| Use of Model Process Systems to Devise and Evaluate Improvements in Process Analysis and Control |
|
Project
Objectives
|
|
The objectives of CPACT project 1 were to set up a facility, which could be used to:
Two reactions were chosen as model processes to be used for (a) – (c) above, the esterification of crotonic acid and 2-butanol, and the esterification of itaconic acid and 1-butanol. The first of these two reactions is a homogeneous reaction using toluene as a solvent. The second reaction poses challenges as it is a heterogeneous reaction (itaconic acid is insoluble in the toluene solvent), and itaconic acid is a dibasic acid leading to the formation of a monoester and a diester, with the monoester the preferred product. Achievements A batch reactor facility has been constructed at Strathclyde. This consists of 2 triple jacketed 5 L reactors, each capable of operating as a batch reactor. Sensors allow the temperature, oil jacket inlet and outlet temperatures, stirring rate, pressure, reaction level, mass of reactants in vessels and condenser fluid temperature to be continually monitored. The temperatures of the oil and reactor, stirring rate and feed rate of reactants can be controlled. A new control program has been written in LabVIEW. The new control program is more flexible than the one provided with the reactors, and will allow any information derived from analytical data to be used for control. Analytical instrumentation that has been obtained during the course of the project include a Kaiser HoloPROBE Raman spectrometer, a Bomem MB155 FTIR/NIR spectrometer, a Zeiss MCS UV-visible spectrometer, a Foss NIRSystems on-line 6000 NIR spectrometer, and a Resonance low-field NMR spectrometer (see Project 3). Methods for in-line analysis using Raman, near infrared and UV-visible spectrometries have been developed and applied to monitor the esterification of crotonic acid and 2-butanol. A method for at-line analysis by low-field 1H NMR has also been developed in collaboration with CPACT Project 3. Comparison studies of the different techniques for esterification reaction monitoring have been carried out. Programs have been written to take spectroscopic data from the Raman, Bomem NIR and Foss NIR spectrometers into the LabVIEW control system. Once in LabVIEW, chemical information can be obtained from pre-made calibration models. The next step will be to allow this chemical information to be used for control purposes, detect anomalies, study kinetics etc. Kinetic studies have been carried out on the crotonic acid/2-butanol esterification reaction which have allowed the reaction to be modelled. A sampling loop has been designed and constructed which will allow on-line NMR measurements to be made. Novel pH probes have been developed. The sensors are made from thermally grown iridium oxide and are more robust than traditional glass pH electrodes. Faster response to changes in hydrogen ion concentration in non-aqueous solvents has been demonstrated, compared with measurements obtained using a glass electrode. At the University of Hull, work has been carried out to determine the potential of Raman spectrometry for monitoring the acid catalysed esterification of ethanoic acid and ethanol, and to assess the potential of microwave spectrometry for determining moisture and fat contents of tobacco and foods. The deliverables from the project have been 5 IMB reports (with another 6 in preparation), 4 oral presentations at conferences (outwith CPACT), 4 poster presentations (outwith CPACT), and 3 publications in peer reviewed journals. Other papers are in preparation. Objective (a) has been achieved. The reactor facility was commissioned in 1998 and analytical instrumentation has been acquired. This has allowed objective (b) to be completed. Work on objective (c) has mainly been carried out by CPACT Project 2, with the transfer of knowledge from Project 2 to Project 1 being less than originally anticipated. Objective (d) has partly been achieved with the design of the LabVIEW control system, however more work is required in this area. Deliverables On-Line Data Communication between
NIR-Bomem Grams and LabVIEW Process Control Software Packages This report describes the work carried out on a third year summer project, by a Computer and Electronic Systems student, Ian Power. Ian was placed at the CPACT reactor facility for two months, during which he developed the software link, and this report is based on his project report submitted for his undergraduate credits. It describes the project environment, development of software programs and the results achieved. It also gives recommendations for further work. LabVIEW Monitoring and Control
System for the CPACT Batch Reactor Process This report provides an outline of the batch reactor system followed by a description of the instrumentation and any conditioning of the monitoring and control signals to achieve the necessary measurement and actuation levels. The data acquisition of these signals and their presentation and use within the system is described in detail. The final programme is fully explained to give the reader an understanding as to how the monitoring, control and data logging aspects of the programme were developed. A user manual is given in an appendix to guide users through the running of an experiment using this control system. This report should also serve as an aid for any future modifications. Esterification of Itaconic
Acid: A Report Industrially, the monoester finds wider applications, although controlling the reaction for optimum monoester yield poses a great challenge. From this perspective, the partial esterification of itaconic acid could serve as a model process for analysis and control under CPACT Project 1. Design of a Fast Sampling Loop
Sampling velocity and isokinetic sampling are presented as crucial factors in the withdrawal of samples, once the features of the sampling probe are chosen. Isokinetic sampling means using a sampling velocity equal to the local system velocity at the point of sampling. Previous work related to the withdrawal of samples from mixing tanks is described and used to justify the choice of optimum features in the design in order to get a highly representative sample from the process. Using this information taken from literature, it can be concluded that the optimum point of sampling is in the stirrer plane. The orientation of the probe should be radial, facing the stirrer blades. In the case of heterogeneous processes, isokinetic sampling should be used. When dealing with chemical reactions and reactive systems, there are also other special factors that need to be considered in the design of a sampling system. These are presented as a modification in the kinetics of the process and an abnormal evolution of the reactive mixture while going through the loop. A chemical model of a fast sampling loop considers the difference between the reactive evolution of the contents of the reactor and that of the sample that goes through the loop. The sampling loop can be modelled as a Plug Flow Reactor connected to the batch reactor. In theory, the kinetic evolution of the content of a batch reactor and the sample in the loop should be identical. However, with real systems, this might not be the case. This is why chemical modelling of the loop is required. The theory behind all these considerations is introduced as a chapter in this report. The design of the pump needed to obtain isokinetic sampling for different stirrer speeds is also studied. Finally, the considerations needed for the insulation of the loop, if this is needed, are considered at the end of the report. The particular values for the specifications of the pump and insulation parameters in the system of interest are described as an appendix, together with some thoughts about the implementation of the NMR analyser in the sampling system. However, it should be borne in mind that this report deals only with the considerations needed for the design of the sampling system and, therefore a careful study of the implementation of a particular analyser was not carried out. The next stage in the design of the sampling system will be the acquisition of that material needed for the construction of the loop considering the facts described in this report, the availability of the components in the market, and the cost of the equipment. Finally, proper construction and testing of the sampling loop will be carried out. Strathclyde Reactor Systems Comparison of In-line NIR, Raman and UV-visible Spectrometries, and At-line NMR Spectrometry for the Monitoring of an Esterification Reaction. Document
Ref: 01/P1/6 Background Experimental
The Raman spectrometer used was a Kaiser HoloPROBE spectrometer employing a 785 nm laser, the NIR spectrometer was a Bomem MB155, the UV-visible spectrometer was a Zeiss MCS and the NMR spectrometer was a development instrument manufactured by Resonance Instruments, with an operating frequency of 29 MHz for 1H. All reactions were carried out in a 1 L reactor at 90 °C, using toluene as a solvent. Optical spectra were converted to 1st derivative format to compensate for baseline offsets. Raman and UV-visible spectrometries employed a univariate calibration model for predicting the concentration of 2-butyl crotonate, whereas NIR and NMR employed partial least squares (PLS) regression. Off-line GC was used as the reference technique. Results Model-based Prediction of Crotonate Composition:
A Brief Report The current study is for a batch reactor and assumes that the reaction product, the ester, is measured on-line. The on-line results are normally available every 15 minutes. The results shows that it is possible to predict the future composition horizon of the crotonic acid esterification reaction from past known ester composition values, using a simple linear time series models of second order. However, the models developed are only able to predict at the particular reaction temperatures for which the model has been developed. A further investigation into a general prediction model which has temperature has a parameter of the model indicated that although there is a possibility of applying temperature compensation to the prediction models, further study and more validation would be required, particularly on the prediction horizon and the prediction model itself. Modelling and Simulation of Crotonic Acid
Esterification Process A MATLAB/Simulink model has been developed to describe the kinetics of the crotonic acid esterification process. The initial model development includes the modelling of the chemical kinetics using both mathematical equations and experimental analyses and modelling of the process and control system. Two unknown kinetic parameters, activation energy and the pre-exponential constant, were calculated experimentally. The parameters for the energy balance and heat losses also had to be calculated experimentally. Other parameters were obtained from literature and from reactor specifications. In order to validate the model, the model prediction for a reaction at 90ºC was compared with real data from a reaction, as shown in the Figure below. The initial results are promising and indicate that the simulation could be used successfully for further control studies.
System Identification, Control Design and
Implementation for the Batch Reactor Process
The model of the heater-chiller unit was found by empirical modelling which proved to be more accurate than input-output identification due to the nonlinearities within the heater-chiller control code. This model was combined with the work previously performed on the reactor to produce a ‘thermal model’ for the process which can be used for the heating and cooling of the reactants.
This was achieved using MATLAB and Simulink routines applied to the models developed. The tuning was initiated through traditional PID tuning methods and then fine-tuned on the actual system. Additional LabVIEW control routines were developed to reduce the control on saturation. This was dealt with in a similar manner to the anti-reset-windup and part of the previous code was used to limit the control action when the thermal difference was approaching the limits. The final section of the report details the implementation of this control system in LabVIEW and produces results from the batch reaction process which show the control system working effectively. Internet
Applications and Implications using the LabVIEW control and monitoring
system Two particular methods were examined: 1. FTP File Transfer to an external server worked very well. Two individual front panels were run on the same computer that was controlling the reactor and monitoring all of the spectrometers at the same time. The images of the front panels were created on that computer but saved through the local area network onto another computer. This second computer was used to send updated image files to the Internet server once every minute. Monitoring the arrangement via the Internet showed that the system worked correctly and that the system updated all the pages correctly even where one of the front panels was hidden from view by another panel on display. 2. Real time display by http where the computer monitoring the spectrometer outputs and controlling the reactor is also configured as an http server. This worked well where the server computer was running only the control, monitoring and server programs. However, the system processor load was very high with 100% processor activity being regularly encountered. Attempting to run other programs at the same time caused a general slowing down of all the programs even with only one external computer taking the output from the computer via the Internet. For this reason it would be necessary to use a multi-processor computer or send the output only to a larger Internet server for distribution. The actual rate of data being received at the Internet viewing computer was in the order of 16 kilobytes per second and hence a fast Internet connection would be necessary to allow this to work. Connecting Different Computer Controlled
Spectrometers to one Network Computer Using LabVIEW The main body of the report outlines the choices and decisions made in implementing the network connections. The Appendices provide detail on specific LabVIEW VIs.
|
