Author ORCID Identifier
Semester
Summer
Date of Graduation
2024
Document Type
Dissertation
Degree Type
PhD
College
Statler College of Engineering and Mineral Resources
Department
Chemical and Biomedical Engineering
Committee Chair
Debangsu Bhattacharyya
Committee Co-Chair
Xingbo Liu
Committee Member
Xingbo Liu
Committee Member
John Hu
Committee Member
Stephen E. Zitney
Committee Member
Yuhe Tian
Abstract
Coal is currently the third-largest source of electricity production in the U.S. However, renewables are projected to be the primary source of electricity production in the future. Though coal-fired power plants (CFPPs) are rapidly retiring in the U.S., many CFPPs worldwide are expected to stay operational for a while due to factors like easy and cheap access to comparatively inexpensive coal. Therefore, non-renewable sources like coal are still expected to be prevalent and be used to ensure the stability of the electric grid in the foreseeable future. The existing coal-fired power plants generally designed to operate at base-loaded conditions are subjected to load-following conditions. Damage to equipment in power plants by corrosion is a leading cause of equipment failure and subsequent forced outage. This work primarily aims to estimate equipment corrosion and its remaining useful life using a nonlinear state estimator across multiple sensor networks. Subsequently, economic analysis is conducted considering market elasticity to assist in decision-making for sensor network placement investments. One key boiler component that undergoes corrosion damage is the waterwall (WW). The high metal temperature in the WW section of supercritical pulverized coal-fired (SCPC) power plants, coupled with the presence of corrosive gases like SO2 and O2, leads to equipment damage primarily by Type II hot corrosion. The key factors influencing the corrosion rate are identified, and a mechanistic model is developed to calculate the thickness of the fireside corrosion scale formed due to the hot corrosion in the waterwall section of the boiler in CFPPs. The proposed corrosion model is validated using the electrochemical corrosion sensor data collected from an operating industrial boiler. As corrosion dynamics are significantly slow compared to temperature and gaseous species concentration dynamics, a multi-scale model is developed where data-driven models are developed for the algebraic states that affect corrosion rate. Equations for the algebraic states and the differential equation of corrosion depth are used to form the differential-algebraic equation (DAE) system. The DAE system is solved by using the dynamic boundary values of the algebraic states to yield spatial and temporal profiles of all states along the WW. Noise is prevalent in almost every physical quantity measurement. Also, the novel corrosion sensors can be costly and impossible to place at all locations, especially due to the high temperature and harsh operating conditions in the WW section. This dissertation investigates corrosion estimation by placing sensors for the algebraic state variables, the measurement of which is comparatively more mature and considerably cheaper than the corrosion sensors. To this end, a nonlinear estimator is investigated. Unlike many systems that are represented by differential only or differential-algebraic equations, in this particular system, differential equations do not affect the algebraic states, but the algebraic states affect the differential states. A modified unscented Kalman filter (UKF) algorithm is developed for nonlinear state estimation of this specific class of differential-algebraic equations system. The performance of the UKF algorithm is evaluated with different configurations of the sensor network by simulating the boiler with various operating conditions, as expected during load-following operation. iii Implementing condition-monitoring-based predictive maintenance strategies increases power plant’s availability by preventing forced outages. Utility owners would like to ensure a positive return on investment when investing in sensors in ailing coal-fired plants. This dissertation investigates the economic impact of corrosion monitoring on equipment in coal-fired power plants through sensor placement for measuring corrosion and operating conditions like metal temperature and concentrations of O2 and SO2. The estimated time taken for the metal surface to rupture due to corrosion is calculated under multiple distinct corrosion networks. The performance of each network is analyzed by comparing the corresponding estimated time to failure to the true time taken to failure. The electricity produced by coal-fired power plants in the future in the U.S. under uncertainties in the energy market is calculated using an energy system model, The Integrated MARKAL-EFOM System (TIMES), and the database EPAUS9rT. Changes in energy trends, like the increased availability of coal-fired power plants due to sensor placement, are incorporated into the database through scenarios. All technologies and corresponding factors affecting coal-fired power plant’s electricity production are identified, and the individual degree of impact of each of these technologies on electricity production is demonstrated. A set of alternate energy futures, which include near-random scenarios involving key technologies, are created using the Latin Hypercube Sampling. The electricity produced from coal-fired power plants with and without improved availability under alternate futures is calculated and compiled to estimate the improved revenue gained by plants due to the anticipated reduction in forced outages due to corrosion monitoring. The initial investment cost for sensors under each network is calculated and compared to the improved revenue gained to calculate the change in net present value (NPV) under the increased availability of coal-fired power plants. The sensitivity of the change in NPV to factors like the type and cost of the sensor, the fraction of coal-fired power plants with sensors installed, and the number of operating plants in the installation year are analyzed.
Recommended Citation
Somayajula, Chandra Sekhar, "State Estimation and Economic Analysis for Electrochemical Sensor-Based Corrosion Monitoring" (2024). Graduate Theses, Dissertations, and Problem Reports. 12571.
https://researchrepository.wvu.edu/etd/12571
Embargo Reason
Publication Pending