vartanian headshotCharlie Vartanian is a Sr. Technical Advisor at the Pacific Northwest National Laboratory where he focuses on integration of energy storage with power systems. Charlie has over 25 years of power industry experience deploying advanced grid technologies, performing electric system studies, and contributing to technical standards development. He has worked previously for Mitsubishi Electric, A123 Systems, Southern California Edison, and the U.S. Navy. During his 15 years at Southern California Edison, Charlie’s activities spanned traditional T&D planning through grid R&D. He is a currently Secretary of the IEEE 1547.9 Energy Storage Interconnection working group.

alam headshotDr. Jan Alam is a power systems engineer at Pacific Northwest National Laboratory (PNNL) where he is working since 2016. He is a thrust-area lead within Energy Storage Industry Acceptance program at PNNL, sponsored by U.S. DOE. He is also a key-contributor in PNNL transactive energy systems and grid services valuation domains. He is managing an effort to develop capabilities at PNNL for grid integration of marine renewable energy resources. Before joining PNNL, he was engaged in solar PV and energy storage research in Australia and collaborated with multiple Australian utilities. Dr Alam also worked as an electric power industry professional in Bangladesh (2005-2010) and held various positions from plant maintenance engineer to consulting engineer working with the government agencies responsible for power sector development.

In this interview, Charlie and Jan answer questions from their webinar, Storage Control for Power System Oscillation Damping and Inertia, originally presented on June 18, 2020.

What does acronym SGSS mean in slide 27?

Smart Grid Storage System, a product related term formerly used by A123 Systems

Shouldn't the inverter operate below maximum output power to react to such sudden frequency drop?

Limiting dispatched power output to reserve some headroom is one option. Another option is to use short term overload capability in the ESS, primarily the inverter.While short term O.L. of existing inverter technology is limited (1.2 P.U., for under a minute), the timeframe for providing inertial support is fairly short (0-15 seconds). And, future Si-C (silicon carbide) based inverters will have higher O.L. capability.

In slide 13, can you implement a multi-mode control (open loop under some conditions and closed loop under others)?

Yes. One implementation is frequency droop with deadband. If monitored frequency is within an acceptable deadband, the ESS performs normal dispatched service (Freq Reg, Peak Shifting), but if monitored system frequency is outside of the deadband, then droop response is performed in form of FR, FFR, or ‘synthetic inertia’ as described in the hybrid generator+ES control discussed later in the webinar.

What usually the physics of the storage? Li-I batteries?

The electro-chemical response and quick bi-directional charge/discharge capability of the main types of batteries used for grid are fast enough to provide FFR and inertial support. Li-ion is particularly well suited. But, PbA, Ni-CAD, Redox Flow and even super caps are technically compatible.

Frequency stability depends on inertia and damping. So, can we solve the problem on the damping side, considering that inverter-based systems act much faster and hence have much larger bandwidth?

Yes. And, the speed of response can be a benefit. BUT, the fast high-bandwidth (high oscillatory) response needs to be designed-in. The speed and bandwidth could be destabilizing if not managed with some technical care.

Which type of BESS are you assuming? Short time storage might be super-capacitors, but what storage capacity and charging currents are required?

The electro-chemical response and quick bi-directional charge/discharge capability of the main types of batteries used for grid are fast enough to provide FFR and inertial support. Li-ion is particularly well suited. But, PbA, Ni-CAD, Redox Flow and even super caps are technically compatible.

The limitation in terms of speed of response is actually the sensing and logic that decides when and how to inject power. The next time-limiting subsystem of an ESS is the inverter. In general, the batteries, caps, or flywheels are relatively fastest subsystems of ESS’s.

What about fly-wheels - are these a candidate?

Yes, however, the limited energy would limit the total set of services that could be provided.

Is there a requirement of BESS size in the power system? And is there any requirement of system power factor?

There is a minimum size requirement for ESS projects connected to BPS via an ISO/RTO interconnection process. But, these minimum size limits have been dropping from MW, to 100kW’s, and now consideration of 10’s KW.

And, aggregators have been bundling smaller ESS’s for participation in A/S markets.

In general, FRR, and more generally inertial support are feasible at any capacity. The limitations may be more based on the economics and regulatory aspects of getting compensated for adding this grid-supportive capability. Or, embedding them as a condition of interconnection, e.g. PREPA Interconnection rule’s MTR’s minimum H requirement.

Do you see Battery Storage systems providing synthetic inertia as a realistic long term solution, or will this need to be coupled with synchronous condensers?

Yes. And long term, there’s a possible scenario where more sync condenser capacity may be added thru re-purposed retired generating plants, PLUS more inverter based resources deployed within a power system having FRR and synth inertia capability. As renewables penetration continues to rise, some system may need all available help for inertial support.

How different is business case impact between droop control and synthetic inertia? In other words, in which case synthetic inertia would be necessary over just droop control?

Synthetic inertia mimics an intrinsic feature of rotating generators whereas droop control is more of a reactive control action in response to a change in system. Therefore, ideally any synchronous ac grid can benefit from artificially created inertial support. The degree of usefulness of synthetic inertia will depend on the available inertia in a system which is a dynamic feature. The business case impact will vary because of similar reasons. Diverse response characteristics of various rotating generators add to the complexity.

Question from an European participant: Is there currently a market for frequency restoration reserves in any of the NERC councils? What are the market incentives for energy storage?

Changes have been ordered by FERC for newly interconnecting generation facilities to install, maintain, and operate equipment capable of providing primary frequency response as a condition of interconnection which can impact the compensation schemes in place. Please see details here.

Energy storage incentive structures for the provision of grid services are not uniform across all ISO/RTOs in the USA. The unique features of energy storage are also not fully captured. FERC order 841 was intended to enable better participation of energy storage into electricity markets. Please see recent news on order 841.

Is there any time limit before system become unstable?

Depends on system, relevant assets, prevailing operation, and type of contingency. Range could be from a few tens of seconds to minutes or tens of minutes for cascaded events.

How do you ensure that BESS control for grid damping events doesn't fight against AGC systems in the grid e.g. WECC?

The spread between time constants and gains between inverter based and traditional assets has, and can be set so the faster responding assets stay well outside of interfering. And, traditional SIS’s that include dynamic stability studies can assure there are no negative controls interactions that trigger oscillations in other machines, or worst case the whole power system.

In application, could PV+Storage operations be biased to allocate a fixed portion of storage, both charge and discharge, to be available for frequency response?

Technically yes. However, the decision could well depend on the operational philosophy, opportunity cost, applicable standards, regulatory requirements, etc.

I would say that a SG dominated power system (with high H) tends to be more oscillative in terms of frequency (due to interaction between generators), how has this issue been solved historically in classic ac grids?

Integrated power system stabilizers (PSS) in SG units and damping controllers built in FACTS devices (e.g., TCSC, SVC) have been used for decades to improve oscillatory behavior in power systems.

Inertia Response: As understood, this synthetic inertia response from ES, is based on grid Frequency and ROCOF "measurement". However, the system ROCOF is triggered by a sudden load imbalance in the system and respective Synchronous machines response. Question: How do you see the effect of "Synchronizing Power" coefficient acting here? What about the delays in ES response? Wouldn’t this create a situation that under a Large Load imbalance, the synchronous machines would get git by large power and enter an unstable point before the ES synthetic inertia response takes effect?

The choice of input variable for control of synthetic inertia is not limited to system frequency only, rotor angle could well be included among other options depending on the event. Therefore, it could be a function of synchronizing power coefficient as well.

DC battery modules and modern power electronics in ESS are extremely fast and able to provide necessary response. In some installations, power electronics response rates may be limited for certain reasons. When configured for synthetic inertia applications, those settings could be tuned to the need.

Where are the BESS locations? Is it near to the RE?

For the synthetic inertia example in this webinar, the ES closely AC coupled with the associated traditional generator synchronous rotating generator. I.e. both on the same medium voltage AC bus before step-up for delivery at project level at transmission voltage.

For ES+RE hybrid systems, there is the additional flexibility of even closer coupling on DC side ahead of conversion from AC/DC, or even remote scenarios. This flexibility brings more technical challenges to work out, but also possibility of greater efficiencies and impacts. A great R&D area for exploration.

What is the unit of inertia used in power systems analysis, planning and operations?

“H” measured in units of MW-seconds

Is the inertial contribution of a generator dependent its dispatched or operating power level, e.g. its ‘Pgen’ value?

No.  It’s based on the rated MVA of the machine. The machines “Pmax” in context of Load Flow model input.

Name two types of synchronous machines.

Any two of the following: Synchronous Generators, Synchronous Condensers, or Synchronous Motors.

Name one impact to power system performance due to the loss of overall system inertia.

Any of the following:

  • Wider frequency deviations for grid disturbances
  • Inability to meet NERC frequency deviation limit criteria
  • Worst case – frequency collapse
  • Increased wear and tear on connected synchronous machines

Can an energy storage system directly supply “H” inertia?

Yes. But limited ESS’s that interfaces with the grid through a synchronous machine, like large scale pumped storage ESS’s.

Can an inverter-based ES system directly supply “H” inertia?


Do DC power systems need inertia?

No. If all resources and loads are DC.

Do loads provide inertia to the power system?

In limited but important cases: directly coupled synchronous motors. Large resistors provide part of the ‘braking’ side of the overall inertial response, e.g. BPA’s Chief Joseph Brake.

Stay Informed - Join the IEEE Smart Grid Community Today

To view past interviews, please visit the IEEE Smart Grid Resource Center.