Smart Grid for Buildings

As a steward of the environment, the California Energy Commission (CEC) desires to reduce not only electricity consumption but also electricity demand (California Energy Commission, 2013). When peak demand gets too high, the existing primary utility infrastructure is not large enough and in the short term can result in less efficient secondary infrastructure being used or in outages. In the long term, larger infrastructure might be needed which could deplete environmental resources further. Various systems, methods, and technologies exist to help reduce demand. Aided with historical electricity demand data, historical weather data, weather forecasts, and utility infrastructure data, the California Independent System Operator (CAISO) and others can detect seasonal patterns as well as predict high demand periods in near real time. Investor Owned Utilities (IOUs) use time-dependent electricity rate schedules to encourage customers to shift their electricity needs to less impacted time periods. CAISO and the IOUs issue warnings when they predict a certain day to be one of the highest days of demand for the year. They also have demand response (DR) programs in which customers can save money when reducing demand during those days. The technology studied here consists of local hardware and cloud-based software that optimizes the operation of a building or campus for the lowest monthly electricity cost. If the building’s electricity rate schedule rewards customers who reduce their demand and the customer has certain flexibilities, the technology will enable significant reduction or shifting of peak electricity demand. For instance, an ideal customer would have a real time electricity pricing schedule or have the ability to purchase wholesale electricity from multiple providers. An ideal customer would furthermore have production assets like solar, storage assets like batteries or thermal storage tanks, or be otherwise able to shift their grid provided electricity needs. In this case, the technology provider would have more variables to add to their optimization equation and more parameters to control in order to save the customer money and thereby likely reduce demand during expensive peak periods. In this study, the technology was installed in one small commercial building owned and operated by SDG&E. They were interested in evaluating the effectiveness of the product and showcasing it to visitors. The building has a central building management system (BMS) and centralized lighting control that facilitated integration of the technology. The building has solar photovoltaic panels, electrical car charging stations, a storage battery, small scale thermal energy storage, daylighting controls, and is connected to the grid. So, there were ample controllable assets. The main limitations of the study are sample size (only one building and two successful simulations to date) and the fact that the vendor does not have an interest in participating in IOU DR programs or pilots until real time pricing programs are fully developed. Furthermore, the size and constraints of the building limited the demand reduction capabilities of the technology and thus does not represent a typical case. During the most representative simulation, the technology reduced demand and saved utility cost as expected. Figure 1 shows that during the 2014-03-14 simulation, demand was most successfully reduced compared to the calculated customer business-as-usual baseline1 from 13:00 until 18:00. During this 5-hour period, the maximum 1-hour interval power drop was 11.7 kW and the minimum was 8.2 kW. Assuming participation in SDG&E’s Technology Incentives (TI) DR program and using the minimum load drop, the one-time savings would theoretically have been about $2,460. It is worth noting that there is a spike in the 15-minute interval power in the evening from about 19:30 to 20:15 with a maximum of 26.88 kW. While it is important to prevent shifting of demand to evening hours according to CAISO’s DR roadmap (California ISO, 2013), the increase in cost under the SDG&E AL-TOU rate is not significant. During this time slot, only 19:30 to 20:00 in the winter is classified as On-Peak and otherwise the time slot is classified as Semi-Peak. In addition, this peak is not higher than the measured peak for the day so it would not increase monthly demand charges. SDG&E’s TI program requires the customer to participate in one of the following on-going DR programs for three years: Base Interruptible Program (BIP), Capacity Bidding Program (CBP), Critical Peak Pricing – Default (CPP-D), or an eligible pilot. Calculating annual DR savings and simple payback period based on one successful simulation in which load was dropped for 5 hours is not possible. However, theoretical calculations are nonetheless provided to give the reader an order of magnitude idea of the potential cost savings. Assuming participation in CBP, a 100% successful day-of nomination of 8.2 kW/event hour for 22 hours for every eligible month, and an approximated flat kWh incentive of $0.06/kWh, the annual savings would be as shown in Figure 2. In regards to simple payback period, the core project first cost was $121,314. However, according to the vendor this amount was atypically high for various reasons. A similar but more typical project with less complex building systems and located somewhere that real time pricing is available such as New York City would have cost roughly $20,000 - $50,000. For simple payback calculations using all three of these first cost possibilities, the TI incentive, and the “1-4 Hour” CBP Annual Savings for argument’s sake, the simple payback period would be as shown in Figure 3. Several items are worth mentioning with regards to the assumptions. First, the number of actual CBP event hours per month is determined by SDG&E (up to a maximum of 44 hours per product per month) and could either reduce or increase the annual savings. Second, the measured drop occurred in March which is not within the CBP program. Third, potential annual savings could possibly be higher for the other SDG&E DR programs but it was not calculated. Fourth, revenue sharing was deducted from the CBP annual savings per a typical vendor requirement but the percentage used here is an assumption. It is negotiated on a project by project basis between the owner and aggregator. The response time of the initial 1-minute interval load drop of 9.6 kW was 1 minute (from 11:54 AM to 11:55 PM) and the largest load drop of 19.2 kW occurred in 42 minutes (from 11:54 AM to 12:36 PM) as shown in Figure 4. The researchers concluded that the technology functions as specified and that it is easy for facility staff to use. They were disappointed that the vendor does not participate in Investor Owned Utility (IOU) DR programs and that the IOUs do not offer real time pricing yet. Since the IOUs are yet to provide real time pricing, the vendor is reluctant to establish a current presence here. However, real time pricing appears to be on the horizon. When it does appear, this technology could prove to be a valuable DR and cost-savings system for a variety of building types and sizes.