Are you an EPFL student looking for a semester project?
Work with us on data science and visualisation projects, and deploy your project as an app on top of GraphSearch.
This thesis presents a methodology for energy management in large companies and its implementation through a web application and through a prototype of a simulation platform. By combining existing tools in an innovative manner and by making use of recent web technology developments, the methodology adopted provides engineers and managers with tools capable of guaranteeing an efficient and sustainable energy management. Although the methodology presented in this work is based on the experience acquired in the food industry, it can be easily applied in other industrial sectors. The methodology is based on two fundamental approaches commonly used to analyse energy consumption in industrial contexts: the top-down approach and the bottom-up approach. The top-down approach is used in the first place to identify the factories and the specific areas within the factories in which the largest improvement potentials can be achieved. In turn, the bottom-up approach builds on the results from the top-down approach to identify and quantify the energy saving potentials. The top-down approach is implemented through a web application in collaboration with an industrial partner. This application encompasses a modular factory model –accessible to engineers in factories through a user-friendly interface– which enables each factory to define its energy usage, allocate energy costs among the different energy consumers and compute key performance indicators. For a rational cost allocation in multi-service energy conversion units, an exergy-based methodology is presented. The efficiency of energy conversion units defined in the factory model, such as the boilerhouse or the air heaters, is assessed using thermodynamic models. The latter are simplified parametric models derived from accurate thermodynamic models developed in a general flow-sheeting and simulation software to comply with computation time and reliability requirements of the web application. The different factory models defined in the web application can be browsed as part of the proposed top-down approach: starting from a high level overview of the factory –targeted mainly at managers– users can then focus on a specific area of the factory. Strategies are developed to guide users in identifying factories or specific areas within the factories with the largest improvement potentials. They include the use of mechanism to rate the quality of a performance indicator as well as a benchmarking module that allows to compare performance indicators across factories worldwide. In sum, the modular and adaptive aspects of the web application guarantee its long-lasting use. In order to quantify energy saving potentials in the energy conversion units defined in a factory model, "what if?" scenarios are performed in a web-based simulation platform prototype developed in this thesis. This platform acts as a decision-support tool by providing graphical representations of profitability and risk analysis. The platform can be accessed by human users through a web browser while other applications, such as the web application described above, may use the simulation functions through a web service. Statistical tools that can help engineers in defining the factory model described above are also presented. They are used to correlate energy consumption with factors such as production volumes or the climate. Tests to validate the developed correlations are also described. The application of this technique in a factory shows that more than 50% of the energy consumption does not have a direct correlation with production factors and allows to identify improvement potentials. Finally, the concept of a bottom-up approach to identify and quantify energy saving potentials in the different production processes of a factory is presented. A triple representation of the requirements of a process is introduced and applied to process integration in a concrete example. The 80/20 rule is also applied to reduce the complexity of the problem. The optimal integration of cogeneration engines and heat pumps using multi-objective optimisation is also presented.