Modélisation de processus | EPFL Graph Search

La modélisation de processus (en anglais, business process modeling ou BPM) consiste à structurer et à représenter les activités d'une organisation, généralement en utilisant une notation graphique pour représenter visuellement l'enchaînement des activités. La modélisation peut s'appuyer sur des méthodes et outils spécialisés, et mettre en œuvre des cadres de références de processus. Cette discipline utilise parfois le sigle BPM en référence au « Modélisation de procédure d'entreprise », mais cet usage tend à disparaître, car il peut être confondu avec le « Business Process Management », dont la modélisation est une des activités. La distinction fondamentale entre ces deux notions de BPM réside dans le fait que pour la seconde, on s'intéresse à donner à l'entreprise les moyens de piloter et de maîtriser ses processus-métiers, tandis que la première ne consiste qu'à les modéliser (toujours avec un objectif venant de l'extérieur : optimisation de chaîne de production, expression de be

Model-based scale-up of a continuously operated consolidated bioprocess based on a microbial consortium for the production of ethanol

David Béat Dempfle

This work demonstrated that consolidated bioprocessing is a promising concept for conversion of lignocellulose to ethanol at industrial scale. CBP offers great cost saving potential, is feasible to be operated continuously and may be scaled up due to extensive knowledge of the process from a chemical engineering point of view.Cost savings of up to 27.5% of the total costs compared to conventional bioethanol production from lignocellulose, as stated by a techno-economic assessment, make continuous CBP the strongest lever to reduce processing costs of lignocellulosic ethanol. A cost sensitivity analysis identified scale and yield as the main cost-pushers from a process point of view, whereas the price level of the plant location has the highest impact on the investment conditions.To prove the feasibility of continuous CBP, and therefore the mentioned cost savings, experiments with a maximum titer of 3.258±0.007 g/L, a productivity of 0.025 g/(Lh) and constant enzyme production over 750h were conducted. Furthermore, the continuous experiments showed that experiments with identical volumetric oxygen transfer rate k_L a, but different oxygen fluxes per membrane area, showed titer differences of ca. 80% (1.83 g/L vs 3.26 g/L) in favor of setups with the large membrane surface. A rigorous process model was developed for continuously operated CBP. 9 species were considered (oxygen, glucose, total concentration Trichoderma reesei, secondary mycelia concentration of T. reesei, enzymes, cellulose, cellobiose, yeast density and ethanol). 8 of these 9 species needed to be modelled spatially resolved to account properly for mass transfer limitations. The fungal biofilm thickness d_f was found to be a critical parameter with an optimum for every membrane configuration. Smaller d_f reduced the fungal biofilm volume and thus, the enzyme production and larger d_f increased the diffusion path length causing shortage in nutrient supply as well as lower enzyme concentrations in the bulk. The enzyme synthesis rate of the secondary mycelia was fitted by reducing it by ca. 40% (0.67 FPU/(mLd) vs 1.152 (FPU/(mL*d) compared to batch models.With the process model as centerpiece, a scale-up framework consisting of the model and a set of non-dimensional parameters was developed to scale CBP from 2.7L laboratory scale to 130L pilot scale. The majority of the non-dimensional parameters could be kept constant during scale-up as requested by the similarity paradigm. Necessary changes were simulated with the model. Again, the fungal biofilm thickness d_f proved to be a relevant parameter for scale-up considerations. In general, the relative loss of biofilm volume is compensated by longer residence times since the residence time has by far the least impact on the process economics.Finally, popular rate-controlled separation techniques were investigated regarding their potential with CBP, since they offer the potential to reduce yeast inhibition, to avoid separation limitations by the azeotrope and they are not bound to the vapor-liquid-equilibrium of highly diluted ethanol-water mixtures. However, most of the mechanisms fail to handle the solids of the fermentation broth. Pervaporation, being the most promising concept for in-situ product removal, would be a cost-saving alternative to distillation for batch operation, but is limited by an unfavorable vapor-liquid equilibrium due to low bulk concentrations during continuous operation.

EPFL2022

Methods and Performance Assessment of PMU-based Real-Time State Estimation of Active Distribution Networks

Styliani Sarri

The evolution from passive to Active Distribution Networks (ADNs) is producing large changes in the operation of these electrical systems. In particular, violations of grid operational constraints, higher dynamics and limited amount of controllable resources represent main limiting factors in the optimal operation of ADNs in presence of massive stochastic distributed generation. In order to deal with these issues, the emergence of ADNs requires the definition of suitable Energy Management Systems to achieve specific operation objectives (i.e., optimal voltage/congestion controls, updated protection schemes etc.). These functions are significantly improved if the system state is known with high accuracy, high refresh rates and low time latencies. Unfortunately, typical refresh rates of traditional State Estimation (SE) processes designed for transmission networks are in the order of few minutes, whereas the time frames of the above functionalities are between few milliseconds to few seconds. Hence, it becomes necessary to define, develop and validate the three-phase Real-Time State Estimation (RTSE) processes characterised by high refresh rates (i.e., in the range of several tens of estimation per second), small latencies (i.e., in the range of few tens of ms) and high accuracy. In this direction, the ADNs SE is facilitated by the emerging technology of Phasor Measurement Units (PMUs) which allow acquiring accurate, time-aligned phasors with typical streaming rates in the order of some tens of f.p.s. Additionally, PMUs measurements of synchrophasors allow formulating the SE problem in a linear way. PMU measurements can be acquired and stored in a real-time (RT) database, provided by Phasor Data Concentrators (PDCs) suitably coupled with the RTSE. This enables, in theory, the performance assessment of the whole RTSE chain. However, the assessment of the RTSE accuracy with real PMUs in a real grid is impossible since the true state is hidden. It is, yet, possible to overcome this limitation by using a Real Time Simulator (RTS) and design a RT setup that allows knowing the true state. This RT setup should be GPS-synchronized, to enable the RTSE accuracy and time latencies assessment. Within the above context, this thesis focuses on the definition of SE methods together with their formal and numerical performance assessment. In particular, the first part of the thesis discusses the formulation of static (e.g., weighted least squares - WLS) and recursive (e.g., Kalman Filter - KF) algorithms fed by synchronized phasors and/or other traditional measurements provided by remote terminal units. Then, the thesis discusses the formal comparison of the accuracy of KF vs. WLS and proves that the former behaves better if its process model is correct. For the case of linear SE, the thesis discusses the method for the verification of the exactness of the so-called measurement noise covariance matrix. The subsequent part of the thesis provides the numerical validation and performance assessment of the RTSE process via offline simulations. This analysis is conducted by using IEEE benchmark distribution and transmission networks as well as real distribution feeders. The last part of the thesis focuses on the experimental validation of the RTSE chain via an experimental RT setup. In this last part, the thesis describes the structure and the individual components simulated in the RT experimental setup as well as the whole validation procedure.

EPFL2016

Integrating Rate Based Models into a Multi-Objective Process Design & Optimisation Framework using Surrogate Models

Levent Sinan Teske

In the development of energy and chemical processes, the process engineers extensively apply computer aided methods to design & optimise these processes and corresponding process units. Such applications are multi-scale modelling and multi-objective optimisation methods. Multi-objective optimisation of super-structured process designs are expensive in CPU-time due to the high number of potential configurations and operation conditions to be calculated. Thus single process units are generally represented by simple models like equilibrium based (chemical or phase equilibrium) or specific short cut models. In the development of new processes, kinetic effects or mass transport limitations in certain process units may play an important role, especially in multiphase chemical reactors. Therefore, it is desirable to represent such process units by experimentally derived rate based models (i.e. reaction rates and mass transport rates) in the process flowsheet simulators used for the extensive multi-objective optimisation. This increases the trust engineers have in the results and allows enriching the process simulations with newest experimental findings. As most rate based models are iteratively solved, a direct incorporation would cause higher CPU-time that penalises the use of multi-objective optimisation. A global surrogate model (SUMO) of a rate based model was successfully generated to allow its incorporation into a process design & optimisation tool which makes use of an evolutionary multi-objective optimisation. The methodology was applied to a fluidised bed methanation reactor in the process chain from wood to Synthetic Natural Gas (SNG). Two types of surrogate model, an ordinary Kriging interpolation and an artificial neural network, were generated and compared to its underlying rate based model and the chemical equilibrium model. The analysis showed that kinetic limitations have significant influence on the result already for standard bulk gas chemical components. A case study applying the previous version of the process design model and the revised version (with rate based model introduced as a set of five surrogate models) will demonstrate that the prediction uncertainties of the process design & optimisation methodology are reduced due to the integration of the rate based model of the fluidised bed methanation reactor. It will be shown that the different process design models predict considerably different optimal operating conditions of the Wood-to-SNG process. This emphasises the importance of the integration of rate based models into the process design models. The presented approach has been developed for the fluidised bed methanation reactor, however, it is a generic approach which can be applied to other process unit technologies as well. Future investigations will target other technologies to further improve the process design & optimisation predictions and support project development.

EPFL2014