Q9Power System Planning

(a) Seven Modular Programs of the WASP Package

WASP (Wien Automatic System Planning Package), developed originally at the International Atomic Energy Agency, is a widely used computer package for least-cost generation expansion planning of a power system over a specified study horizon. It determines the optimal sequence, type, size and timing of new generating units to be added to an existing system so that forecast demand is met at minimum total discounted cost while satisfying specified reliability criteria. WASP is organized into seven distinct modules, each handling a specific part of the overall expansion-planning computation.

• LOADSY (Load System Description): processes the peak demand forecast and the shape of the annual load-duration curve for each year of the study period, providing the demand data required by later modules.
• FIXSYS (Fixed System Description): describes the existing and already-committed generating units in the system (their capacity, availability, operating cost characteristics) that are 'fixed' and not subject to the expansion-planning decision.
• VARSYS (Variable System Description): describes the characteristics (capital cost, fuel/operating cost, availability, technical parameters) of the candidate new plant types that may be added during the study period, from which the optimization selects.
• CONGEN (Configuration Generator): automatically generates all technically feasible combinations (configurations) of fixed and candidate plants for each year of the study period, subject to user-specified constraints.
• MERSIM (Merge and Simulate): merges the load and system data and performs a probabilistic production-cost/reliability simulation (using a loss-of-load probability approach) for each configuration generated by CONGEN, computing expected energy not served, operating cost and reliability indices.
• DYNPRO (Dynamic Programming Optimization): uses dynamic programming to search across the configurations and years simulated by MERSIM to identify the expansion sequence that minimizes total discounted cost over the entire study horizon while meeting the specified reliability constraint in every year.
• REPROBAT (Report Batch): generates the final consolidated output reports summarizing the optimal expansion plan, costs and reliability indices selected by DYNPRO for review by the planner.

(b) Steps in Optimal Power System Planning Using WASP

Optimal power system generation expansion planning aims to determine, over a multi-year study horizon, the least-cost combination and timing of new generating capacity additions that will meet forecast demand while satisfying a specified minimum reliability level. The WASP programme implements this through a structured sequence of steps.

• Load forecasting: prepare the peak demand and load-duration curve forecast for every year of the study period (input to LOADSY).
• Characterize the fixed (existing/committed) system: compile technical and cost data of all currently operating and already-committed units (input to FIXSYS).
• Characterize candidate expansion plant types: compile capital cost, fuel cost, heat rate, availability and other technical data for every type of new plant that may be considered for addition (input to VARSYS).
• Generate feasible configurations: for each year, generate all technically permissible combinations of existing and candidate plants subject to specified addition/retirement rules (CONGEN).
• Simulate operation and reliability: for every configuration, simulate system operation probabilistically to estimate production cost, fuel consumption, and reliability indices such as loss of load probability and expected energy not served (MERSIM).
• Optimize the expansion sequence: apply dynamic programming across all years and configurations to identify the sequence of capacity additions that minimizes total discounted cost (capital plus operating plus cost of unserved energy) over the whole horizon while meeting the reliability constraint in every year (DYNPRO).
• Report and review: generate the final report of the optimal plan (REPROBAT) and review it against practical, financial and policy considerations before adoption, iterating the input assumptions if necessary.

The strength of the WASP modular architecture lies in the separation of concerns between describing the system (LOADSY, FIXSYS, VARSYS), generating and simulating alternatives (CONGEN, MERSIM) and finally optimizing and reporting (DYNPRO, REPROBAT); this modularity allows a planner to change one input, for example updating only the load forecast in LOADSY or adding a new candidate technology in VARSYS, and rerun only the affected downstream modules rather than repeating the entire study from scratch, which considerably speeds up sensitivity analysis around key planning assumptions.

In applying WASP to a real system, the planner typically performs the full LOADSY-through-REPROBAT sequence multiple times under different scenarios, for example varying fuel price assumptions, discount rate, or the reliability criterion threshold, comparing the resulting optimal plans to understand how sensitive the recommended expansion sequence is to these uncertain inputs; a plan that remains broadly similar across a wide range of reasonable assumptions is generally considered more robust for actual investment decision-making than one that changes drastically with small changes in assumptions.

Although newer, more flexible expansion-planning tools using mixed-integer linear/stochastic programming have since been developed, WASP remains widely referenced and taught because its clear modular structure closely mirrors the logical steps that any generation-expansion study must follow - describing the system, generating alternatives, simulating their performance, and optimizing the selection - making it an effective conceptual model for understanding least-cost planning even where a utility ultimately uses a more modern software package for its actual studies.

The MERSIM module deserves particular attention because it is the computational core that connects system description to system performance: by convolving the probability distribution of forced outages of every generating unit in a candidate configuration against the load-duration curve for that year, MERSIM produces the probabilistic production-cost and reliability indices that DYNPRO subsequently uses to compare configurations; the quality of the final optimized plan is therefore only as good as the accuracy of the forced-outage-rate and load-duration-curve data supplied to MERSIM.

Historically, WASP-based studies were prepared by national or regional planning agencies (with support from bodies such as the International Atomic Energy Agency, which originally developed the package for nuclear and other generation planning studies in developing countries) and were used as the technical basis for long-term national generation expansion programmes, illustrating how a single well-structured software package can support planning decisions at the scale of an entire national power system.

Overall, the WASP methodology exemplifies the general principle underlying all optimal generation expansion planning: reduce a very large combinatorial decision problem into a structured sequence of manageable computational stages, use probabilistic simulation to honestly capture the effect of equipment unavailability on cost and reliability, and use a rigorous search technique such as dynamic programming to guarantee that the final recommended plan is close to the true least-cost solution given the specified input assumptions and constraints.