Can Solar Help Me Avoid Demand Charges?

In this blog post, we are going to take a closer look at demand charges and the impact that an on-site solar PV system may have upon them. While many people simply assume that their measured demand will be reduced by the average output of the PV system, they are often disappointed to find that this is not the case. The amount by which solar PV systems reduce demand charges are influenced by a variety of factors and we are going to explore several of them in this post where we will take a close look at a customer behind Southern California Edison (SCE). 

As a quick refresher, SCE determines demand, in kW, based on the maximum 15-minute rate of consumption. For rate TOU-8, demand is measured on a 24x7 basis to determine the facilities demand charge and separately measured during the mid-peak and on-peak periods to determine the demand charges for these respective periods. The on-peak period only exists during the summer months of June - Sept. The graphic below shows a screenshot of an SCE invoice and identifies each demand charge. 

As you can see, the on-peak demand charge during this billing cycle is significant at $22.99/kW and even a modest reduction in on-peak demand will yield significant savings. Since the SCE on-peak period is defined as noon - 6pm on summer weekdays, you'd think that solar PV would be a sure thing to reduce this charge. Sometimes it is.......and sometimes it isn't as we'll see below. The first graph below shows a month where the solar PV system performed consistently and reduced the on-peak demand by about 500 kW from what it otherwise would have been. Given current on-peak SCE demand charges of $25.33/kW (June 1 2014 tariff), this reduction of 500 kW would yield a savings of $12,665 for this billing cycle. Note how this load profile is classic "Duck Curve" and how easy it is to distinguish between weekdays and weekends. 

The X axis shows hours 1 - 24 and the Y axis shows hourly kWh usage. Each line depicts the hourly load for a given day of the month and the legend to the right shows the line colors associated with each day.

The X axis shows hours 1 - 24 and the Y axis shows hourly kWh usage. Each line depicts the hourly load for a given day of the month and the legend to the right shows the line colors associated with each day.

The graph below shows July '13 where a couple of cloudy afternoons ruined the demand charge savings for the month. You can see a handful of days where the load rebounds between 3 - 4pm to about 2,900 kW when solar output is diminished due to cloud cover. Although there are still some savings relative to where the load would be without solar PV, the savings in demand are significantly less than the average output of the PV system.

The X axis shows hours 1 - 24 and the Y axis shows hourly kWh usage. Each line depicts the hourly load for a given day of the month and the legend to the right shows the line colors associated with each day.

The X axis shows hours 1 - 24 and the Y axis shows hourly kWh usage. Each line depicts the hourly load for a given day of the month and the legend to the right shows the line colors associated with each day.

As this example demonstrates, even when a utility has an on-peak period that coincides with solar PV output, demand charge savings will still exhibit a high degree of variability from month to month. As a result, savings from demand should be treated as extra gravy, but the basis for any financial analysis around behind the meter PV has to be driven by avoided kWh charges. These are more certain (presuming your solar PV system works properly) while reductions in demand can be fickle.

If you have questions about this stuff, call us.

A Closer Look at Demand Charges on NGrid-RI's G-32 Rate

At Energy Tariff Experts, (ETE) we've found that demand charges are a source of confusion for many energy professionals and end users. The NGrid-RI (née Narragansett Electric) medium commercial rate, G-32, provides a great example of the nuances that one might encounter when reviewing billed demand charges. The image below is a dissected excerpt from page 2 of the G-32 invoice. The bill date is from early 2013 and although there has been a modest rate change since, the rate structure and bill format have not changed. 

RI G-32 invoice.png

The distribution billings have been divided into the following groups: fixed charges (blue box); volumetric charges billed by kWh (purple box); and demand charges billed in either kW or kVA (red box). Note that the billed demand value is 255.2 kVA which does not match either the kW or kVA measured demand for the billing period. The measured demand for the billing period is shown on the middle right section of the bill, near the callout box that calculates the Power Factor for the Peak period of the billing cycle. In order to determine why the billed value for demand does not match the measured value for demand, one must go to the Tariff to see how the demand charges are calculated. The NGrid-RI rate tariff can be found here (see link on right to open PDF). Below is the relevant excerpt from the tariff that explains how the billed demand is determined.

G-32 determination of demand.png

You'll note that demand is calculated as the greater of four options. End users with a Power Factor less than 90% will be subject to the provisions of item b. In the case of the end user whose bill is shown above, their billed demand was determined in accordance with item c. In July of 2012, this end user had a billed demand of 340 kVA and their Power Factor was well below 90%. Therefore, per the provisions of item b,  the billed demand value of 340 kVA is inclusive of the Power Factor penalty for Power Factors less than 90%. In the bill shown above, you'll note that 255.2 kVA is equal to 75% of 340 kVA.  Due to the provisions of item c, this end user's peaky load, exacerbated by a poor Power Factor during their peak demand month of July 2012, has resulted in an elevated demand charge that has carried forward for multiple billing cycles.

There is another important quirk in the way demand is billed on this rate. The Transmission Demand Charge line item reflects billings for the full 255.2 kVA while the Distribution Demand Charge line item only reflects billings for 55.2 kVA. In essence, demand is billed on a step-wise basis on the G-32 rate. For all billed demand below 200 kW/kVA, only the Transmission Demand Charge applies.  As a result, the marginal cost for each unit of demand below 200 kW/kVA is $2.92 (on this bill, the rate has subsequently increased). Once billed demand crosses the 200 kW/kVA threshold, the marginal cost increases because the Distribution Demand Charge then applies to all incremental demand over 200 kW/kVA. On this bill, the marginal cost of demand over 200 kW/kVA rises to ($3.40 + $2.92 = $6.32) (both rates have subsequently increased).

The step-wise nature of the incremental costs associated with demand charges on the G-32 rate has implications for an energy efficiency strategy. Eliminating demand over 200 kW/kVA and improving Power Factor will return a savings of $6.32 kW/kVA. These savings may be larger if the load is peaky and the provisions of item c in the demand charge calculation apply. Once demand is brought below 200 kW/kVA, the marginal cost of a unit of demand drops by approximately half, and the economics of eliminating demand are not as good. If an end user can accomplish a demand reduction to the 200 kW/kVA level, they will then see a better return on investment by focusing on energy efficiency measures that reduce kWh usage.

If you are trying to devise a high ROI energy efficiency strategy and are trying to gain a better understanding of demand charges, call us. We can probably help you.