WEB BASED REMOTE EXPERIMENTS FOR CHEMICALENGINEERING EDUCATION—THE ONLINE DISTILLATION

COLUMN

A. KLEIN� and G. WOZNY

TU Berlin, Institute of Process and Plant Technology, Berlin, Germany

This paper deals with the development of a real laboratory plant with a web basedprocess control system for training of students over the Internet and demonstrationpurposes. The system is described in detail on the example of a fully automated

distillation column for the separation of ethanol–water mixture. The process–user-interfaceis implemented using the tomcat container with a JSP application and the OPC interfaceAPI. Furthermore a real-time video streaming is used for the online visualization of theplant. Finally, the successful evaluation of the tool is exemplarily presented for atroubleshooting task as a didactical scenario.

Keywords: web based learning; remote experiments; practical training.

INTRODUCTION

In the field of chemical engineering education, practicalexperience plays an important role. Currently, many insti-tutions for engineering education are making great effortsto provide a web-based access to real experimental facili-ties to their students (Carnevali and Buttazzo, 2003; Erbeand Bruns, 2003; Zeilmann et al., 2003; Michaud et al.,2001; Henry and Knight, 2003). Since practical trainingis also a very time consuming task for both, students andteachers, and therefore often neglected, the Department ofProcess Dynamics and Operation (d|b|t|a) of the BerlinUniversity of Technology is searching for new ways ofgiving access to the laboratories to share experimentswith other groups and integrating this field to the existingcurriculum (Wozny et al., 2004).

Such remote laboratories are expected to have a signifi-cant pedagogical benefit, especially compared to virtuallaboratories, which are based on simulations. In particular,remote labs are considered advantageous for being closer toreality, providing real-time, noisy data and a higher degreeof freedom for the learners. This is assumed to increase theinvolvement of the learners, enabling a ‘thrilling’ learningexperience and stimulating higher order thinking skills.

Another advantage of remote labs over on-lab exper-imentation is seen in the greater accessibility. Learnersare enabled to conduct experiments from wherever theywant and they are able to use experimental facilities fromanywhere without moving. Thus, didactical scenarios with

a student who is resident in United States conducting exper-iments in a European lab, supervised by an Australia-basedprofessor are feasible (Urbas et al., 2005). But the missinghands-on experience puts the online experiments at a slightdisadvantage with local experiments. This is seen in themissing achievement of skills to prepare the equipment,run the experiments and so on. Complexity is reduced, andthe students will never have to conduct real troubleshoot-ing, since they are not the ones to adjust the equipment ifsomething goes wrong.

In contrast, simulations show a further reduced complex-ity since the system will only respond according to its pre-defined model. Nevertheless, complexity reduction is abasic didactical principle, and the power of simulation isseen in the interactive illustration of theoretical concepts.

There are different didactical applications of onlineremote labs possible (Michaud et al., 2001):

. teachers can use it during lectures for demonstrations;

. students can use it during scheduled lab sessions as anexperiment sharing tool;

. students can use it outside class as a flexible self-trainingtool.

This paper focuses on the technical realization and theimplementation of an online column within a teachingcourse for students.

THE ONLINE DISTILLATION COLUMN

In this paper, a web based laboratory is presented used toaccess and conduct experiments with chemical processplants (Klein et al., 2004). The process used for this experi-ment consists of an ethanol–water distillation column withTe

chset

CompositionLtd,Salisbury

Doc:

{ICHEME}ECE/Articles/Pagination/ECE06015.3d

Printed:

25/9/06

Page#:

6page(s)

Opp:UNDEFINED

REF^&_name;

Task:UNDEFINED

REF^

&_task;

�Correspondence to: Mr A. Klein, TU Berlin, Institute of Process and PlantTechnology, Sekr. KWT 9, Strasse des 17. Juni 135, 10623 Berlin, Germany.Email: andreas.klein@tu-berlin.de

1

0000–0000/06/$00.00+0.00# 2006 Institution of Chemical Engineers

www.icheme.org/ece Trans IChemE, Part D, Month 2006doi: 10.1205/ece06015 Education for Chemical Engineers, 00(0): 1–5

20 glass-trays and a video streaming application for visual-ization which is set up at the Department of Process Designand Operation (d|b|t|a) at the Berlin University of Technol-ogy. The ethanol–water mixture is a standard learningexample in teaching process behaviour and the process isused as a part of a bio-ethanol-project. The distillationtower is 3.5 m high and has a diameter of 35 mm. It is auto-mated and controlled by the process control system (PCS)ABB Freelance 2000TM. To facilitate interoperability withother applications, Freelance includes an OPC-serverwhich is based on the DCOMTM standard. The technicalaspects of the communication between the plant and theuser will be described below. The plant is equipped with21 sensors that measure the temperatures on every secondtray, flows (products and feed) and pressures (pressuredrop and absolute pressure). Four actuators (feed flowrate, feed temperature control-on/off, heat duty and recyclestream) are mounted in the plant which can be controlledon-site via the PCS Operator Console as well as via OPCover the Internet by a remote user.

Besides the control of the actuators and the observationof the sensor data, which are visualized by a Java-appletin the browser, users can visually observe the columnusing a video stream also. The camera is movable betweenfive fixed positions and shows the interesting parts of thecolumn like the reboiler, feedstage and condenser. Thelive streaming video signal is captured with a broadcasterand sent to the user via a streaming server. The usedmedia format is quicktime, which is broadcasted by theSorenson Broadcaster and served by the open-sourceDarwinTM streaming server by AppleTM. As shown inFigure 1, the camera positions are driven by the PCS andare also controlled via Internet.

TECHNICAL BACKGROUND

For remote experimentation, the process can be controlledin real-time via internet using a Java applet called PE/SSE.1

Real-time data trend diagrams are displayed for the obser-vation of different sensors like temperature, flow, pressureand concentration. Additionally, the streaming web cam pic-ture is displayed. The user is connected to the process con-trol system via Java and OPC Technology (OPC). Figure 2shows the communication structure in detail.

On the client-side, the proprietary developed JavaTM-based experimental PCS called PE/SSE that is able to runthe user interface (GUI) embedded in web browsers is used(Klein et al., 2003). The data-exchange between the GUIand the PCS core is based on an open internet protocol forprocess data transport (PDTP) (Urbas, 1999). To connectFreelance2000TM to the PCS core a bridge is implementedbetween the client side and the server-side process controlsystems. This architecture ensures scalable and reliableaccess to our laboratory resources. As shown in Figure 2,the bridge between client- and server-side PCS consists of:

. Online distillation column with Devices/Remote I/Oand industrial ABB Freelance2000TM PCS.

. OPC-server: This server is part of the ABB Freelance2000TM PCS.

. OPC-to-JAVA-client (JOPC): This client provides aJavaTM api to access the OPC-server.

. PDTP-client: This newly designed compound translatesPDTP-variables used by the PDTP-broker into OPC-to-JAVA-client function calls.

. PDTP-broker: The purpose of this compound is tomanage the communication between the PE/SSE-GUIsof users and other gateways. The OPC-variables whichhave to be sent to the PDTP-broker are defined in aconfiguration-GUI.

Besides the PCS, we have developed an interfacebetween the broker and commercial simulation environ-ments like gPromsTM.2 With these interfaces at last, itdoes not matter if the data originate from a real processor a simulation, if they are PDTP compliant. Furthermore,the broker handles the user connections between PE/SSEand process (real or simulation). Due to the rights manage-ment of the broker, only one user can act as an operatorwhile a lot can view the scenario. Thus, with this systemvirtual classroom scenarios can be run. But the operatormust not be the same person during the plant operation.The operator can give the process free and one of theother participants can take over. This mechanism is handledwith help of an ‘Admin’-Button, which works like a ‘deadman’s handle’. If no one presses the button in a definedtime frame the plant shuts down. All persons who are work-ing with the system only need s an internet connection, butcan be logged in from all over the world.

The system has been implemented fully scalable withmodularized process side broker architecture. The frontend of the online column has been adopted closely fromDIN 19227 and thus is similar to the user interface designof commercial PCS with some additional abilities (Figure 3).

Using its open internal object model, many add-ons fordifferent purposes, e.g., integrated adaptive informationand support depending on process data or backgroundinformation are implemented. For the configurationof PE/SSE, a library of templates is implemented. Sincethe object model behind the GUI bases on the JavaBeans-standard, it is easy to add new or edit existing components.

Due to the fact that users deal with a real remote plant,security issues are key aspects of remote experiments.Thus, remote experiments have to imply an inherent safeprocess design as well as to provide maximum securityagainst hacker attacks. The plant is mounted in a fireprotective fume hood. In the online experiment, all valvesand the reboiler are designed inherent safe so thesystem in case of trouble will fall back to a stable stateon its own. Additional security is given by using a weakreboiler which causes slow system dynamics and expendsreaction time for the operator. To create feasibleshutdown-scenarios and to find possible safety risks, aHAZOP Analysis was conducted. Its results on the onehand have been the basis for the design of the processfallback strategy, and on the other hand for the automaticshutdown scenario of the PCS.

Safety against hacking is provided by the usage ofrestrictive firewall settings and network structures, as wellas several underlying systems of the experiments

1Process Operation Education and Training/Small Systems Edition.

2gPromsTM by PS-Enterprise modelling and simulations software forprocess engineering.

Trans IChemE, Part D, Education for Chemical Engineers, 2006, 00(0): 1–5

2 KLEIN and WOZNY

components. Additionally, Log files are automaticallytraced to reduce hazard of being hacked.

EVALUATION OF THE TOOL

An evaluation of the tool has been conducted (Gausset al., 2004). The students in the role of operators have toidentify and to fix two disturbances affecting the properoperation of the column. Therefore, the students have todetect the irregularities in the related trend diagrams andgather the correct malfunction cause. At least, the studentsshould initiate the correct reaction and bring the processback into a steady state. The troubleshooting scenario sup-ports the students to go deeper into the relation of theoreticalmodelling and the physical behaviour of the real process. Asthe students operate the process, they get a new perspective

on theoretical learning content. The results of the evaluationstudy indicate that the learning scenario was highly acceptedby the students (Gauss et al., 2004; Klein et al., 2004). Trou-bleshooting performance was influenced by intrinsic motiv-ation and information processing velocity of the learners.In this controlled study the acceptance of the scenario andits effects on learning outcome are positively evaluated.

At the d|b|t|a, online experiments are integrated intoexisting courses like process dynamics and processcontrol, where the lecturer can access the experimentfrom a lecture, as well as into practical exercises, wherefor instance students can handle the column themselvesto see the effect of certain disturbances on thehydrodynamic, energy balance and the product quality.The results of the experiments can be taken over intoMS EXCEL and there be analysed. With this experiment,

Figure 1. Web interface and online column.

Trans IChemE, Part D, Education for Chemical Engineers, 2006, 00(0): 1–5

WEB BASED REMOTE EXPERIMENTS FOR CHEMICAL ENGINEERING EDUCATION 3

practical process knowledge can be transferred to thestudents inside the learning environment (e.g., http://lms.fakiii.tu-berlin.de/).

CONCLUSIONS

Active working with online experiments is assumed to

support students acquiring applicable knowledge for practi-

cal situations. The seen disadvantage of online experiments

not to impart hands on experience is balanced by the

greater accessibility of online experiments. Due to the

higher utilization of experiments, even for the owner it isworthwhile to open his local capacities to a broader com-munity. In contrast to online experiments, online simu-lations show a further reduced complexity since thesystem will only respond according to its predefinedmodel. Technically, there are several methods to offerexperiments online. One of the methods using a

Figure 2. Communication structure: OPC-Server—PE/SSE—Bridge.

Figure 3. Comparison PE/SSE-GUI and ABB Freelance PCS.

Trans IChemE, Part D, Education for Chemical Engineers, 2006, 00(0): 1–5

4 KLEIN and WOZNY

commercial underlying PCS with a proprietary Java andOPC-based has been presented here.

A major advantage of the method described in this articleover other methods, is the ability to put context sensitiveinformation on the process of modelling as well as to con-trol and operation theory of the processes elements, or theprocess itself directly on the elements of the PCS-screen.

Because users handle a real remote plant, security issueshave to be considered carefully with remote experiments.The process design should address three levels ofsecurity—play for safety—narrow damage down—avoidenvironmental impact—and thus be considered in a safetyanalysis like the HAZOP. The results of the analysis leadinto a inherent safe process design and some IT infrastruc-tural actions to ensure process safety as well as safetyagainst web attacks.

The implemented experiment was positively evaluated in aformative study and has been introduced into two lectures—process dynamics and process control. Due to the goodmotivational effects, other online experiments have beenimplemented using standardized access technologies.

NOMENCLATURE

API application programming interfaced|b|t|a Department of Process Dynamics and OperationDCOM distributed component object modelGUI graphical user interfaceHAZOP Hazard and Operability StudyI/O input/outputJSP JavaServer pagesOLE object linking and embeddingOPC openness, productivity, collaboration (OLE for process

control)PCS process control systemPDTP process data transfer protocolPE/SSE process operation education and training/small systems

edition

REFERENCES

Carnevali, G. and Buttazzo, G., 2003, A virtual laboratory environment forreal-time experiments. In Proceedings of the 5th IFAC InternationalSymposium on Intelligent Components and Instruments for ControlApplications (SICICA 2003), Aveiro, Portugal, 9–11 July, 39–44.

Q1 Conrad, W., Baumann, E. and Mohr, V., 1980, MTP: Mannheimer Test zurErfassung physikalisch-technischen Problemlosens (Hogrefe,Gottingen).

Erbe, H.-H. and Bruns, F.W., 2003, Didactical aspects of mechatronicseducation. In Proceedings of the 5th IFAC International Symposiumon Intelligent Components and Instruments for Control Applications(SICICA 2003), Aveiro, Portugal, 9–11 July, 45–50.

Q1Friedrich, H.F., Hron, A. and Hesse, F.W., 2001, A framework for design-ing and evaluating virtual seminars, European Journal of Education,36(2): 157–174.

Gauss, B., Klein, A., Urbas, L., Zerry, R. and Wozny, G., 2004, Learningby troubleshooting—a suitable didactical scenario for online exper-iments?, INCOM 2004—11th IFAC Symposium on Information ControlProblems in Manufacturing, 5–7 April 2004, Salvador, Brazil.

Q1Grund, S. and Grote, G., 2002, DERIVE final evaluation results. Retrieved9 December 2003, from http://www.derive.uni-bremen.de/pdf/D55.pdf.

Henry, J. and Knight, C., 2003, Modern Engineering Laboratories at adistance, Int J Engng Ed, 19(3): 403–408.

JOPCClient TM, OPI Odense Production Information, Denmark.Klein, A., Zerry, R., Hausmanns, C., Urbas, L. and Wozny, G., 2003, Java

based process control of a distillation column via internet using OPC;ECCE-4, Granada [Spain], September 2003, 21–25.

Klein, A., Hausmanns, C., Urbas, L. and Wozny, G., 2004, Internet-basedlearning—the online laboratory distillation column, EDEN 2004—Annual Conference, 2–6 June 2004, Budapest.

Michaud, F., Gentil, S. and Barrault, M., 2001, Expected benefits of web-based learning for engineering education: examples in control engineer-ing, European Journal of Engineering Education, 26(2): Q2151–168.

Q1Oswald, W.D. and Roth, E., 1987, Der Zahlen-Verbindungs-Test (ZVT)(Hogrefe, Gottingen).www.opcfoundation.org

Q1Unz, D., 2000, Lernen mit Hypertext: Informationssuche und Navigation(Waxmann, Munster).

Urbas, L., 1999, Entwicklung und Realisierung einer Trainings- und Aus-bildungsumgebung zur Schulung der Prozessdynamik und des Anlagen-betriebs im Internet (Development and realisation of a trainingsenvironment for process dynamics and operation via internet),VDI-Fortschrittsberichte Reihe 10 (614).

Urbas, L., Gauss, B., Klein, A. and Wozny, G., 2005, Online remotelaboratories. In: Wozny, G. (ed.) ELearning in Process andChemical Engineering—A Practical Overview. (VDI Verlag, Berlin,Dusseldorf).

Wozny, G., Klein, A., Zerry, R., Hausmanns, C. and Urbas, L., 2004,Improvement of learning of process technology using modern infor-mation technology, PRES 2004, 7th Conference on Process Integration,Modelling and Optimisation for Energy Saving and Pollution Reduction,22–26 August 2004.

Zeilmann, R.P., Gomes da Silva Jr, J.M., Bazanella, A.S. and Pereira, C.E.,2003, A web-based remote laboratory for control education. In Proceed-ings of the 5th IFAC International Symposium on IntelligentComponents and Instruments for Control Applications (SICICA2003), Aveiro, Portugal, July 9–11, 51–56.

The manuscript was received 7 June 2004 and accepted for publicationafter revision 28 August 2006.

Trans IChemE, Part D, Education for Chemical Engineers, 2006, 00(0): 1–5

WEB BASED REMOTE EXPERIMENTS FOR CHEMICAL ENGINEERING EDUCATION 5

ECE06015Queries

A. KLEIN and G. WOZNY

Q1 Where are the following cited in text?Conrad et al. (1980)Friedrich et al. (1980)Grund and Grote (2002)Oswald and Roth (1987)Unz (2000)

Q2 Please incorporate web address into text.