Artificial intelligence entropy generation modelling of steam turbine control valves at extremely low steam mass flow

Abstract Technological and ecological environmental regulations in the EU are increasingly focused on reducing greenhouse gas emissions. Consequently, older coal-fired thermal power plants are being compelled to cease operations or switch to more environmentally friendly primary energy sources. By replacing coal technology with gas technology, the production of steam, which drives the existing steam turbine, is significantly reduced. The existing coal boiler is replaced with a heat recovery steam generator that utilises the exhaust heat from the gas turbine, thus reducing the nominal steam input to the steam turbine from 64 kg/s to only 10 kg/s. This raises doubts about the feasibility of retaining the existing steam turbine, which is designed for higher flows. To address this, a simulation model has been created using artificial intelligence and real process data. The model calculates entropy generation and consequently the loss of available work energy (exergy) at the inlet throttling valves of the steam turbine under extremely low flow conditions. The generated entropy is closely related to the losses of available work energy or exergy. The model results show that due to the extreme throttling of the steam passing through the inlet control valves, exergy losses can amount to up to 1.12 MW. This is work energy permanently lost, which could otherwise have been used for electricity production.