Operators of the power system are facing a paradox – on the one hand, the grid is becoming smarter which gives operators an extra “edge” in managing the grid, but, on the other hand, policy changes associated with the promotion of renewable power and customer choice are creating increased uncertainty in both supply and demand. Yet operators must still deliver power with an average service reliability of 3 x 9s (the LOLP or LOEP target in system planning models), while also maintaining short-term stability and security of supply, even as the inertia of the system decreases.
Increases in supply uncertainties are caused by the steadily increasing penetration of renewable energy production as a result of RPS mandates. Delivery uncertainties are on the rise due to lagging transmission construction leading to congestion.
Demand uncertainty is increasing as self-optimizing end-users install distributed energy and storage, smart appliances that use M2M controls and chargers for EVs, and take advantage of time-differentiated pricing.
Electricity dispatch is based on minute-to-minute forecasting and clearing of customers’ “net load”. That is, the end-use customer’s load minus any local power generation or discharging of energy storage devices.
But the utility or system operator usually does not have visibility into, or interconnection with, this generation behind the meter – creating demand uncertainty and an inability to take advantage of the distributed generation in reliability emergencies.
Price Volatility and Price Elasticity-Created Uncertainty
The market as structured also creates short-term operational challenges for the grid operators. Prices in wholesale markets can change rapidly over short periods of time, e.g., 15 minutes, leading to sharp changes in the availability of supply and in the level of demand, driven by price elasticity. These impacts of price volatility, combined with the increased percentage of intermittent resources, creates the need for additional fast-acting reserves to maintain the grid operator’s target service reliability level.
How can the power system operator cope with the increased physical- and market-driven uncertainties?
We suggest that a “least cost” coping approach can consist of a combination of (1) investments in the “smart grid” (smart sensors, advanced controls, data analytics), and (2) investments in flexibility products (to be defined below).
We have discussed (1) above, capital investments in smart grid automation and control equipment, in previous dialogs. Approaches include:
- Better situational awareness to get more out of assets -- more measurement, smart sensors
- More advanced controls, reacting faster, “circuit breakers” to break high-gain feedback loops that create propagating instabilities
- Increased communications
- Integration of utility systems (OMS, GIS, CIS, SCADA, DMS, CIS, etc.)
- More interconnections
- Larger balancing areas
- Utility-dispatchable DG
- Behavior change through market mechanisms (TOU) and incentives
- Production curtailment
- Customers’ power use management – DG, UPS, smart appliances
- Better short-term load forecasting
- DSM 2.0 programs
Flexibility Products: An Alternative to Traditional Power System Capital Expenditures
Moving to (2) above, we’ll identify some flexibility products, and attempt to define a conceptual supply/demand market for them based on their costs (price of supply) and benefits (market-driven demand).
How many times recently have you heard the call for increased operational flexibility from utilities or ISOs? – “the grid needs to be more flexible.” But what does that mean in concrete terms?
Let’s try to make flexibility a “concrete”/tangible good. Let’s even try to price it within a market for flexibility products and services. Can/should we apply option pricing algorithms (e.g., Black-Scholes, or valuation of “puts” and “calls”) to determine the value of flexibility products associated with grid operations? We should be able to apply the same methodologies that energy traders use to calculate analogous option values for every hour of every day for energy, capacity, ancillary services, and congestion products in wholesale markets.
Operational flexibility is not new to the power sector -- we have been incorporating flexibility in grid planning and operations for decades to manage the stochastic uncertainties in power grids – examples of these traditional stochastic uncertainties include forced outages, unpredictable user behavior, and weather. For example, system planners use Monte Carlo simulations to incorporate the effects of forced outages. We will continue to use traditional hedges (i.e., flexibility products) against stochastic behavior – examples include increased amounts of ancillary services/reserves, contingency planning (N – x), interruptible rates, DLC, and conventional DR.
Obviously maintaining additional optionality costs more. So, to maintain same 99.97% reliability as grid supply/demand uncertainties increase, costs will go up.
What additional non-traditional flexibility products/services are available to the grid operator? Some examples of flexibility "supplies" (with different associated prices) that comprise the flexibility supply curve are:
- Bulk and distributed energy storage plants
- Ramping products; reserves that ramp more quickly
- Inertia products
- Generator tuning services
- Power System Stabilizer (PSS) is an application that enhances damping to extend power transfer limits. Enhanced damping is required when a weak transmission condition exists along with a heavy transfer of load. A PSS works in conjunction with the excitation system of a synchronous machine to modify the torque angle of the shaft to increase damping.
- Regulatory/market structure changes: connect wholesale and retail markets to eliminate perverse effects of mis-matched pricing approaches
- Utility and/or ISO access to customer-owned DG, i.e. the Virtual Power Plant (VPP) Concept
- Advanced DR
- Improved situational awareness allowing plant to be run closer to design capacity
- Strategically placed smart sensors for situational awareness and alerts
- Data analytics
- Smart controls for increasing grid reliability, stability and security
- Improved short-term forecasting: load, weather
The Flexibility Supply/Demand Curve
Can we calculate the cost of the above flexibility products? Probably, so therefore we can graph the supply curve for a set of “flexibility products” versus price (i.e., cost + profit, if any).
On the demand side, can we plot how much demand there could be to purchase various types of flexibility at different prices? Not so easily, we suppose, but wouldn’t grid operators continue to buy flexibility until the marginal benefits of the next increment equaled the marginal cost of same?
So, it gets down to calculating how much flexibility is worth, i.e., the value of the optionality that is creates. This value could have two components: the expected “avoided costs” of events that disrupt reliability and the arbitrage value of the optionality in the absence of the need to use the optionality to mitigate a reliability event.
Two Final Challenges
Who pays? Maintaining the existing level of reliability is going to cost more.
If maintaining the traditional reliability level is going to cost more, what is the trade-off between traditional capital investments to maintain reliability versus the same grid operator investing in non-traditional flexibility products?
It seems that an efficient set of market mechanisms might offset the cost of flexibility with better utilization of existing assets. If true, such a path could reduce the enormous costs of the transition to the smart grid by using such non-traditional cost-saving mechanisms.
One last point. We have assumed that reliability levels remain constant. If we let reliability levels “float”, like a national currency, would people be interested in buying the level of service reliability that they want from a menu? Presumably they could have the option to buy more/less reliability (and pay more, or less) at certain times of their choosing. They could also invest in behind-the-meter reliability-enhancing equipment. Instead of a “flexibility” supply/demand curve, we would then have a reliability supply/demand curve.
As always, comments welcome and appreciated