Coupled approach combining pinch analysis, advanced exergetic analysis, and optimisation for the improvement of energy efficiency in industrial processes : contribution of expert systems

The European Climate Law aims to achieve carbon neutrality by 2050, with a focus on decarbonising industry. A cornerstone of CO2 emissions reduction, improving energy efficiency consists of the retrofit of industrial processes to make them more energy efficient and then promoting the connection of these processes to renewable and recoverable energies, which are considered less intensive than fossil fuels. To improve existing processes in this direction, it is necessary to develop tools to carry out advanced energy diagnostics of industrial processes in order to develop energy efficiency solutions and facilitate decision-making relating to their industrial implementation. The COOPERE (COmbine Optimisation of Processes, Energy Recovery and Exergy analysis) approach, developed by Gourmelon et al. (2017) at the Laboratoire de Génie Chimique, provides engineers with a systematic methodology for identifying and reducing inefficiencies in industrial processes. Based on a combined use of the pinch method and exergy analysis, this approach composed of three stages; The first consists of data collection and process modelling. The second stage is based on a pinch analysis of the process to retrofit the heat exchanger network without modifying the internal structure of the process. Finally, the third stage combines exergy analysis, pinch analysis and multi-criteria optimisation to propose solutions for retrofitting the overall process (core of the process and exchanger network). Although very promising, this approach still has some limitations that may hinder its application in an industrial context. These limitations mainly concern the third stage mentioned above. While the exergy analysis developed by Gourmelon et al. (2017) is very effective in quantifying and characterising the exergetic inefficiencies of each section of the process, it still relies on an exhaustive treatment of all the inefficiencies identified, significantly complicating the subsequent optimisation phase and increasing the CAPEX of the proposed solutions. On this point, the need to prioritise improvement actions based on a more detailed analysis of the irreversibilities observed in the process appears essential in order to make the approach more pragmatic. Moreover, the proposal for a structural modification of the process aimed at reducing these inefficiencies often remains delicate because it requires solid expertise on the part of the engineer. The work presented as part of this thesis aims to supplement the COOPERE approach with a methodology combining Advanced Exergetic Analysis, Rule-based Reasoning and Optimisation to overcome these limitations. Firstly, advanced exergetic analysis allows for more precise characterisation of irreversibilities; a generic approach has been implemented to evaluate avoidable/unavoidable irreversibilities on the one hand and endogenous/exogenous irreversibilities on the other, thus allowing for a more relevant prioritisation of the process sections to be optimised. These conclusions are then used to feed an expert reasoning system based on rules and coupled with an ontology aimed at assisting the engineer in proposing several alternatives for structural modifications to the process. Finally, a multicriteria structural and parametric optimisation approach based on a meta-heuristic and integrated into the simulator is used to determine a set of several alternatives (process structure and operating parameters) that would minimise criteria such as maximum recoverable energy, electricity consumption, investment cost, etc.