What’s Happening with EV Sales?

forecasting1With gas prices hovering just above $2 a gallon, there is less excitement and curiosity about electric vehicles (EV) as there has been in the past.  In fact, 2015 was the worst year for EV sales; total EV sale were down 5 percent in 2015 compared to 2014, with only 116,099 vehicles sold compared to 122,438.  But, there may still be hope for the EV market as 2016 is shaping up to be a banner year, with 93,197 vehicles sold through August, compared to 72,270 YTD August 2015 .

What forces drive EV sales?  Without question gasoline prices influence EV sales, with the possible exception of one particular EV model.  Presumably, consumers are not purchasing a $90,000 Tesla Model S with the goal of saving money on their annual gasoline bill.  The U.S. average price for a gallon of Regular gasoline has not been above $3.00 since October 2014 and has not been above $3.50 since July of that year.

With monthly data available on both EV sales and gas prices, it is possible to calculate a correlation between the two.  Analyzing data from January 2012 to August 2016, I found a correlation coefficient of -0.46, which is not as strong as one might want for forecasting purposes, but it is a relationship none the less.

So the question is: Do EVs make economical sense right now with gas prices as low as they are?  The answer depends on which models you compare.  Generally, there does not appear to be much incentive to trade in the gas-guzzler for an EV.  EVs are almost always more expensive than their gas counterparts, but upfront costs can be recouped with lower annual operating costs.

I attempted to compare 2 similar vehicles: the VW Golf and VW e-Golf.  For this comparison, I assumed 15,000 annual miles driven, gas price of $2.25 a gallon, and electricity rate of $0.15 kWh.  Based on these assumptions, it will cost $653/year to drive 15,000 miles in the e-Golf and $1,164 in the gas powered Golf (based on published vehicle efficiency ratings, www.fueleconomy.gov).  The e-Golf is significantly cheaper to operate, but this comes at a cost; the average e-Golf retails for $32,295 compared to $22,960 for the traditional gas powered model.  Even after factoring in a federal tax credit of $7,500, the e-Golf is still $1,800 more.  If you were to finance both vehicles over 5 years, at 3 percent, add in electric costs for the e-Golf and gasoline for the Golf, both vehicles would cost roughly the same to operate.  The story changes if gas prices increase; at $3.00/gallon, the e-Golf would save you almost $2,000 over the 5-year period.

Although the EV market is still in its infancy in most areas, EVs may in a few years have the impact that solar does today and that is why tracking the EV market is important to the electricity industry, as evidenced from Itron’s most recent survey showing more and more utilities incorporating EVs into their forecasts.

 

1“Monthly Plug-in Sales Scorecard”, Inside EVs. http://insideevs.com/monthly-plug-in-sales-scorecard/.


Who is Forecasting Long-Term Solar Generation?

In this last forecasting brown bag presentation on solar load forecasting, we asked participants who had developed a long-term solar load forecast before 2013 and after 2015. As expected, very few had done a forecast before 2013 and majority put together something after 2015. During the last Vermont state forecast we did in 2014, solar wasn’t even a major topic until, of course, the month before the forecast was due. But what is a reasonable approach?

We started by collecting monthly data on installed systems and number of customers for each state starting in 2010. Then we compared saturation rates – what we found is that those states with the highest return on investment had the highest level of saturation. People make rational economic decisions after all! Well, at least some people do. Armed with this information, we estimated a regression model for Vermont that relates system saturation to system economics using a simple payback to capture system economics. And guess what? It worked. We were pleasantly surprised; when we used a cubic specification the model fit was awesome. We have used this model in several service areas – some with high saturation-levels (Nevada) and some with very low saturation (Indiana) and it seems to work, most of the time. This model approach was laid out in the brown bag presentation.

If you google “Forecasting New Technologies” you will find dozens of approaches. Most of these entail fitting an S-shaped curve to your own or like technology data set. If you have tried a Bass Diffusion model or Fisher-Pry logistic curve fit model or something else, we would love to hear about it. We all need to forecast solar generation – let’s share approaches!


Accelerating the Integration of Variable Generation into the Electric Grid

Recently, I was invited to speak at the Utility Variable Generation Integration Group (UVIG) Conference, which was focused on improving the integration of variable generation into the electric power system. UVIG was established in 1989 and currently has more than 160 members from the U.S., Canada, Europe, Asia and New Zealand. The conference was well-attended with participants from Europe, North America and Asia.

This year’s conference topics included a tutorial on stochastic forecasting methods as they apply to quantifying wind and solar generation forecast uncertainty, how to run a solar and wind forecasting trial, integrating stochastic generation forecasts into an EMS system, and industry trends with distributed solar photovoltaic (PV) forecasting.

During the conference, I participated on a panel that discussed ways to incorporate distributed PV into a load forecast. My co-presenters, Jim Blatchford of the California ISO and Dr. Tom Hoff of Clean Power Research, and I discussed how we incorporate the PV forecasts that Clean Power Research develops into the real-time load forecast model Itron developed for the California ISO. If you would like to find out more about the how we do this, feel free to reach me at frank.monforte@itron.com.

If you are interested in learning more about wind and solar generation forecasting, you should consider attending a future UVIG conference.


Improving Net Load

Renewable Portfolio Standard requirements and decreasing costs of photovoltaics (PV) are resulting in significant amounts of solar PV systems being installed in California. Current load forecasting methods often fail to address the growing uncertainty in net load (demand for power less behind-the-meter solar generation) forecasts.

Itron is excited to work with the California Energy Commission, Clean Power Research, the California ISO and utility partners Sacramento Municipal Utility District, Southern California Edison and Pacific Gas & Electric Company on a new project to help reduce net load forecast uncertainty by producing high-accuracy solar generation forecasts and linking them to net load forecasts. Project goals and objectives will be met by applying research in the California ISO (CAISO) and electric utility operations.

Clean Power Research and Itron already supply the CAISO with solar forecasts and net load forecasts separately. Under this research effort, improved forecast methodologies will be developed and implemented starting in mid-2015 with full implementation by mid-to-late 2017.

Success in this area is anticipated to result in reductions in regulation service costs of more than $10 million per year by 2020. Additionally, these improvements will lower regulation service needs by over 2 million MWh per year by 2020, avoiding the emission of up to 2.7 million tons of greenhouse gases (GHG) per year.

Read Clean Power Research’s blog post to learn more about the project. To find out more about Itron’s role in improving Net Load Forecasts please contact Dr. Frank A. Monforte at frank.monforte@itron.com.


The Buzz about Solar Forecasting

Installed solar generation capacity is growing worldwide and this movement’s effect on load forecasts is significant. Energy service providers and electricity market operators are striving to understand how solar generation impacts their short- and long-term load forecasts. Itron is at the forefront of developing statistical modeling approaches to address this problem. I’ve written a white paper titled “Forecast Practitioner’s Handbook: Incorporating the Impact of Embedded Solar Generation into a Short-term Load-Forecasting Model” which describes a statistical modeling framework to incorporate the load impact of embedded solar generation.

When we talk about solar generation, we must differentiate utility solar installations, where the electricity generated feeds directly into the grid, from non-utility installations (also referred to as embedded solar generation, i.e., rooftop solar), where generation offsets on-site consumption. Both pose unique forecasting challenges. Utility solar installations impact the measurements of net load, which is defined as load minus utility solar generation. In this case, accurate forecasts of utility solar generation are required to forecast net load. Embedded generation directly impacts measurements of load since this generation occurs behind the meter. As a result, embedded generation impacts how we model load. The joint impact of utility and non-utility solar installations is increased volatility of net load. This, in turn, has added complexity to near-term forecasting of ramping regulation requirements.

There are a number of initiatives underway – the Department of Energy’s Sunshot Initiative being the most active - that focus on utility solar generation forecasting. The primary focus of these initiatives is developing tools that provide accurate utility solar generation forecasts. Further, these initiatives have been applied to the area of forecasting embedded solar generation. Clearly, these initiatives are delivering high value to the industry. Unfortunately, the impact of embedded solar generation on loads and consequently on load forecasting has received little to no attention. The purpose of Dr. Monforte’s white paper is to provide guidance on how to incorporate the impact of embedded solar generation in a load forecast.

As you think about forecasting the impact of embedded solar generation into your forecast, you need to make assumptions about:

(a) solar insolation, which is how much sunlight hits the panels on any given day of the year and time of day,
(b) the average operating efficiency of the solar panel population,
(c) average cloud cover, and if you are generating a long-term forecast,
(d) the growth of the embedded solar generation in your service area.

The white paper breaks these pieces down into manageable tasks and begins with an example that illustrates the impact embedded solar generation can have on existing load forecasting models. This is followed by an overview of the language of solar generation and a presentation of practical steps to develop engineering-based explanatory variables that capture the load impact of embedded solar generation. The modeling constructs presented in the paper can be used in both short- and long-term load forecast models.


My white paper can be downloaded from Itron’s website. Click here to download.

http://energy.gov/eere/sunshot/sunshot-initiative


The Solar Lease Deal

Working in the energy industry puts a big target on your chest when it comes to solar leasing deals. In the span of weeks, I have had several friends approach me asking whether they should sign up for a solar panel lease deal from a local company.

With this new assignment, I sharpened my pencil (really a spreadsheet), collected their last 12 months of bills, and calculated their electric costs with and without the solar deal. It’s important to understand that each person is different as shown by their historic energy consumption. For my friends, the cost savings from leased solar ranged between $20 and $200 per year under current rate conditions. The trade-off for the immediate savings (and no upfront costs) is the 20-year contract agreeing to purchase the solar power output at a fixed rate escalating with inflation.

Solar House-01In the few cases I reviewed, the savings weren’t great. Apparently my friends don’t use much air conditioning in our mild coastal climate. But, if you double their energy consumption or add an electric vehicle, the savings would more than double due to the quirkiness of an increasing tiered rate structure.

For the analytical mind, calculating savings is a straight-forward exercise in math once you understand the electric rate structure. But, what captured my attention was why my friends with such low energy consumption were being targeted by solar lease companies? Are the solar lease companies making such large margins on their sales that they profit by saving customers $20 to $200 per year?

In the following weeks, I was also contacted by a company offering me a solar lease program. Like my friends, my analysis showed low savings (I also don’t have air conditioning). Then it hit me. We are being offered the solar deals because our zip codes imply that we are credit worthy enough to sign the 20-year lease.

There are many factors that we should consider when forecasting the penetration of rooftop solar. We should examine payback periods, local rates, and distribution issues. We should also add one more factor into the equation, we should consider which of our customers have a big target on their chests called credit worthiness.


DON'T MISS A POST!
I agree to have my personal information transfered to AWeber ( more information )
Opt in to receive notifications when a blog post is published. Don't miss the thought leadership, insight and news from Itron.
We hate spam. Your email address will not be sold or shared with anyone else.