Renewable energy resources such as solar and wind, produce power in a manner that generally does not contribute to frequency control of interconnected power systems. For wind turbines, the reason for this is that the generators used to convert wind to electric energy have small inertias that dissipate rotational energy more readily than conventional steam turbines. Also, wind turbines are operated such as to generate optimal power from the available wind, and hence do not have much spinning reserve. For inverter-based solar generation, the solid-state controls have no rotating component at all. (Solar thermal power is usually produced with synchronous generators and thus contribute to frequency control as most thermal-type power plants are able to do.)
However, both wind turbines and solar inverters have the important characteristic of fast, programmable controls. The question then comes up: Is it possible for these power sources to participate in frequency control response of interconnections? This is an intriguing question that merits some further investigation.
[As a note, there is an ongoing Protocol Revision Request (PRR) from ERCOT for Wind-powered Generation Resources (WGRs) Primary Frequency Response that addresses this specific question for wind turbines. But this Blog addresses a more general issue of fast controls that includes solar inverters and possible back-to-back DC controls.]
Frequency Response Defined
Per NERC standard BAL-003, Frequency Response occurs within the first few seconds following a change in system frequency (disturbance) to stabilize the Interconnection. Traditionally, this response is provided by governor action that adjusts the energy input into rotating generators’ prime movers, and by load acting as a resource or shed by under-frequency relays.
[Note: NERC’s Resources Subcommittee (RS) recently initiated a Standards Authorization Request (SAR) for BAL-003 to put a measurement process in place so engineers can objectively analyze the adequacy of Frequency Response and underlying issues to enable informed decisions. The Frequency Response Standard drafting team is proposing a standard with performance goal that each Interconnection can withstand at least an N-2 event without encroaching upon the first tier of Under Frequency Load Shedding (UFLS).
For the typical interconnected system, the system operators determine the system’s reserve response in a measure such as MW/0.1 Hz change in frequency. The critical contingency is usually a severe but credible event such as the loss of one or two of the largest generators in the system, and this contingency sets the spinning reserve requirement in MW. Some of the reserve can be supplied by demand response; i.e., load shedding, but the rest is provided by other generators.
The NERC standards are steam generation-centric in that the responses are highly dependent on synchronous machines to define the operating frequency during normal operation and governors to maintain frequency under contingencies. The droop response specified by NERC is one that is geared to the capability and limitations of large synchronous machine frequency regulation.
Frequency Response from Fast Controls
We postulate that fast controls, as may be found on wind farms or inverter-based solar farms, can be set such that they mimic the response of synchronous machines. This is possible by allowing some capacity to remain in reserve at these devices and then releasing the reserves in a controlled fashion that simulates rotational inertia.
For wind turbines, response time to frequency excursions is much faster since these tend to have smaller inertia than bulk steam power plants. The “spinning reserve” available from a wind farm, comprising of the available wind capacity not converted to electric power (i.e., vanes are feathered or are not utilizing the full energy available from the prevailing wind), can be delivered or released at a much faster rate than steam turbines. This opens the possibility for “pulsing” the wind output to minimize any overshoot in recovering frequency, acting like an SVC to damp out voltage oscillations.
For inverters, pseudo-inertia may be achieved in an even simpler fashion since these have no inertia at all and so can be controlled to the limit that their ramp-rates can manage. The typical ramp-rate is 10% of capacity per second, much faster than needed for frequency control.
Granted, all the above takes some programming effort, but we believe this is doable.
In order for a fast controlled resource to to meet the PRR requirements, as well as comply with the specifications of NERC standards, it needs to demonstrate control of frequency within the specified parameters for Primarily Control; i.e., similar to the frequency response of a steam unit.
Once the concept of using fast controls to supply frequency control takes root, then there are all manner of new applications that may grow from this. Some ideas include:
- Optimize the allocation of energy. By carefully determining the amount of Basic Energy and energy for frequency response, the contribution of generation under fast controls can be optimized in terms of the value received by the plant owners. As daily and seasonal changes impact the energy markets, the allocation can be adjusted to maximize benefits.
- Not only inertia can be simulated but also other control parameters such as gain, feedback signals and deadband. To the limits of control logic, the fast control resource can be made to look as close to a traditional reserve resource as possible.
This is an intriguing idea that is being pursued in many areas of the industry. Overall, we view this as part of the overall transformation of the energy supply landscape as more and more renewable resources are integrated in power systems. Eventually, the standards themselves may change, and the concept of rotational inertia replaced by some new paradigm such as “frequency stability.”
If only for these dramatic changes, it is an interesting time to be a power engineer.