Home

Chapter 1
Chapter 2
Chapter 3
Chapter 4
Chapter 5
Chapter 6
Library

Useful Links

4.1.3 Sequential Decision Processes and Adaptive Planning

The control of markets and electricity grids must be done on a continuous basis, which necessitates ongoing decision-making regarding the supply availability, demand management, unit commitment, dispatching, ancillary services and regulation.   For practical reasons, this decision process is carried out by updating plans at discrete points in time as opposed to continuously adapting to changing conditions.   The time interval between updates of a plan is called the planning interval and the number of intervals into the future over which the plan is specified is called the planning horizon .   At the beginning of each planning interval, an asset manager's updated plan is based on the current status of the asset and its environment and the latest forecast of demand, supply and prices. Moreover, opportunity costs influence the plan, as current decisions depend on future plans.   The periodic updating of plans forms a sequential decision process (SDP) , which is a general term that identifies any process through which a sequence of decisions is made over time in a way that each decision is adapted to the legacy of all previous decisions and to the outcomes of uncontrollable influences on the performance measures that are pursued by the decisions.

A widely-used heuristic approach to solving SDP's is known as the rolling horizon planning procedure.   This procedure is executed through the combination of three planning techniques:

•  Rolling the plan: Plans are updated at regular intervals.  

•  Planning over a horizon: Each plan extends over a number of future time periods.

•  Adapting the plan: At each update of the plan, the plan is adjusted within limits that are determined by the system's constraints on the rates at which resource flows can change.   The planning horizon for each plan consists of a horizon over which the plan must be "frozen" followed by a horizon over which adjustments are allowed.   The boundary between the fixed portion of a plan and the adjustable portion of a plan is called the planning "fence".  

Of course the discrete representation of the time scale within a SDP for a process that changes continuously introduces an approximation.   However, the notion of developing a plan in finer and finer detail as one proceeds through the levels of the hierarchical planning method described in Section 1.2 applies to the time scale as well.   Higher-level, more strategic decisions are given longer planning horizon and longer planning intervals than lower-level tactical or operational plans.   As one moves down the hierarchy of decisions, the planning horizons and the planning intervals are made shorter.  

Another approximation that is inherent in a rolling horizon and adaptation procedure stems from the use of a deterministic forecast for each plan update.   The accuracy of this forecast increases as the forecast horizon decreases.   Consequently, the adaptation options with the shortest time fences enjoy the most accurate forecasts and can be viewed as "fine tuning" actions with respect to the "coarse tuning" of the plans produced by the longer-fence options.

Table 1.3a shows the basic scope and definition of the five levels of hierarchical planning, which make up power system management.  

Decision Domain

Planning Horizon (typical)

Planning Interval (typical)

Configuring

> 1 year

> 1 month

Planning

1 day - 1 year

1 day

Scheduling

36 hours

1 hour

Dispatching

1 hour

5 minutes

Controlling

0.5 hour

< 5 seconds

Table 1.3a: Planning horizons and periods

In the case of electricity scheduling and dispatch there are four options for specifying and updating a plan. Each option is constrained to be exercised within the capacities that are set by the capacity reservation decisions made at a higher level of the decision-making hierarchy (see Section 1.2 ).   Tables 1.3b - 1.3e define these options.   The update intervals, planning horizons and time fences given in Tables 1.3c and 1.3e are typical values in the operation of a large power system.  

Scheduling option

Capacity constraint

Demand constraint

Day-ahead unit commitment

Day ahead offers

Day-ahead bids

Imbalance commitment

Imbalance offers

Imbalance bids

Regulation reserve commitment

Regulation reserves offers

Regulation forecast

Spinning reserve commitment

Spinning reserves offers

Control error forecast

Table 1.3b: Capacity and demand constraints on scheduling options


Scheduling option

Update interval

Planning horizon

Time fence

Day-ahead unit commitment

24 hours

36 hours

12 hours

Imbalance commitment

24 hours

30 hour

6 hours

Regulation reserves

8, 16 hours

9, 17 hours

1 hour

Spinning reserves

8, 16 hours

9, 17 hours

1 hour

 

Table 1.3c: Scheduling option parameters

Dispatch/Control option

Capacity constraint

Demand constraint

Day-ahead dispatch

Day-ahead commitments

Day-ahead commitments

Real-time dispatch

Imbalance commitments

Demand forecast

Ancillary service regulation

Regulation reserve commitments

Regulation error

Voltage/frequency control

Spinning reserve commitments

Control error feedback

Table 1.3.d: Capacity and demand constraints on dispatching options

Dispatch/control option

Update interval

Planning horizon

Time fence

Day-ahead unit commitment

8, 16 hours

9, 17 hours

1 hour

Real-time dispatch

1 hour

1.5 hours

30 minutes

Ancillary service regulation

5 minutes

30 minutes

5 minutes

Voltage/frequency control

4 seconds

30 seconds

4 seconds

Table 1.3.e: Dispatching option parameters

In the re-structuring energy markets of the United States, generation-unit commitment decisions are made through a combination of self-scheduling decisions made by generation asset managers and market clearing of bids and offers for electric power through markets that are managed by ISO's.   See Module 3 (under construction) for an explanation of the workings of these markets.

4.1.4 Potential Approaches to DG Power Management

The integration of DG capacity into the management of a regional power grid presents some new opportunities and risks.   Much work and experimentation needs to be done before the proper role of DG in the power systems of the future can be determined. In this section we examine some of the alternatives in the use of DG.

Four different categories of business entities appear likely to consider the implementation of DG technology as an element of a strategic plan.   These include:

  • Investor owned utilities (IOUs) and publicly owned utilities that may want to install DG units for supplying peak demand in areas that are located behind congested transmission lines.  
  • Manufacturers of DG systems, which have already advanced the technologies for DG on several fronts and, with the exception of large gas turbines, appear to be pacing their capacity growth by market growth.  
  • The new generator and consumer (NGC) that sees operation of DG units as a potential substitute for some or all of its purchases of electricity from utilities.   Included in this category are industrial sites, apartment complexes, government agencies, military bases, universities, hospitals, shopping malls, and the like.   On the horizon, NGCs may also become wholesale energy suppliers by interconnecting DG units.
  • The contractor industry that performs one or more of the functions of designing, building, installing, and operating DG units.   The viability of contractors depends on the rate of adoption of DG by the above categories of business entities.  

The most common business model for DG adoption and growth may consist of one that partners new generators and consumers with DG contractors.   Such partnerships relieve the NGCs of the need to develop extensive expertise in power generation and management, which would distract them from their core business operations.   Instead, NGCs can engage the services of contractors who have expertise in constructing and operating DG units.  

A foray into power generation by an NGC is a risky venture, even with the assistance of a qualified contractor.   Several categories of uncertainty engender financial as well as non-financial risks for the NGC.

  • Technological uncertainty:   DG owners must deal with the chance that the DG technology will not perform as reliably or as efficiently as its specifications.   In particular, interconnected DG systems may actually reduce the reliability of a distribution grid due to the inability of grid operators to control unit dispatches under rapidly changing conditions.
  • Fuel cost uncertainty:   The price of natural gas, coal, and oil will affect the financial performance of any DG unit that uses any of these fuels.   Owners of renewable energy technologies (such as wind turbines) will not need to worry about the cost of energy resources, but the price they receive for surplus power will depend to a large extent on the price of conventional fuels that provide competitive benchmarks prices.
  • Load uncertainty:   The growth and volatility of electricity demand within the transmission grid that serves the NGC must be considered.   Ironically, efforts to reduce the cost of electricity in the form of demand-side response could reduce the value of DG units that are most beneficial in supplanting expensive peak-load power from utilities.
  • Electricity price uncertainty:   The financial performance of a DG unit depends on the cost of electricity from utilities that the unit supplants.   Future prices of electric power in the United States remain highly uncertain due to variability in fuel costs, regulation, and technological change.
  • Regulatory and public policy uncertainty:   The viability of DG projects depends, to a certain extent, on the actions of government entities.   NGCs need to consider the chance that any tax incentives, subsidies, or easements associated with a DG implementation may be offered or repealed by future legislatures and executive branches.   Given the spotty history of utility system restructuring and deregulation, it is difficult to predict the effects of government policies on evolving electricity markets.  

In spite of the risks associated with these uncertainties, there are opportunities for the profitable and otherwise successful use of DG.   One approach to the implementation of DG merits consideration for its potential to produce profits for DG owners as well as benefits to grid reliability, security and environmental impacts. This approach aggregates the operation of DG units into a single management entity that can dispatch power to the grid.   Through this management DG owners would employ electricity and waste heat for their own needs when power remains cheap, and would sell surplus power to the grid when spot prices for electricity are high.   However, in order to participate in the market for power, DG units must be dispatched at times and locations where they are most needed.   This kind of dispatch requires coordination of all DG units in a distribution grid; hence, the need for a supply aggregator to manage multiple DG units.  

Through this managerial hierarchy, DG owners would contract with a supply aggregator to sell electricity to the power grid in much the same way that consumers contract with load aggregators (retailers) to purchase electricity from the grid.   The supply aggregator trades off the cost of generation from the DG units against the spot prices available from the wholesale market.   Furthermore, the supply aggregator has the expertise to protect DG owners from the financial risks of price volatility through the trading of commodity derivatives.   Given the special expertise and close attention that such trading requires, DG owners would find the services of a contractor beneficial in terms of financial performance as well as a relief from the need to develop expertise outside of their core competencies.   The supply aggregator of the power systems of the future would design, install, maintain, and operate DG units for the owners as well as trade the power produced by the units and manage risks associated with operation of the units within an integrated distribution grid.  

An experiment in New York, funded by the New York State Energy Research and Development Authority (NYSERDA) and the U.S. Department of Energy, suggests that such dispatch can occur efficiently.   In this experiment, fifty backup generators (providing a capacity of 35 MW) were linked to centralized control points for dispatch to the power grid.   The centralized dispatcher served as an aggregator that purchased bulk electricity from utilities and sold it in the day-ahead market in different parts of New York.   When real-time prices exceeded the cost of generation from the DG units, the aggregator dispatched the DG units and earned revenues.   This aggregated DG system proved to be technically feasible and produced savings of $1.5 million in one year through the use of islanded DG units for load curtailment. The success of the NYSERDA/DOE experiment should motivate serious interest by business and government regulators in the potential for aggregated dispatching for DG.

 

<< Back