Search

Solar and wind based electric utility power sources have grown rapidly. The power output by these sources is transient, intermittent, and occasionally unpredictable. An integrated power grid management strategy based on multiple sources including natural gas and coal burning plants is essential to compensate for the intermittent power output of renewable resources. This strategy utilizes conventional fossil fuel power plants that operate under variable load conditions (defined as cycling). Cycling of power plant systems results in suboptimal efficiencies and exacerbates excessive wear, fatigue, and creep of the system components. The resulting financial burdens are overwhelming conventional power plants putting in jeopardy the security and the stability of the U.S. energy grid. In an effort to understand the impact of cycling in both physical and monetary aspects, a hybrid physics based and data driven model is developed. The model helps in assessing the performance and in estimating the costs of cycling operations. For the first time, transient energy and exergy analyses are conducted of a steam generator and its components in a real powerplant under cyclic load conditions. A steam generator system located at the Coal Creek Station (CCS) power plant in North Dakota, USA is used as an example for assessing the effects of cycling on the efficiencies. The analyses utilize data with one-hour resolution collected over a six-year period from 2014 through 2019. The exergy analyses lead to the identification of the lowest-cost operating conditions at all loads. The exergy analyses include a combination of energy and entropy analyses. The lowest-cost operating conditions are not identified by the energy analyses alone. The results of the present optimization analyses show that cycling decisions based on minimizing exergy destruction may lead to significant reductions in the future costs of power generation and to lead to profitability at all loads. A cost analysis was implemented to determine the cost of exergy destruction. The cost of exergy destruction aids in determining the cost of inefficiencies (loss of usable energy) resulting from cycling operations. Costs of cycling are formulated and determined for all cycles that occurred for a steam generator system located at CCS from 2014-2020. The 2014-2019 data were used for training and the 2020 data were used exclusively for testing the proposed surrogate Artificial Neural Network (ANN) model for the physics-based cost model. The ANN model is trained using the estimates from the physics-based model to estimate the costs with limited information. The ANN model was trained and validated with the physics-based model using cycles from 2014-2019 and was tested using the cycles in 2020. The overall approach is recommended for implementation at fossil fueled power plants to assess past cycling costs via the physics-based exergy efficiency model and to plan for future cost-effective cycling operations via the ANN model.