August 1, 2011

Make Your Plant Ready for Cycling Operations

By Steven A. Lefton and Douglas Hilleman, PE, Intertek-Aptech

Cost-of-Cycling Estimates

The Intertek-Aptech staff has developed two approaches to developing cycling cost estimates, such as the one summarized in Table 2, so that results can be compared and validated: the top-down estimate and the bottom-up estimate. Unit- and plant-specific information and industry data on similar units are also used in these analyses.

Top-Down Estimates. Top-down estimates use historical unit operating data and historical cost data to determine the costs of cycling operations (hot, warm, and cold starts and shutdowns; load following; and ancillary services such as regulation and ramping). Fundamentally, we take wear and tear costs and other cycling-related costs and statistically determine the costs of cycling using detailed multivariable regression techniques that examine cost versus total cycling damage.

The first step to determining operating costs is to examine the actual validated plant maintenance costs, primarily found in work orders. The costs we take into account include the operating, maintenance, and capitalized maintenance costs incurred by the unit, while those unrelated to equipment maintenance are eliminated. Other cycling-related costs—such as the cost of fuel and chemicals for water treatment used for startups, and costs related to the increased outages caused by cycling—are also accounted for. These costs are then analyzed, processed, and tallied to create annual “candidate” cycling costs for the unit. All of these costs are candidates for our analyses, which determine the relationship between costs and the unit’s total cycling operational damage.

The second step is to add cycling damage to the maintenance cost estimates. The damage the unit accumulates from cycling is determined by examining the plant’s operating records. Specifically, the hourly average power output of the unit’s generator is analyzed to count cycles (all types of cycling and load-following) and determine the historical damage that the unit has accumulated versus baseload operation. We calibrate the damage during load transients and starts to plant signature data obtained during typical transients and starts.

Finally, we take the accumulated damage and historical costs to calculate a statistical “best estimate” of the cycling costs and calculate the upper and lower bounds using statistical regression techniques. In sum, we develop probabilistic estimates of the effects of cold, warm, and hot starts; shutdowns; and load-following operating modes.

We prove these estimates are valid by backcasting historical costs as a function of cycling costs. We know past costs are the best predictors of future costs when unit cycling remains constant. However, a significant increase in cycling and ramp rates can significantly accelerate equipment damage and increase future costs.

Bottom-Up Estimates. The other cost-estimating process is the bottom-up analysis. It is referred to as a bottom-up because it starts with the detailed work order history and a review of failure events and analysis completed earlier. The bottom-up review includes seven to 10 years’ worth of plant work orders, when available. The review includes a detailed analysis of work orders by a subject matter expert (SME), often including the plant personnel, to classify the costs as either related to cycling or to normal plant (baseload operation) costs. Actual failure reports, overhaul records, and prior inspection reports are also checked to determine the root cause of previous failures so that the costs can be properly classified.

This analysis uses actual plant signature data from representative starts and shutdowns and gives consideration to known design and operations issues. Load-following and actual plant ramp rates are examined, including the absolute temperature change from carefully selected data points to validate the type of damages.

Comparing Cost Estimates

The result of these cost analysis approaches are cost estimates for various plant systems and sub-systems that identify the actual (usually the minimum) cost of cycling (based on recent historical costs). Those estimates can be used to predict future cycling costs. These cost estimates may be adjusted by latent damages factors determined by the SMEs, based on failure rates, inspections, interviews, or signature data. The estimates identify high-risk failures without detailed condition assessment inspections of many components. The estimates may drive limited specific condition-based inspections to validate the risk of failure or time for replacement. The estimates also include specific countermeasures with a valid cost-benefit at several levels with a return on investment for budget planning.

When the two cost estimates are completed, the customer has a report of detailed operating costs with recommended design improvements, equipment upgrades/replacement, and operating process improvements. The plant owner has hard data that specifies the costs of cycling the plant and a plan for managing cycling costs in the future with hardware and operating plant improvements. These plant improvement costs are usually ranked by cost-benefit and implementation costs. Some operating process or procedure improvements can be realized at little or no cost. Usually, those plant modifications with the highest benefit-cost ratio are quickly put into the company’s capital planning cycle.

The electricity market has appreciably changed over the past decade, especially with the introduction of large amounts of nondispatchable wind and solar power in some regional markets. Cycling a plant is not necessarily a bad practice. But the decision to do so should be made by an owner who has full knowledge of all the available options and estimates of the real costs that must be paid, today or in the future, as a result of that decision. Give plants the tools they need to manage cycling costs, and they will produce winning results.