Residential profile is the public profile provided by DTE on their website for residential customers.

Transcription

1 Michigan Public Service Commission DTE Electric Company Analysis of average net metering inflow and outflow Schedule: GG-1 Total Average Demand Total Production needed for Net Production of 1KW panel Size of System to provide full production Import kwh Import Hours Export kwh Export Hours 8,436 kwh 8,436 kwh 1,260 kwh 7 KW 5,185 kwh 5,909 Hours 5,185 kwh 2,851 Hours Total Bi-directional use 10,370 Self Consumption KWh 3,251 kwh Increased Net work use Kwh 1,934 kwh Increased Net work use 22.9% Amount of energy consumed in same hour 38.5% Amount of Energy exported 61.5% Max Export 5.29 Max Demand 2.53 ratio of peak production to demand Source Notes Solar production Detroit City data on the NREL PVWatts All defaults and 1 KW of panels no defaults were changed Residential profile is the public profile provided by DTE on their website for residential customers. Date range is from 3/27/2016 to 3/26/2017 this was when the model was pulled from the external website and modeling work was started.

2 Page 1 of 11 MPSC Question No.: ELPCDE-2.50c Respondent: R. Mueller Refer to Exhibit A-16 Schedule F9, representing the SAC calculation for an average residential customer for c) Confirm that DG energy that is generated and used on site does not enter the Company s distribution grid and is not delivered to the customer by the Company. If deny, please explain in detail. The energy that is consumed on site does not enter the company s network but it can influence the company s equipment. The energy that is consumed on site influences the voltage and current on the company s equipment just by its presence as a basic physics principle. If the locally generated energy was not present, the resistance to energy on the system would be less at the customer site, and as such would flow more freely. The existence of the locally generated energy produces a resistance similar to a back flow on a water pipe when the storage tank in a basement is mostly full. This influence can raise other customer s voltage who share the secondary with the customer with the generation. Because of the manufacturing limits of equipment, even if the customer has a breaker to prevent flow back into the distribution system, they can (and probably are) back flow into the system for 6 to 8 cycles (roughly 0.1 seconds) frequently as motors start and stop at the location or clouds clear, etc. Therefore, even if energy produced and consumed at the site does not enter the distribution system, there are momentary and continuous impacts to powerflow of the distribution system which would not be present if the generation was absent. n/a

3 Page 2 of 11 MPSC Question No.: ELPCDE-2.50h Respondent: R. Mueller Refer to Exhibit A-16 Schedule F9, representing the SAC calculation for an average residential customer for h) Please provide all analyses that the Company performed that demonstrate a relationship between the size of a DG system and the cost of serving a customer with that DG system. The question assumes a straight-line relationship between PV output and cost, which is not appropriate. While there is a relationship, there are many factors that play into the cost. In a similar fashion, as costs to serve are averaged over a class of customers with similar demand, establishing twoway demand costs for distributed generation customers makes sense. Utility best practice is that the load without the generation is planned for and then the generation without the load is planned for. This mitigates any issues with sudden load tripping and the full generation hitting the grid that is inadequate or the generation tripping and the load demanding full support from the grid. In general, a customer with a larger DG system needs a larger transformer and larger conductors just as a customer with a larger load needs a larger service. There may be required changes to relays, fuses and other protective devices as size goes up or there is increased penetration of generation in a local area this becomes more common. They may require capacitor banks and statcoms, again as the size (or number in close proximity) the likelihood increase. Similar comparisons can be drawn for a dozen more classes of equipment and cost categories. n/a

4 Page 3 of 11 MPSC Question No.: ELPCDE-2.50i Respondent: R. Mueller Refer to Exhibit A-16 Schedule F9, representing the SAC calculation for an average residential customer for i) Please provide all analyses that the Company performed on how the relationship described in h) may change based on the underlying usage of the customer, both in terms of energy and demand. Unless the customer s minimum load always (even in vacation mode or shut down mode) exceeds the capacity of the DG system then the customer is probably backfeeding onto the system. Unless the customer has a breaker to prevent backfeed, the system is always at risk of the full capacity of the local generation feeding into the distribution system. Utility best practice is that the distribution system is always sized so that if local generation is lost or local load is lost that the system can handle that load or supply from that system. This prevents the customer who accidently loses their generation from potentially burning out the distribution system from inrush current and also protects the system from potential back feed that is more than the system was designed for. Best practice for the safety and stability of the system is to plan for both the load being absent with full generation and for the generation being absent with full load. This protects everyone. If best practice is not followed, then protective equipment is required at the site including relays and breakers to trip the site anytime there is an issue with either the generation or the load disappearing. n/a

5 Page 4 of 11 MPSC Question No.: ELPCDE-2.51a Respondent: T. W. Lacey/ K. O. Farrell / C. Serna Page: 1 of 2 Suppose there are two customers who have identical usage, peak demand at the time of class and system peaks used in the cost allocation process, and non-coincident peak demand used in the cost allocation process. Suppose the first customer installs a small DG system that reduces their total annual kwh usage by 10% and never exports any energy to the Company s distribution grid. Suppose the second customer installs a more efficient air conditioner and LED lights, and reduces their annual kwh usage by 10%. Further, suppose that the usage, the peak demand of the two customers at the time of class and system peaks used in the cost allocation process, and the non-coincident peak demand of the customers remains identical to each other after these actions. a) In this example, confirm that according to the Company s cost allocation methodology that both customers incur the same cost to serve on the system before and after their actions. If deny, please explain in detail. Not necessarily on a cost allocation basis. Schedule 300 uses the sum of the individual customer max demands. There is a possibility that the customer who installed a small DG system sets their max demand when they are not generating, while the customer who has installed an efficient air conditioner and LED lights is reducing their demand consistently by 10%. In this scenario, the two customers contribution to Allocation Schedule 300 would not be equal. Assuming the only changes to allocation schedules are schedule Power Plant Energy Production (and assuming Schedule 300 as explained above, did not change) then confirmed for cost of service. However, the Company s cost allocation methodology like all cost of service allocation methodologies does not perfectly reflect the true costs driven by each individual customer. DG customers, as described in the example provided, put more stress and thus drive more costs than customers who reduce usage through energy efficiency. The DG customer is still using the same amount of energy (from two sources) so his or her inrush current requirements are the same as before installing DG. Furthermore, the DG customer drives additional costs due to the ramping nature of their generation which changes from minute to minute due to cloud cover passing through. In addition, the Company also needs to maintain backup capacity to serve the DG customer s entire load, which has not changed simply due to the installation of DG and DTE might need

6 Page 5 of 11 MPSC Question No.: ELPCDE-2.51a Respondent: T. W. Lacey/ K. O. Farrell / C. Serna Page: 2 of 2 to serve the entire load with little notice if the distributed generation equipment might be offline for any reason. The customer implementing energy waste reduction tools is actually using less total energy (from one source) than before taking action, so I would expect his or her system requirements to be less, and thus may result in lower system costs than a DG customer. n/a

7 Page 6 of 11 MPSC Question No.: ELPCDE-2.51b Respondent: C. Serna Suppose there are two customers who have identical usage, peak demand at the time of class and system peaks used in the cost allocation process, and non-coincident peak demand used in the cost allocation process. Suppose the first customer installs a small DG system that reduces their total annual kwh usage by 10% and never exports any energy to the Company s distribution grid. Suppose the second customer installs a more efficient air conditioner and LED lights, and reduces their annual kwh usage by 10%. Further, suppose that the usage, the peak demand of the two customers at the time of class and system peaks used in the cost allocation process, and the non-coincident peak demand of the customers remains identical to each other after these actions. b) Please explain why is it appropriate to charge the first customer a SAC to replace lost distribution revenue but it is not appropriate to charge the second customer a SAC to replace the same lost distribution revenue. In general DG customers reduction in energy provided by the utility is greater than a customer installing energy efficient lights and appliances. There are differences between the two customers in the example provided above. Unlike a customer with a small DG system, the customer who is trying to become more energy efficient does not rely on the grid to export power, their load reduction is more consistent over the day (less variability), and even though DG systems are producing, DG customers continue to rely on the grid for intra-minute increments when appliances turn on and require more voltage than can be supported by their DG systems. See also, direct testimony of C. Serna, page 61, line 3, through page 62, line 2. n/a

8 Page 7 of 11 MPSC Question No.: ELPCDE-2.72a Respondent: R. Mueller Refer to Serna Direct at 53. a) Please explain why the impact of changes in inverter-based generation output would affect legacy protective equipment differently from other large, rapid changes in household usage (such as turning an oven on or off, having an electric water heater cycle on or off, or having an air conditioning compressor start and stop). Water heaters and air conditioners tend to cycle on the order of minutes to hours. The compressor on an air conditioner will tend to cycle on, run the temperature down in a building by 2 to 4 degrees (depending on specific equipment settings) and then turn off. It is not unusual for an air conditioner to cycle only a couple of times an hour. Water heaters are similar, raising the temperature of the water in its tank several degrees each time it operates. Similar multi-minute cycles exist in most household equipment. Inverter based PV has a tendency to change output by 40% or more in less than 1/60 th of second. This rapid change of output can impact the life of distribution equipment. In the worst situation, the inverter based resource is exporting to the distribution system at full output and the load is pulling power from the distribution system at the lowest output, cycling the direction of power flow through the transformer and potentially other nearby equipment hundreds to thousands of times an hour, on equipment that was designed to support 1 way power flow. Clean Power Research has done a number of studies over time that show the variation in power. While this variation in power has effects on distribution equipment, the variability in voltage has a bigger impact on customer s equipment that shares the same secondary as was documented in a series of mid-1990s IEEE papers particularly motors like those found in air conditioners, furnace fans, and refrigerators. N/A

9 Page 8 of 11 Refer to Serna Direct at 53. MPSC Question No.: ELPCDE-2.72b Respondent: R. Mueller b) Generally, what types of equipment (e.g. meters, transformers, conduits, etc) on the Company s system are able to accommodate reverse power flows without any modifications, and what types of equipment on the Company s system are not able to accommodate reverse power flows without any modifications? The Company has a large range of equipment that ranges in age from more than 50 years to less than a year. The equipment installed more than 2 years ago was not procured or engineered to support two-way power flow, nor was it engineered into the grid to support two way power flow. With the level of PV penetration in the grid it was not prudent to ask for more money to spend the engineering time or effort to find and test equipment for which no standards existed (none of the industry standards had specifications on what needed to be tested for two-way power flow until recently) and then to train engineers on how to design for two-way power flow. The Company took care to join the EPRI groups and projects that were leading edge research for two-way power flow design and has encouraged new engineers to learn about issues with two-way power flow. The company has under taken an effort to liaison with companies with more PV and understand what they have changed in equipment standards, training and implementation. The Company learned a number of lessons from benchmarking Southern California Edison on what needed to be done to relays and load tap changers. We learned from Hawaiian Electric about mechanical voltage regulators and additional relay impacts, we learned from Sacramento Municipal Utility District about issues with some types of transformers and other equipment. In each case over the last two years the Company has worked to update internal standards and procurement to support multidirectional power flow. At an affordable rate of upgrades it will be decades before the Company s grid has been updated to fully support two way power flow. The Company is still learning from others with more two-way power flow than the company has today, about issues that are impending as penetration levels increase. N/A

10 Page 9 of 11 MPSC Question No.: ELPCDE-5.98a Respondent: R. J. Mueller Refer to the Company s response to ELPCDE 2.50c. a) Suppose a house with a DG system is simultaneously producing 3 kw of power and simultaneously consuming 3.5 kw of power, so that its instantaneous net draw on the Company s system is 0.5 kw. Suppose another house without a DG system has an instantaneous net usage of 0.5 kw. Would the instantaneous impact on voltage, current, and resistance to energy on customers who share the secondary with these customers be the same or different? Please explain in detail. The net draw would be similar, but the inverters have proven to have harmonics issues, that outweigh the harmonics issues of similar sized residential UL approved equipment. The second item is based on studies by Sandia, there is a high probability that in Michigan on partly cloudy to cloudy days (roughly 300 of 365 according to NOAA) variations in the production on a 30 second or shorter basis would be frequent, something that similar sized residential appliances don t typically do. The issue is not power flow on average, but instantaneous power flow changes and the frequency there of, and the contribution of harmonics to the system. Longitudinal studies are underway in California to quantify the loss of life to residential appliances that meet UL standards based on these kinds of issues. While DTE is not part of these studies, they are expected to report out in 2022.

11 Page 10 of 11 MPSC Question No.: ELPCDE-5.103b Respondent: C. Serna/ R. Mueller Refer to the Company s response to ELPCDE 2.51b. b) Confirm that feeders that may serve dozens or hundreds of customers experience intra- minute variations in load due to the variation in usage of individual customers, even if none of these customers have DG systems. If deny, please explain. Feeders typically serve many customers each with inter-minute variations in load. Distributed generation customers are the only customers who, in addition to exhibiting variations in consumptive load, outflow energy into the distribution system. This electrical outflow is a fundamentally distinct variation in load from the load characteristics of a non-dg customer. In general, see Serna Direct at 52.

12 Page 11 of 11 MPSC Question No.: ELPCDE-5.107a Respondent: R. J. Mueller Refer to the Company s response to ELPCDE 2.73c. a) Are transformers and secondary lines designed to handle the sum of the absolute peaks of each individual customer served by that equipment, regardless of when those peaks occurred, or are they designed to handle the diversified demand of the aggregated load of the customers served by that equipment? The electrical system has been designed for diversified aggregated load. Solar PV DG systems do not produce energy that matches the profile of diversified load. Solar PV production magnitude is directly tied to availability of sunlight. As per response to ELPCDE-5.99a, DG systems will produce significantly more energy than diversified loads would be expected to be produce over certain hours. As penetration of Solar PV increases, the net powerflow during certain hours becomes less diverse as it is dominated by reverse powerflow following the solar production curve.