Control system considerations for forward-looking owners.
By: Adam Baker
I’ve compiled a list of every single decision I work through in the process of designing how a control system for a solar power plant should work. Some of these decisions, our customers never think about. Some come up from time to time based on the interconnection agreement (IA) and site location. Some decisions don’t seem important now, but allow owners to embrace and prepare for technology we know is coming (rather than try to cobble together a retrofit later).
I’ll start with the most obvious considerations, and progress to the more detailed ones.
As controls for active power and reactive power are almost completely independent, I’ll list considerations for each mode separately, starting with active power.
Active Power Mode
1. Power Purchase Agreement (PPA)
This legal document lists what you’ll be paid per kWh for energy that comes from the site.
2. Interconnection Agreement (IA)
This legal document defines the rules on how much power you can deliver, voltage, voltage conditions, and other site operation parameters such as plant MW setpoint and number of inverters.
Now that we know how many MW we want to make and # of inverters, it’s simple division, right? Some owners believe they’re ready to design a controls system with just these two documents! Not so fast. A (decent) control system is much more complicated.
3. Plant Capacity
Is plant capacity the same as the maximum active power output of all the inverters in aggregate? Early in my career, most sites I worked on had more AC capacity than the interconnection allowed export of power. A 100MW site with 120MW of inverters, for example. This allowed us to turn up output to compensate for clouds, but also meant we had to curtail the inverter output almost all the time to comply with our IA. If plant capacity is equal to what your inverters can produce, you have a relatively simple (for now) control system.
4. Ramp Rate
The IA often spells out ramp rate rules. 10% per minute of plant nameplate capacity is the norm. This means if you have a 100MW site, it should ramp at 10MW per minute for active power, and when doing a controlled shutdown of the plant. When clouds come, or during periods of impeded irradiance, you can’t always control how fast the site ramps down. The control engineer needs to identify ramp rate rules, and if the owner wants the ability to ramp faster.
5. Frequency Droop
IEEE rules require inverters to turn down output in over-frequency conditions. But there may be a good reason you’d want frequency droop handled by the plant control system. Perhaps you have an over capacity of AC and want to turn up the plant’s total output. Perhaps you wish to use your battery storage source to respond to underfreqency conditions.
Choose your frequency droop path (inverter vs. controls system) wisely, because it’s extremely difficult to change after the fact.
6. Command Sources
Command sources might come from the local HMI, an energy management system, or feed from the balancing authority. This means you may have multiple active power setpoints. Your plant control system should be configured to choose (by priority) which to follow, and when.
7. Feedback Source
Closed loop control requires a feedback source. Your plant control system needs a fallback if there is loss in metering telemetry. When the primary high side meter isn’t available, is there a secondary meter configurable to calculate high side output? If you have no meter feedback, can you aggregate all inverter information, add kW and kVARs, and calculate output?
8. Inverters Available vs. Unavailable
If an inverter isn’t communicating with the control system, you can’t control it. Unavailable inverters mean your ramping and active power parameters change based on what you can and can’t control.
If you’re trying to ramp up or change the site setpoint, the amount of ramping you need to do with other inverters is larger than your normal 10% per minute. For example, if half the inverters aren’t running, you’d need to ramp the rest at 20% per minute, giving you an overall site ramp rate of 10% per minute.
9. Uncontrolled Source of Power
If I know 20MW of inverters are running, but I can’t control them, I have to redo all my calculations for ramp rate and target setpoint. This uncontrolled source of power will contribute whether I want it to or not at the point of interconnection.
That uncontrolled source of power is subject to the same variability from clouds as the controllable part of my site. This means if I have 20MW now, I might have 25MW in 5 minutes, or 10MW. It’s important for a plant control system to make decisions based on the actual output of the site, including both controlled and uncontrolled components.
10. Transitioning Inverters
Most sites have a requirement that inverters wait 300 seconds between the time they’re commanded to start and the time they actually start. This means if the control system commands an inverter to start, I won’t be able to get power from that inverter in the short term (up to 5 minutes).
When that 5-minute timer expires, that inverter will jump to the command it’s been given. That’s why I like to keep track of an “in transition” mode so I can preemptively ramp down the site knowing that additional inverters will come on at the end of that 5-minute delay.
11. Inverter Dynamic Response Profile
Each inverter has a slightly different delay before getting to a commanded output. For example, if I command an inverter from 0% to 100% output, it might take a couple hundred milliseconds, but could take longer. In your control algorithm, it’s important to account for delays between when issuing a command and when the inverter does what it’s told.
Once we’ve considered inverter status and behavior, are we ready to send an inverter command? In many cases, yes, but there are a few forward-looking capabilities you might still want to integrate, even if you don’t have them in place right now.
12. Cloud Affected Ramp Rate Limiting
This is a serious factor in very small grid applications (Hawaii, Guam, Puerto Rico). In addition to normal site ramping, they are also required to ramp back up from the lowest output after cloud departure at 10% per minute. Controlling this is not a straightforward process. It’s not as easy as turning down inverter response and ramping slower than 10% because any underperformance due to slow ramping ends up costing the owner a lot.
13. GHI vs. POA
If your tracker points 60° to the east in the morning but clouds are diffusing irradiance, you may be able to deliver more energy by flattening trackers instead of pointing toward the sun. During times of high diffuse irradiance, horizontal irradiance is higher than plane of array irradiance from a module pointing toward the sun. The control system needs to override typical tracker position based on what’s best for site energy output. Luckily, it should have access to met station data, and can compare GHI and POA irradiance.
14. Battery Storage Charge/Discharge
If you’re planning on integrating battery storage, your algorithms will change depending on if you’re battery is charging or discharging. Are you trying to time shift energy or limit ramp rate? How exactly are you going to use the stored energy? Do you want to charge off solar during the day or night?
Now that we’ve considered all these controls, we should have the ability to determine what the inverter command should be right now for this moment in time in order to optimize the system.
Now we can dive into the more complicated topic of reactive power control!
Reactive Power Mode
If you’ve been involved in distributed generation sites 5MW and smaller, not tied to transmission, this will be a new topic to you. If you’re involved in large generator interconnection agreements, reactive power modes are something you’re familiar with. However, I will probably bring up topics you haven’t yet considered.
15. Automatic Voltage Regulation Mode
I’m seeing more and more sites required to do voltage control, and running in voltage control full time. There are three different voltage regulation modes: fixed VAR output, power factor mode, and a voltage control mode. You need to know which modes are required for your site, and which you’re going to operate in. (This topic usually involves a long, deep discussion with your solar controls system engineer.)
16. Ramp Rate
How fast do you want to ramp VARs? You want to respond to voltage deviations more quickly, but don’t want instantaneous VARs. I usually recommend if your active power ramp rate is 10% per minute, the ramp rate for VARs should be 10% per second.
17. Required PF Range
The normal PF required range in IA is +/– .95, but I’ve seen broader. In those cases, you might need supplemental VARs beyond what inverters can produce, oversized inverters, or more inverters than you really have a need for from an active power standpoint.
18. Command Source
Like with active power, there can be multiple sources for what the voltage or AVR setpoint should be. It might come from an HMI, energy management system, balancing authority, or override algorithm in the control system. A setpoint can come from multiple locations, and the controls can interpret a setpoint based on line conditions.
19. Inverters Available vs. Unavailable
Do you have enough inverters available to produce your required amount of MVARs, or do you need to use static VAR compensation?
20. Active or Reactive Power Priority
In my experience, inverters are set up for reactive power priority 100% of the time. From a grid stability standpoint, VARs are more important than active power, even if you’re not getting paid for VARs. You are required by the IA to support VAR requirements even when the site is running, and that is more important than delivering power. So for inverters, reactive power priority is the norm.
21. Thermal Derating
If you’re in a high temperature location, will inverters have a thermal derating that affects their ability to deliver VARs, power factor, or active power?
22. Battery Storage
You can use battery storage to deliver kW or VARs, but you may not be able to use it for both. Battery storage can supplement what you’re making with inverters. Or, if your site is big enough and there’s enough need on the transmission line to supply ancillary services, you may have battery storage on the site purely for VAR support. Will you use battery storage for active power or VARs?
23. Line Loss Compensation
A large site many miles away from the substation may not get real-time metering information. If you’re trying to maintain PF at that substation, site output might not agree with what’s getting to the end of the feeder line. You may need to do some calculations on what the voltage drop is through that line to make sure you’re in compliance with the IA.
24. Static VAR Sources
If you have capacitor or inductor banks, how are you going to use them? How many capacitor bank steps do you have? Make sure you spread the duty cycle around so you engage all capacitor banks in a relatively balanced fashion. (For that reason, you may not want to have a bunch of different sizes.)
25. Dynamic Response and PQ Curve Shape
Almost all of today’s inverters have semicircular or rectangular PQ curves, so you can make VARs with no or low active power.
26. Night VARs
Most central inverters support night VARs, a capability that lends itself well to ancillary services.
Finally! Now we have all the information we need about VAR sources, device capabilities, and how we need to deliver those VARs.
Now we’re ready to send commands to each inverter for what their VAR output ought to be.
In summary: controls can (and should be) complex
There are a LOT of decision points you can factor into how the control system should work. Most owners don’t think about more than the basics when designing a PV plant, but you don’t want to box yourself in a design where you can’t add functionality later.
Remember that a simple control system will cost a lot less, but may not give you the performance you really want from your solar site.
There are very few integrators that know how to design a control system that includes all these decisions. If all these considerations are looking a little intimidating, or if you have questions about the design of your solar site, I would love to chat with you about your challenges.
Adam Baker is Senior Sales Executive at Affinity Energy with responsibility for providing subject matter expertise in utility-scale solar plant controls, instrumentation, and data acquisition. With 23 years of experience in automation and control, Adam’s previous companies include Rockwell Automation (Allen-Bradley), First Solar, DEPCOM Power, and GE Fanuc Automation.
Adam was instrumental in the development and deployment of three of the largest PV solar power plants in the United States, including 550 MW Topaz Solar in California, 290 MW Agua Caliente Solar in Arizona, and 550 MW Desert Sunlight in the Mojave Desert.
After a 6-year stint in controls design and architecture for the PV solar market, Adam joined Affinity Energy in 2016 and returned to sales leadership, where he has spent most of his career. Adam has a B.S. in Electrical Engineering from the University of Massachusetts, and has been active in environmental and good food movements for several years.