Model-based fault diagnosis for hybrid systems: Application on chemical processes

The complexity and the size of the industrial chemical processes induce the monitoring of a growing number of process variables. Their knowledge is generally based on the measurements of system variables and on the physico-chemical models of the process. Nevertheless, this information is imprecise because of process and measurement noise. So, the research ways aim at developing new and more powerful techniques for the detection of process fault. This article presents a method for the fault detection based on the comparison between the reference model evolution and the real system generated by the extended Kalman filter. The reference model is simulated by the dynamic hybrid simulator, PrODHyS. It is a general object-oriented environment which provides common and reusable components designed for the development and the management of dynamic simulation of industrial systems. The use of this method is illustrated through a didactic example relating to the field of Chemical Process System Engineering.     

Optimal planning and scheduling of batch plants with operational uncertainties: An industrial application to Baker's yeast production

In this work, the mixed integer linear programming (MILP) model developed in Orçun et. al 1996 for optimal planning and scheduling of batch process plants under uncertain operating conditions is further improved to deal also with discrete probability functions. Furthermore, the logic behind integrating the processing uncertainties within the MILP model is implemented on the variations in the production volumes that can be faced in some batch processes such as Baker's yeast production. The modified model is tested on Baker's yeast production plant data to illustrate the effect of uncertainties on the production planning and scheduling. The results show that the plant production will be improved by 20% when the optimal production planning and scheduling is utilized by fine tuning the degree of risk the management can resist. An example on how a process design engineer may utilize such an MILP model for optimal planning and scheduling of batch process plant and identify plant problems, such as the bottleneck operations, is also included. A simulation type analysis on how to improve the processing site, i.e. the effect of introducing an extra operator to the bottleneck operation, is also demonstrated in this work using the available plant data.     

Dynamic simulation and control of gas turbines and compressor systems

A dynamic simulation program has been developed to model processes involving fluid compression, expansion, heat generation and heat transfer. Controller design procedures are embedded in the computer code. This allows easy transfer of information between simulation, used to sample the process response to various perturbation, and optimization of controller design and tuning. Equations are developed for some equipment models, based on 1-D flow approximation. The architecture of the simulation package based on the object paradigm is described. The program has been validated on test cases. In particular, compression stems have been evaluated for different strategies.     

Sequencing a generalized two-stage flowshop with finite intermediate storage

The scheduling of batch chemical facilities may be modeled as a generalized flowshop problem, for which the objective is to minimize the products' completion time. In particular, this paper examines the scheduling difficulties that arise if interstage storage is used within the production facility. A branch and bound solution procedure is presented which uses a simulation model to evaluate sequence compeltion times. An initial upper bound completion time is determined based on a heuristic scheduling rule.     

Parameterization of population balance models for polythermal auto seeded preferential crystallization of enantiomers

For analysis, design and model-based control of crystallization processes typically population balance models or reduced models derived therefrom are used. Usually the kinetic parameters in these models are determined analyzing measured concentration trajectories and/or particle size distributions using parameter estimation procedures. In the case of preferential crystallization of enantiomers the analysis of experiments is complex since there are two “competing” crystal populations. In this field often batch processes are performed using seeds of the desired enantiomer. Currently, it is in particular challenging to quantify and optimize a new concept: e.g. the so-called “auto seeded programmed polythermal preferential crystallization” (“as3pc” [Coquerel, G., Petit, M.-N., Bouaziz, R., 2000. Method of resolution of two enantiomers by crystallization. United States Patent, Patent number: 6,022,409]). In order to design and optimize this process the temperature dependent kinetic constants for crystal growth, nucleation and dissolution have to be known. In this work a reduced model for this auto seeded process is presented. The general identifiability of the model parameters is investigated along with some suggestions on how to reparameterize the kinetic terms involved. The values of the identified key parameters are estimated using conventional least square optimization using experimental data determined for the model system threonine/water. Parameter confidence and cross correlation are discussed and finally the model is validated using experiments not used for parameter estimation.     

Experimental validation studies on a multi-dimensional and multi-scale population balance model of batch granulation

In this study, a dynamic model is presented for the granulation process, employing a three-dimensional population balance framework. As a first attempt to account for the multi-scale character of the process, the nucleation and aggregation kernels used in the population balance model are derived using mechanistic representations of the underlying particle physics such as wetting kinetics and energy dissipation effects. Thus, the fundamental properties of the powder and the liquid were used as parameters in the model to predict the granulator dynamics and granule properties. The population balance model is validated against experimental data from a calcite/PVOH-H₂O recipe obtained using a lab-scale drum granulator for granule size, fractional binder content and porosity. A reasonably good agreement between experimental and simulation results were obtained for the granule size distribution under different experimental conditions. In addition, accurate model predictions were made for the evolution of the average properties (i.e., size, fractional binder content and porosity) for various operating conditions.     

Global Reaction Kinetics of CO and CO2 Methanation for Dynamic Process Modeling

For the design and optimization of methfanation processes detailed modeling and simulation work is advisable. However, only a few kinetics published in literature rely on wide temperature and pressure ranges, which are prevalent at modern methanation applications with dynamic operation. Especially the simulation‐based design of methanation processes with commercial catalysts is difficult due to legal restrictions regarding the publication of kinetic data of those catalysts. In this work, rate equations for the dynamic modeling and simulation of methanation processes operating with commercial Ni/Al₂O₃ catalysts are selected, adapted, and tested in a dynamic reactor model. The results suggest that the catalyst's nickel content is an indicator for the choice of a rate equation. Testing of the equations in a reactor model meets published data for CO and CO₂ methanation and own measurements.     

多回路网络控制系统的模糊动态调度

本文首先阐述了采样周期对于网络控制系统性能的影响,然后针对多回路网络控制系统中,回路之间争抢网络资源,导致系统性能恶化的情况,在原有EDF调度算法的基础上,提出含有模糊调度器的多回路模糊动态调度算法。该算法根据系统中各回路的误差和误差变化率,利用模糊控制的方法实时调整各回路的优先级,从而实现对网络控制系统的调度。最后,利用TrueTime工具箱建立了包含模糊动态调度器的网络控制系统仿真模型,并将其与无调度器的网络控制系统进行对比。仿真结果表明,本文所设计的多回路模糊动态调度算法能够合理地安排各回路使用网络的先后顺序,具有较好的控制性能。     

A nested online scheduling and nonlinear model predictive control framework for multi-product continuous systems

In the pursuit of integrated scheduling and control frameworks for chemical processes, it is important to develop accurate integrated models and computational strategies such that optimal decisions can be made in a dynamic environment. In this study, a recently developed switched system formulation that integrates scheduling and control decisions is extended to closed-loop operation embedded with nonlinear model predictive control (NMPC). The resulting framework is a nested online scheduling and control loop that allows to obtain fast and accurate solutions as no model reduction is needed and no integer variables are involved in the formulations. In the outer loop, the integrated model is solved to calculate an optimal product switching sequence such that the process economics is optimized, whereas in the inner loop, an NMPC implements the scheduling decisions. The proposed scheme was tested on two multi-product continuous systems. Unexpected large disturbances and rush orders were handled effectively.     

Nonsmooth model for dynamic simulation of phase changes

Dynamic modeling of processes involving phase changes can be challenging due to changes in the model equations caused by appearance and disappearance of equilibrium phases. Dynamic simulation of these processes requires the ability to detect the change in the number of phases and adapt the model to the new phase regime on the fly. In this work, an easy‐to‐use nonsmooth model for dynamic simulation of processes with vapor‐liquid equilibrium is presented. The presented model does not introduce any auxiliary variables or equations, nor does it require solution of an optimization problem to determine the new phase regime during the dynamic simulation. It can therefore be used for comprehensive simulation of, e.g., distillation columns, where the number of phases present can change during startup and shutdown. The nonsmooth model is illustrated through examples of an evaporator and a distillation column. © 2016 American Institute of Chemical Engineers AIChE J, 62: 3334–3351, 2016     

A dynamic coupling model for hybrid atomistic-continuum computations

A dynamic coupling model is developed for a hybrid atomistic-continuum computation in micro- and nano-fluidics. In the hybrid atomistic-continuum computation, a molecular dynamics (MD) simulation is utilized in one region where the continuum assumption breaks down and the Navier-Stokes (NS) equations are used in another region where the continuum assumption holds. In the overlapping part of these two regions, a constrained particle dynamics is needed to couple the MD simulation and the NS equations. The currently existing coupling models for the constrained particle dynamics have a coupling parameter, which has to be empirically determined. In the present work, a novel dynamic coupling model is introduced where the coupling parameter can be calculated as the computation progresses rather than inputing a priori. The dynamic coupling model is based on the momentum constraint and exhibits a correct relaxation rate. The results from the hybrid simulation on the Couette flow and the Stokes flow are in good agreement with the data from the full MD simulation and the solutions of the NS equations, respectively.     

Application of gain scheduling for modeling the nonlinear dynamic characteristics of NO<Subscript><Emphasis Type="Italic">x</Emphasis></Subscript> emissions from utility boilers

A hierarchical gain scheduling (HGS) approach is proposed to model the nonlinear dynamics of NO _x emissions of a utility boiler. At the lower level of HGS, a nonlinear static model is used to schedule the static parameters of local linear dynamic models (LDMs), such as static gains and static operating conditions. According to upper level scheduling variables, a multi-model method is used to calculate the predictive output based on lower-level LDMs. Both static and dynamic experiments are carried out at a 360 MW pulverized coal-fired boiler. Based on these data, a nonlinear static model using artificial neural network (ANN) and a series of linear dynamic models are obtained. Then, the performance of the HGS model is compared to the common multi-model in predicting NO _x emissions, and experimental results indicate that the proposed HGS model is much better than the multi-model in predicting NO _x emissions in the dynamic process. This paper was presented at the 7 ^th China-Korea Workshop on Clean Energy Technology held at Taiyuan, Shanxi, China, June 25–28, 2008.     

Towards the integration of process design,control and scheduling: Are we getting closer?

The integration of design and control, control and scheduling and design, control and scheduling, all have been core PSE challenges. While significant progress has been achieved over the years, it is fair to say that at the moment there is not a generally accepted methodology and/or “protocol” for such an integration – it is also interesting to note that currently, there is not a commercially available software [or even in a prototype form] system to fully support such an activity.Here, we present the foundations for such an integrated framework and especially a software platform that enables such integration based on research developments over the last 25 years. In particular, we describe PAROC, a prototype software system which allows for the representation, modeling and solution of integrated design, scheduling and control problems. Its main features include: (i) a high-fidelity dynamic model representation, also involving global sensitivity analysis, parameter estimation and mixed integer dynamic optimization capabilities; (ii) a suite/toolbox of model approximation methods; (iii) a host of multi-parametric programming solvers for mixed continuous/integer problems; (iv) a state-space modeling representation capability for scheduling and control problems; and (v) an advanced toolkit for multi-parametric/explicit Model Predictive Control and moving horizon reactive scheduling problems. Algorithms that enable the integration capabilities of the systems for design, scheduling and control are presented on a case of a series of cogeneration units.     

Crack propagation during Vickers indentation of zirconia ceramics

A quarter finite element model of 3 mol% yttria stabilized tetragonal zirconia polycrystal (3Y-TZP) ceramics undergoing Vickers indentation was established to simulate the evolution of stress and the propagation of cracks inside a sample. The indentation experiment was carried out on the Micro Vickers Hardness Tester. The results of the geometric characteristic parameters, such as the indentation diagonal half-length a, the crack length c and the maximum indent depth h_m, from the indentation simulation and experiment were similar. The types of indentation cracks under various loads were determined according to the Lawn-Evan model, which exactly correspond to the simulation results. In addition, the propagation of indentation cracks was discussed based on the maximum principal stress contour plots at various stages, and the conclusions were verified by the indentation analysis model proposed by Yoffe. As a result, the model developed in this paper can be used in indentation studies to solve the related problem.     

Integration of multivariate SPM and FDD by parity space technique for a food pasteurization process

Multivariate statistical process monitoring (MSPM), contribution plots, and parity space fault diagnosis (FD) techniques are used to detect abnormal operation of dynamic processes and diagnose sensor and actuator faults. The methods are illustrated by monitoring the critical control points (CCP) and diagnosing causes of abnormal operation of a pilot pasteurization plant. An empirical model of the process is developed by using subspace state space system identification methods and normal process data. The process data collected under the influence of different magnitude and duration of faults in sensors and actuators are used to validate the MSPM and FD techniques. T² and squared prediction error (SPE_N) charts are used as MSPM charts. A parity space technique for dynamic stochastic systems and dynamic trends in contribution plots of T² and SPE_N statistics are used for FD. The detection and FD by these techniques show significant improvements over univariate methods.     

Design of integrated batch processes with discrete and continuous equipment sizes

This paper presents a mathematical programming approach for the integrated design of batch processes involving multiple processing steps. Detailed dynamic models are used to describe the behaviour of individual batch operations. The mathematical formulation regards both the design (e.g. equipment items with discrete sizes) and operational characteristics (i.e. control variable profiles such as reflux ratio, cooling water flowrate) of each task as degrees-of-freedom, the values of which are selected to optimise a particular objective function (e.g. profit). The solution may be restricted by inequality constraints which hold throughout the task (e.g. reactor temperature violation) or at the end of the task (e.g. end-product purity specification). The sizes of the equipment items are treated as discrete decisions and are chosen from a selection of standard sizes. The mathematical formulation leads to a mixed integer dynamic optimisation problem and is solved using the outer approximation/augmented penalty (OA/AP) algorithm.     

Control and Regulation of Integrated Mitochondrial Function in Metabolic and Transport Networks

The pattern of flux and concentration control coefficients in an integrated mitochondrial energetics model is examined by applying a generalized matrix method of control analysis to calculate control coefficients, as well as response coefficients The computational model of Cortassa et al. encompasses oxidative phosphorylation, the TCA cycle, and Ca²⁺ dynamics. Control of ATP synthesis, TCA cycle, and ANT fluxes were found to be distributed among various mitochondrial processes. Control is shared by processes associated with ATP/ADP production and transport, as well as by Ca²⁺ dynamics. The calculation also analyzed the control of the concentrations of key regulatory ions and metabolites (Ca²⁺, NADH, ADP). The approach we have used demonstrates how properties of integrated systems may be understood through applications of computational modeling and control analysis.     

A framework for schedule evaluation with processing uncertainty

The scheduling of batch chemical processes has been the focus of a substantial amount of research, focused primarily on building schedules for processes which are assumed to operate in a deterministic manner. This work describes an approach to directly incorporate schedules into a simulator for the purposes of schedule validation and testing of rescheduling methodologies when stochastic events occur. Although other types of uncertainty could be considered within the framework, the deviations presented in this work are those which can be represented as processing time variations and equipment breakdowns. The framework, when used open-loop, is an effective tool for evaluating the expected performance and robustness of various scheduling strategies. Two scheduling tools are compared, one which uses the uniform time discretized model and the other uses a non uniform discretization of time. In addition, user written logic has been incorporated into the simulator which allows the performance of various rescheduling techniques to be compared and evaluated prior to implementation in a processing facility. Numerical results of Monte Carlo simulation studies for both uses of the framework are presented.     

CFD simulation of gas–liquid flow in a high-pressure bubble column with a modified population balance model

In this study, based on the Luo bubble coalescence model, a model correction factor C_e for pressures according to the literature experimental results was introduced in the bubble coalescence efficiency term. Then, a coupled modified population balance model (PBM) with computational fluid dynamics (CFD) was used to simulate a high-pressure bubble column. The simulation results with and without C_e were compared with the experimental data. The modified CFD-PBM coupled model was used to investigate its applicability to broader experimental conditions. These results showed that the modified CFD-PBM coupled model can predict the hydrodynamic behaviors under various operating conditions.