\)\(\def\bivarepsilon{\boldsymbol{\varepsilon}}\)\(\def\bizeta{\boldsymbol{\zeta}}\)\(\def\bieta{\boldsymbol{\eta}}\)\(\def\bitheta{\boldsymbol{\theta}}\)\(\def\biiota{\\boldsymbol{\iota}}\)\(\def\bikappa{\boldsymbol{\kappa}}\)\(\def\bilambda{\boldsymbol{\lambda}}\)\(\def\\bimu{\boldsymbol{\mu}}\)\(\def\binu{\boldsymbol{\nu}}\)\(\def\bixi{\boldsymbol{\xi}}\)\(\def\biomicron{\boldsymbol{\micron}}\)\(\def\bipi{\boldsymbol{\pi}}\)\(\def\birho{\boldsymbol{\rho}}\)\(\def\bisigma{\boldsymbol{\sigma}}\)\(\def\bitau{\boldsymbol{\\tau}}\)\(\def\biupsilon{\boldsymbol{\upsilon}}\)\(\def\biphi{\boldsymbol{\phi}}\)\(\def\bichi{\boldsymbol{\chi}}\)\(\def\bipsy{\boldsymbol{\psy}}\)\(\def\biomega{\boldsymbol{\omega}}\)\(\def\bupalpha{\bf{\alpha}}\)\(\def\bupbeta{\bf{\beta}}\)\(\def\bupgamma{\bf{\gamma}}\)\(\def\bupdelta{\bf{\delta}}\)\(\def\bupvarepsilon{\bf{\varepsilon}}\)\(\def\bupzeta{\bf{\zeta}}\)\(\def\bupeta{\bf{\eta}}\)\(\def\buptheta{\bf{\theta}}\)\(\def\bupiota{\bf{\iota}}\)\(\def\bupkappa{\bf{\kappa}}\)\(\def\\buplambda{\bf{\lambda}}\)\(\def\bupmu{\bf{\mu}}\)\(\def\bupnu{\bf{\nu}}\)\(\def\bupxi{\bf{\xi}}\)\(\def\bupomicron{\bf{\micron}}\)\(\def\buppi{\bf{\pi}}\)\(\def\buprho{\bf{\rho}}\)\(\def\bupsigma{\bf{\sigma}}\)\(\def\buptau{\bf{\tau}}\)\(\def\bupupsilon{\bf{\upsilon}}\)\(\def\bupphi{\bf{\phi}}\)\(\def\bupchi{\bf{\chi}}\)\(\def\buppsy{\bf{\psy}}\)\(\def\bupomega{\bf{\omega}}\)\(\def\bGamma{\bf{\Gamma}}\)\(\def\bDelta{\bf{\Delta}}\)\(\def\bTheta{\bf{\Theta}}\)\(\def\bLambda{\bf{\Lambda}}\)\(\def\bXi{\bf{\Xi}}\)\(\def\bPi{\bf{\Pi}}\)\(\def\bSigma{\bf{\Sigma}}\)\(\def\bPhi{\bf{\Phi}}\)\(\def\bPsi{\bf{\Psi}}\)\(\def\bOmega{\bf{\Omega}}\)\({S_k}\) represents the sales for each period k, and n is the number of periods averaged. This model was run using the values 2, 3, 5, and 10 for n. \(\begin{equation}\tag{1}{S_k} = \mathop \sum \limits_{i = k - n}^{k - 1} {{{S_i}} \over n}{\forall _k},k \gt n\end{equation}\)

Using this forecasting model yielded various results for the sales forecast in the period ending in June 2021. Using a two-period moving average, where each period represented six months, seemed to give the most accurate result at 1,754 projected sales for the period ending June 2021. The forecasted result using a three-period average was 1,583, the result using a five-period average was 1,546, and the result using a 10-period moving average was 1,227. While the two-period moving average appeared to give the best result, a decision was made to use the three-period moving average for comparison of different methods. When the three-period moving average was graphed (Figure 2), like the two-period, five-period, and 10-period moving averages, the three-period moving average seemed to fall short compared to the actual values for the most recent year. The squared mean error, which is the square of the difference between the forecasted value and the actual value, and mean absolute percentage error, which is the absolute value of the difference between the forecasted value and the actual value divided by the actual value, were calculated and are listed in Supplemental Table A for each value for the two-period moving average, the three-period moving average and the five-period moving average. These values are used to see how accurate the forecast model was, with an error of zero being completely accurate.

View Figure

• Download as Powerpoint
• Download Figure

The next forecasting model implemented to bring the most recent data closer to the current trend was a weighted moving average. This method allowed the most recent data to carry more weight than the older time periods since the data seems to indicate a quickly changing trend. Using the equation in Equation 2, the weighted moving average was calculated for EV sales in June of 2021. This time a weight \({W_k}\) was added to the moving average equation. \(\begin{equation}\tag{2}{S_k} = \mathop \sum \limits_{i = k - n}^{k - 1} {{{S_i}{W_i}} \over n}{\forall _k},k \gt n\end{equation}\)

The equation was computed with k=3, and W1 = 5%, W2 = 35% and W3 = 60. The weighted average forecast did improve the forecasted sales number to be more in line with current trends. Figure 3 graphs the improved sales predictions. The squared mean error and mean absolute percentage error for the weighted moving average can be found in Supplemental Table B.

View Figure

• Download as Powerpoint
• Download Figure

The weighted forecast gives a predicted value of 1,813 for the period ending in June 2021. The two different methods are graphed together along with the forecasted value for June 2021 in Figure 4.

View Figure

• Download as Powerpoint
• Download Figure

A final attempt was made to use the exponential smoothing method of forecasting. The Solver function in Excel was used to find a value for the smoothing constant (α) that minimized the mean absolute percent error (Figure 5). The Solver function calculated an alpha value for 0.23 using the smoothing constant and calculating for next period using Equation 3. This gave a result of 1,658. \(\begin{equation}\tag{3}{\rm{Forecast\ period}} = {\rm{forecast\ previous\ period}} + \alpha \times \left( {{\rm{current\ period\ value}} - {\rm{forecast\ previous\ period}}} \right)\end{equation}\)

View Figure

• Download as Powerpoint
• Download Figure

Because smoothing methods are not always very helpful in creating trends because they only smooth out fluctuations over time, another method was needed to predict future EV sales. A trend analysis was then run on the previous three years of data. The data was placed into Excel, and using the INTERCEPT() and SLOPE() function, an equation (Equation 4) was calculated for the trend line and graphed in Figure 6. \(\begin{equation}\tag{4}y = 175x + 813\end{equation}\)

View Figure

• Download as Powerpoint
• Download Figure

Equation 4 was then used to find the sales forecast for the years 2022 through 2030. Once these values were found, the next step was to calculate how these sales compared to the overall vehicle registration in Louisiana and use that data to calculate what percentage of new vehicle purchases were EVs. Using National Automobile Dealers Association data for new vehicle purchases, and comparing that to new EV purchases in Louisiana (NADA, 2020), a trend line was calculated in Excel (Figure 7). This trend line was then extended to the year 2030 to create a forecast for the percentage of new EV purchases compared to all new vehicle purchases (Figure 8).

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

Using this forecast trend, by 2030, it is expected that 3.1% of all new vehicle purchases will be EVs. However, looking at the trend line, the slope of the actual value line appears steepest in an upward trend between 2018 and 2019. The actual value point for 2019 also lays higher than the trend line, suggesting that the most recent data might be indicating a steeper trend than this forecast shows. A forecast was then created using only the two most recent data points (Figure 9).

View Figure

• Download as Powerpoint
• Download Figure

Using only the most recent data, the trend for percentage of new vehicle purchases compared to new electric vehicle purchases indicates an expected result of 4.7% of all new vehicle purchases to be EVs.

ANALYZING CURRENT INFRASTRUCTURE IN LOUISIANA

The next step taken to determine the availability of infrastructure for EVs in Louisiana was to analyze how many homes within the state could accommodate at-home electrical charging. While exact statistics on this factor are not readily available, some information can be gleaned from existing sources to compute an estimation. Several factors were considered to determine how many homes in Louisiana could accommodate at-home charging.

The first thing looked at was what percentage of single-family homes exist in Louisiana. A Residential Energy Consumption Survey carried out by the US Energy Information Administration (EIA, 2015) in a report from 2015 shows that 70.1% of people living in the West South-Central census region (which contains the states Texas, Oklahoma, Arkansas, and Louisiana) live in a single-family home. This is similar to other census regions in the south, so while the state of Louisiana could not be isolated, the regional data is a good approximation. Using information from the same survey, respondents reported that of the 70.1% of people who lived in single-family homes, 63.5% of them had access to a garage, with 44.6% of the population having access to a garage for home charging. Another factor considered was the amount of households that owned cars in Louisiana. Using information from the 2019 census data (ACS, 2019), at least 76% of households own at least one vehicle in Louisiana.

For the purposes of this paper, only homes with a garage were considered good candidates for at-home charging. While some homes may be able to connect to home charging without garage access, this option will not be feasible for most users. Using this information, 44.6% of homes in Louisiana have the potential to install at-home charging outlets (Figure 10).

View Figure

• Download as Powerpoint
• Download Figure

At the current rate of growth of EV sales, and considering 44.6% of homes have the potential for at home EV charging, Louisiana should be able to support at home charging, since only 3-5% of new vehicle purchases will be EVs. However, looking at this data alone shows a huge obstacle in reaching the 2030 Biden Administration goal. Without consistent access to at-home charging, there will have to be other resources for EV charging (Traut, Cherng, Hendrickson, & Michalek, 2013).

Since most charging will be done at home in Louisiana, commute distance becomes an important factor. According to Answer Financial, the average commute distance in Louisiana is 44 miles (88 miles round trip) (Leonard, 2018). This is the eighth longest commute in the United States but is still within the range of most EVs on the market. The average EV can travel a real-world distance of 193 miles (Tweedale, 2021). This distance easily allows for a commute to work and back home again on a single charge.

It is also important to consider travel within the state. The location of various publicly available charging stations was found using ChargeHub (ChargeHub, 2021) to see if travel throughout the state was feasible using an EV (Figure 11).

View Figure

• Download as Powerpoint
• Download Figure

The map in Figure 11 shows all the charging stations currently available in Louisiana, with the blue markers indicating Level 2 charging stations and the yellow markers is a Level 3 charging station. The Level 3 charging stations, also referred to as “Rapid Chargers,” can charge an EV in as little as 30 minutes, but for the purpose of this paper, both Level 2 and Level 3 chargers will be considered equally. The information from ChargeHub and Google Maps (Google Maps, 2021) was used to determine distances between major hubs of charger availability. Some sample distances were:

• New Orleans to Alexandria = 190 miles

• Alexandria to Shreveport = 125 miles

• Alexandria to Monroe = 95 miles

• Monroe to Lafayette = 183 miles

• Lafayette to Houma = 102 miles

• New Orleans to Lake Charles = 206 miles

While the New Orleans to Lake Charles distance does fall outside the 193 real world average, one passes through two different EV charging stations hubs (in Baton Rouge and Lafayette) when driving from New Orleans to Lake Charles. Therefore, while the distance of 206 miles may not be possible on a single charge, there is access to other charging stations along the way.

CONCLUSIONS

Electric Vehicle sales in Louisiana lag behind many states in the US. If the sales forecast is correct, the current number of houses that can support EVs and the current infrastructure in place is sufficient enough to support the current market. For the people who choose to own an EV and charge at home, the average Louisiana homeowner should be able to travel to and from work on a single charge. The current infrastructure in Louisiana is also sufficient enough to support long distance travel throughout the state. However, just because past sales in Louisiana have been slow, this does not mean the trend will continue. A new type of rapid charger, more efficient battery, or a more affordable vehicle might disrupt the entire electric car market.

Currently it does not seem likely that the sales of EVs will ever surpass the 44% restraint due to the inability of most people to charge at home. Until more businesses or multifamily housing options offer a way for EVs to be charged outside the single-family home, sales of EVs in Louisiana will remain low.

At the current rate of growth, Louisiana will not meet the goal of 50% of new vehicle purchases by 2030, but for people who do take the plunge and buy an EV, as long as they are able to install a charging station at home, these EV owners should be able to travel to and from work and throughout the state.