<p>Thermal power plants employ regenerative type air pre-heaters (APH) for recovering heat from the boiler flue gases. The APH gets clogged due to deposition of ash particles and products formed by reactions between leaked ammonia from the upstream selective catalytic reduction (SCR) unit and Sulphur oxides (SOx) present in the flue gases. Clogging depends strongly on the APH operating conditions, concentrations of ash and Sulphur oxides present in the flue gases, and the ambient conditions. The gas pressure drop across the APH is the main indicator of clogging. It increases gradually over time, resulting in increased power consumption by the induced draft fan (IDF). Soot blowing is commonly used to clean the APH, but it does not completely remove the deposits, ultimately leading to forced outages. There is a need to have a system that can predict clogging or fouling in advance so that appropriate actions can be taken not only to reduce power consumption but also to design maintenance activities to avoid forced outages.</p><p>We present a deep learning based model for forecasting the gas pressure drop across the APH using the Long Short Term Memory (LSTM) networks. It utilizes real-time sensor data from the APH and the upstream boiler and the SCR units along with coal properties data from laboratory analyses and the ambient conditions. The data is transformed in an appropriate manner for feeding into the LSTM network which utilizes data from the past 30 to 60 days and forecasts gas pressure drop across the APH for the next 15 to 30 days. The forecasting model employs soft sensors for key inputs such as ammonia leakage from the SCR and ambient temperatures, to extend the forecasting period up to 3 to 4 months which enables designing maintenance schedules optimally to avoid forced outages. The model was trained and tested with data from industrial scale APHs. It can also be deployed as part of a digital twin of the APH not only to provide real time visibility and optimizing power consumption by IDF, but also to avoid forced outages of the plant due to unpredictable fouling.</p>

Forecasting of Fouling in Air Pre-Heaters Through Deep Learning

Paper Proceedings: https://asmedigitalcollection.asme.org/POWER/POWER2021/volume/85109

ASME 2021 Power Conference (POWER2021)

Supporting the need among researchers worldwide for additional online and multimedia content, ASME now offers the ASME Conference Video Collection. This platform will collect and provide long-term access to conference presentation videos, encompassing all technology areas within ASME's technical conference and events program. About 30,000-40,000 engineers across all mechanical engineering and related disciplines participate in ASME conferences each year. This includes authors of approximately 6,000 to 8,000 technical papers published by ASME, the majority of whom also submit video presentations for our virtual conferences. The video collection will be available as an add-on to existing product packages or as a separate subscription.