It is clear to us that energy policy and regulations are the key drivers of the business case for SG applications, and not technologies. These policies/regulations promote, for example, RPS mandates, dynamic pricing, demand response, competitive market structures, self-optimizing customers (e.g., distributed generation and storage, smart appliances, micro-grids), electric vehicles, cyber-security, and data privacy. It is a kind-of “policy-push” market, with SG applications in a “catch-up” mode.
In order to implement the new policies and regulations in all of their complexities and not-designed-for impacts on the traditional electricity grid, while still maintaining the current levels of service reliability, stability and security, the grid needs to be smarter, and react faster. We will be operating “closer to the edge”.
The SG is at its core about automation, control, and optimization across the power system operations – both physical and market operations. For example, it comprises smart sensors, intelligent electronic devices, communications systems, M2M interfaces, data analytics, situation awareness and alerts, and control systems.
In its ideal form, the SG is a system of systems that in essence have the potential to optimize power system operations and capacity requirements. To realize this potential, i.e., for the grid to be “smart”, these systems ultimately need to be interoperable since the SG is an interconnected system from generation all the way to end-use of electricity.
The above new policies/regulations are out ahead of the SG in terms of technology, interoperability, and grid operations – the SG is playing “catch-up”. But more importantly, we also need the SG in order to realize the full benefits of these new policies and regulations.
The “catch-up” situation can lead to unintended/undesirable consequences related to the operation and reliability of the power system.
Fortunately, SG applications have the capability, if not yet the readiness, to mitigate these risks, provided they are interoperable.
The Transition to an “Ideal” SG Architecture Will Be Messy -- We Are Going To Feel Uncomfortable
As engineers, we like tidiness. In a perfect world, the transition to a fully-functional SG would be orderly and paced to accommodate new applications while protecting grid integrity: perhaps a three-stage transition -- from today’s operations’ data silos in utilities to a single common information bus, then to many common, integrated buses, and finally to a converged system.
But in a non-perfect world, i.e., reality, the SG will evolve as a hybrid of legacy and new systems -- it will not be an optimized process – there will not be a “grand plan” – clusters of interoperability will appear here and there across the SG.
The transition will take perhaps 30 years -- not for technology-based reasons, but because the “refresh cycle” for utility assets is lengthy – so, there’s time for a whole career for all of us in deploying SG applications!
Interoperability itself will be implemented pragmatically, i.e., a combination (1) customized APIs, and (2) the development of new standards.
The transition process will be piece-meal, scattered -- interoperability added in different parts of the SG over time with priorities driven by (1) new policies and regulations -- local and federal, (2) reliability requirements, and (3) economics. It will be based on thousands of individual, situational decisions, and on “learnings” from real-world experiences.
Because of the piece-meal character of SG application deployments, and the fact that the power system is a fully interconnected system, unintended impacts on power system operations will be created.
Grid operators will be left to figure out how to maintain historical levels of reliability even as threats to reliability increase. In previous dialogs, we have pointed out that as a direct result of the policy and regulation initiatives, the probability of unserved energy costs and higher levels of SAIDI/SAIFI-related penalties will increase. The task (and fortunately, the capability) of SG applications is to decrease that probability.
What Will the SG Look Like in 2020?
Renewable generation (including distributed PV) will have doubled its market penetration due to RPS mandates, resulting in increased intermittency, reduced system inertia, and increased reserve requirements.
Increased levels of interoperability will have been implemented to facilitate other policy- and regulatory-driven SG applications such as:
- Increased integration of wholesale and retail market operations as dynamic pricing is implemented, e.g., TOU, RTP, and CPP
- Self-optimizing customers that have installed distributed generation and storage, and pre-programmed M2M appliances and devices to manage their electricity bills
- Automated demand response used by operators to manage peak load
- EV charging/discharging, creating a significant impact on local load curves
Similar to today, forecasts of “net load” will form the basis for grid operators making generation dispatching decisions – no-one expects to see closed loop control of the power system by 2020. “Net load” is defined as the load that the utility or grid operator “sees”: the end-use customer’s consumption minus customer generation, minus storage discharging, minus demand response, plus EV charging, plus/minus preprogrammed curtailment or consumption.
Recently, Michael Niggli, President and COO of SDG&E, talked about the radical change in the shape of the “net load” curve that the CA-ISO will need to serve in 2020 (in this case defined as demand minus must-take renewable generation). He called this forecast March 27, 2020 CA “net load” shape the “duck belly” curve. It consists of a steep down-ramp starting at 8:00 am in the morning, with a very low minimum load arriving during the middle of the day (as a result of PV generation), and a steep up-ramp starting about 5:00 pm creating the day’s peak at about 8:00 pm – a very different commitment, dispatch, and operating challenge for the grid operator relative to historical norms.
To compound the operating challenge, the “net load” forecasts will become much more uncertain and volatile due to:
- The intermittency of distributed generation
- Unknown or opaque self-optimization algorithms – “local optimization” that may be non-optimal for grid operations
- Price elastic behavior of customers shifting demand
- Mass actions of pre-programmed M2M devices and ADR participants – actions that can be completed within 4 seconds under the new OpenADR protocol
It is clear that maintaining the same levels of reliability, stability, and security in the physical grid will cost more and require better, more ubiquitous, and interoperable, sensors, automation, and control systems – but isn’t that the forte of the SG?
Another challenge for the SG will be interactions between the physical grid and power market operations. These have the potential to create unintended and unprotected events that we need to anticipate, as the following example suggests.
An Unintended Consequence: A Positive Feedback Loop
As a result of increased interoperability between the wholesale and retail markets for power, the interaction of the physical grid with market operations could create unintended and unprotected events.
Here’s one scenario:
- A large forced outage coupled with an unexpected down-ramp in wind generation creates a price spike in supply – the system operator schedules new supply to fill the gap
- Market price increases sharply
- As a result of dynamic pricing, net load drops quickly due to pre-programmed smart appliances, and /or self-optimizing customers switching their distributed PV generation or micro-grids from battery charging to consumption or to arbitraging the ancillary services market, in the process creating reverse power flows in distribution systems
- If grid operations/generation are “slower” than price-responsive supply and demand, a positive feedback loop is created, leading to grid instability
In their intriguing paper, Masiello et al. of KEMA have simulated the above positive feedback loops in the interoperable grid using a system dynamics model -- the perfect type of modeling methodology to simulate interacting feedback loops -- and concluded that there are realistic situations where grid instability can occur – see here.
The above scenario is not too dissimilar to the “flash crash” in Wall Street caused by high-speed traders out-speeding the market’s ability to clear itself. What did they do to remediate it? They put in a circuit breaker – how familiar! But grid operators, except in very extreme circumstances, can’t simply open a circuit breaker when difficulties arise – electricity service is too critical to society’s well-being. What to do?
But Wait! It’s 2020 – “So, SG, Heal Thyself…….”
In 2020, in anticipation of increased operational stresses, dozens of PMUs making synchrophasor measurements will have been deployed throughout the distribution system within IEDs, DFRs, routers, relays, transformers, pole-top insulators, etc.
The tiny chip-sets will be tunable, field-programmable, and very accurate – down to 0.01 degrees of phase angle difference, measuring voltage, frequency and df/dt 120 times per second.
The PMUs will communicate directly with each other, and, through PDCs, with transmission PMU systems. They will provide real-time state estimates.
And most importantly for this situation, the IEDs will have multiple embedded RAS (Remedial Action Schemes) applications with local or peer-to-peer intelligence, enabling them to take speedy autonomous actions (in milliseconds) at the “edge” of the grid that can attenuate incipient instabilities and secure the grid, e.g., triggering of local energy storage to provide frequency regulation, or of ADR for rapid load reduction.
They will also have data and communication interfaces that are interoperable with legacy field communications and SCADA/DMS/EMS utility systems to initiate coordinated follow-up remedial action schemes within a minute or so.
They will likely be able to swiftly sense autonomous changes in end-use customer generation and consumption to improve the accuracy of the “net load” forecast used by grid operators for dispatch.
Bottom line: the positive feedback stability crisis above can be averted -- but none of this can happen without the implementation of broad-based interoperability within the SG between its many currently siloed automation and control applications.
So, Interoperability Is Mission Critical
As we’ve noted, SG business cases are driven by policies and regulations related to intermittent renewables, dynamic pricing, demand response, competitive market structures, self-optimizing customers (e.g., distributed generation and storage, smart appliances, micro-grids), electric vehicles, cyber-security, and data privacy.
Implementation of these policies and regulations requires SG interoperability.
This implementation stresses the power system’s ability to maintain acceptable levels of reliability, stability, and security.
In the case of the distribution system, it creates power flows for which the system was not designed.
We do not fully understand all of the operating ramifications of installing these new systems - we will need to monitor and measure them in the field.
Fortunately, we can address and remediate these expected operating risks by deploying new SG applications, i.e., advanced sensing, communications, automation, and control systems.
As long as ……..these SG systems are interoperable.
As always, comments are welcome and appreciated in the comment box below…...