Impact of Large-Scale PV Integration in Southern Nevada – Case Study [1]

Y. Baghzouz

UNLV

[1] http://www.pnnl.gov/main/publications/external/technical_reports/PNNL-20677.pdf

Project Scope • The research effort evaluates the impact of large-

scale photovoltaic (PV) and distributed generation (DG*) output on NV Energy’s electric grid system in southern Nevada. – It analyses the ability of NV Energy’s generation to

accommodate increasing amounts of utility-scale PV and DG, and the resulting cost of integrating renewable resources.

– It quantifies the impact of variable PV generation output on NV Energy’s system operations, including balancing reserve requirements, and the ability of the existing generation fleet to accommodate increasing amounts of large-scale PV and DG.

(*) DG is defined as PV rated below 5 MW, connected to distribution lines below 25 kV.

Project Tasks

• Task 1: Estimate large-scale PV output and DG output for southern Nevada on a one-minute time scale for an entire year.

• Task 2: Develop composite PV output profiles via statistical methods for ten sites and five case study scenarios in southern Nevada.

• Task 3: Analyze the impact of variable PV & DG on system balancing reserve requirements within NV Energy’s southern Nevada system, and evaluate the generation fleet’s capability to meet the requirements.

• Task 4: Identify the impacts on generation dispatch and costs of integrating PV & DG, and identify options to increase the amount of solar resources that can be installed on NV Energy’s system

Large-Scale PV Systems

• The large-scale renewable projects include those that have signed purchase power agreements (PPAs), but not yet in commercial operation, as well as several projects that exhibit potential to be secured under future PPAs.

• Large-scale plant designs in size from 5 to 300 MWAC and included both fixed-tilt thin-film and single axis- tracking polycrystalline Si systems. Distributed PV systems range from 4 kWAC to 3 MWAC and comprise both roof-mounted and ground-mounted systems.

• At the higher penetration levels, PV and DG could supply up to 20 percent of NV Energy’s load at peak and up to 50 percent or more during off-peak hours.

Proposed large PV installations

Cases Studied:

Scenarios/Cases Studied

Solar irradiance and PV Power

Ground-based irradiance data is not available at any of the ten proposed PV plant sites. A method(*) was developed by SNL to simulate time-synchronized minute-by-minute irradiance at each of the proposed PV sites using hourly estimates of irradiance from geostationary satellite imagery, and min-to-min ground irradiance measurements at six LVVWD sites.

http://www.geos.noaa.gov Map of LVVWD sites with 1 min. irradiance

measurements.

(*)Simulation of One-Minute Power Output from Utility-Scale Photovoltaic Generation Systems, C. Hansen, J. Stein,A. Ellis, Sandia National Laboratories, August 2011

Study Assumptions The study assesses the ability of the system as it existed during the study (2011), except

that load and weather data from 2007 are used. Generation resources reflect a 2011 mix and operating costs.

External sales were set to zero – an important assumption as high PV & DG penetration could result in additional committed generation to avoid dump energy caused by units operating at minimum load.

System upgrades or mitigation is required when generating operating limits are exceeded or when NERC performance standards (under CPS-1 & CPS-2) cannot be met.

Error forecasts are ignored, i.e., load following and regulation requirements are determined by only the variability with load and solar, and not be affected by forecast errors.

Real-time load and solar forecasts are determined from the 10-minute clock average of the time series, which sets the target for real-time dispatch/load following.

Regulation and Load Following

• Since a perfect real-time forecast is assumed, the regulation and load following are calculated by the following formulas (where net load = load demand – PV power):

– Regulation = [actual net load] – [10 minute average of net load with ramps]

– Load Following = [10 minute average of net load with ramps] – [60 minute average of net load with ramps]

• NV Energy Reserves:

– Contingency reserves: 425 MW, of which 50% is spinning and 50% is non-spinning.

– Regulation reserves: 25 MW, of which 100% is fully spinning.

PV Output Profiles • One year of AC power output, at one-minute intervals, were

produced for each PV plant. PV output profiles for a typical “clear” day and for a “cloudy” day under a high PV penetration (i.e., Case 5) are illustrated below.

• Results indicate that the level of variability as a percentage of total installed capacity becomes smaller when PV plants are installed over a large area.

Intermittency in irradiance can vary significantly among individual sites, i.e.,

cloud cover impacts some, but not all PV sites simultaneously.

Sample of PV Output Power

The figure below illustrates one “cloudy” week (i.e., July 20, 2007 to July 26, 2007) of aggregate power for Case 4 (492MW).

Study Results • integration of large-scale PV increases regulation and load following

requirements that must be supplied by NV Energy’s generating resources, mostly from combined cycle and combustion turbine units.

• These higher requirements increase the amount of generation committed in day-ahead schedules, degrade unit efficiency, and accelerate operations and maintenance, all of which increase energy costs. – On average, 1 MW of additional thermal generating capacity must be

reserved for regulation for each 25 MW of PV capacity installed in NV Energy’s southern Nevada system.

– On average, 1 MW/min of additional ramping capability must be reserved for regulation for each 75 MW of PV in NV Energy’s southern system.

Study Results (cont.)

– At the highest PV and DG penetration levels, the system heat rate degrades by about 5% compared to the base case where PV & DG capacity is zero.

– Integration costs caused by increasing operating reserves and modest curtailment of PV output range from a low of $3/MWh for low renewable penetration to a high of just under $8/MWh for the higher penetration cases.

Recommended Measures The results indicate that measures must be taken to mitigate balancing area ramp deficiencies to integrate PV and DG. These measures include:

– increasing thermal generation operating reserve margins and committing less efficient combustion turbine generators to meet the increased flexibility in thermal generation output caused by the integration of PV and DG output.

– The level of mitigation should include sufficient margin to account for other potential impacts and at a reasonable level of risk.

– In the future, other mitigation measures including the installation of fast-response energy storage systems, such as hydroelectric storage (Hoover ramping capability*) and batteries may offer suitable solutions.

(*) NV Energy’s Hoover hydro allocation currently is used as a peak shaving resource. Use of Hoover for frequency regulation would reduce the Company’s ability to schedule capacity during higher cost peak hours.

Regulation capacity requirements – Winter - Case 1

The additional regulation capacity requirements caused by PV in the winter are relatively larger than in the summer which could be explained by the more cloudy weather in the winter in the southern Nevada area.

Regulation Capacity Requirements with increase in large scale PV Penetration : Winter – Cases 1A-5A

As large-scale PV increases from 149 MW to 892 MW, the total system regulation requirements increase significantly, especially in the winter season.

Regulation Ramp Requirements – Winter - Case 1

The ramp requirements for regulation only increase slightly in the daytime hours for Case 1.

Regulation Ramp Requirements – Winter - Case 1A-5A

It can be observed that in Case 5A, regulation ramp requirements are almost double the requirements for the system without solar, which means the requirements caused by PV become dominant.

Load Following Capacity Requirements – Summer – Case 1

The fact that PV reduces load following capacity requirements in some hours means that the direction of 10-minute PV trends during these hours coincide with the 10-minute trends of the load itself.

Load Following Capacity Requirements with increase in large scale PV Penetration : winter – Cases 1A-5A

Through all daylight hours, the requirement increases with solar penetration, but at a different rate in each hour. Overall, however, the increase of load following capacity requirements caused by PV is much less significant than that of regulation.

Load Following Ramp Requirements – Summer – Case 1

In the summer months, PV can reduce the load following ramp requirements when the 10-minute PV generation trend coincides with the load trend. In other hours, PV barely causes any increase in the ramp requirements.

Load Following Ramp Requirements with increase in large scale PV Penetration : Summer – Cases 1A-5A

Overall, the increase of load following ramp requirements caused by PV is much less significant than that of regulation.

Regulation Capacity and Ramp: winter – Case 5

Load Following Capacity and Ramp: winter – Case 5

Sample of Ramping Capability Adequacy Results

Results of maximum ramp deficiency from the ramping analysis are presented in the figure below for Case 5A: The blue and red bars are the hours in which ramping deficiency has been identified. The maximum ramp deficiency reaches 20MW/min in the early afternoon hours.

Generation Re-dispatch to Resolve Ramping Deficiency

To mitigate the ramping deficiency, one of NV Energy’s 50 MW combustion turbines/peaking units at Clark Station was committed in the hours with ramping deficiency, and the combined cycle units on AGC were backed down to provide additional regulation ramping capability. This addition was sufficient to reduce ramping deficiencies to base case levels.

(1) Trend of Yearly Regulation Direction Changes and Mileage as Function of PV Penetration Level

An increase of 1 MW in PV installed capacity can approximately cause 7.9 more direction changes and 170 MW of mileage increase (both up and down) in regulation process.

(2) Balancing Service Ramp Statistics

More solar generation can cause higher frequency of regulation movements.

Case 0

Case 5A

Improved Operations with High Penetration of Renewables

• Larger balancing area

• Shorter scheduling intervals

• Committing closer to real time

• Demand response, storage and smart grid may help

END!