Kathe’s Blog

“Simulation-based Analysis of Masdar City’s Personal Rapid Transit System”

Katharina Müller war erste Absolventin des Bachelorstudiengang Wirtschaftsingenieurwesen mit Schwerpunkt Logistik an der Hochschule Ulm. Ihr Praxissemester verbrachte sie bei einer Speditionsfirma in Dubai und ihre Bachelorarbeit schrieb sie für Siemens Airport Logistics in Singapur. Für ihre Studienleistungen wurde sie mit dem Hochschulpreis der schwäbischen Wirtschaft 2008 ausgezeichnet. Sie entschloss sich dann den Masterstudiengang Systems Engineering and Industrial Management anzuschließen und arbeitete in Teilzeit für das Institut für Betriebsorganisation und Logistik.

Für ihre Masterarbeit zog es sie zurück in die Vereinigten Arabischen Emirate, diesmal in die Hauptstadt Abu Dhabi, welche gerade 15 Milliarden Euro in die Errichtung der ersten CO2-freien Ökostadt der Welt investiert. In Masdar City sollen in wenigen Jahren 50.000 Einwohner und 50.000 Pendler leben und arbeiten. Die gesamte Stadt wird ausschließlich von Strom aus regenerativen Energienquellen versorgt und der Müll soll komplett wiederverwertet werden. Konventionelle Autos wird es in der Stadt nicht geben, sondern ein ganz neuartiges Verkehrssystem, welches so weltweit noch nicht realisiert wurde. Ein Personal Rapid Transit System befördert Passagiere und Fracht voll-automatisiert, nachfrage-gesteuert und ohne umzusteigen zu jeder beliebigen Station des Netzwerkes in kleinen Elektrofahrzeugen. Diese Form des öffentlichen Nahverkehrs bietet den Komfort eines Taxis und kann gleichzeitig durch intelligente Steuerung der Fahrzeuge wesentlich effizienter und energiesparender sein. Bevor eine komplette Modellstadt mit Sitz der UN-Behörde für regenerative Energien (IRENA) auf dieses neue Verkehrssystem angewiesen ist, musste ein Simulationsmodell erstellt werden, um verschiedene Design-Parameter zu testen.

Das neu gegründete Masdar Institute of Science and Technology arbeitet eng mit dem MIT in den USA zusammen und setzt die gleichen wissenschaftlichen Ansprüche an Dozenten und Studenten. Über die Herausforderung der technischen Realisierung, ihre vielfältigen beruflichen und kulturellen Erfahrungen wird Katharina Müller berichten und kann Fragen beantworten zur Vorbereitung von Auslandsaufenthalten, Stipendien und zur derzeitigen wirtschaftlichen Lage in den Vereinigten Arabischen Emiraten nach der Weltfinanzkrise.

Donnerstag, 8. Oktober 09, 18 Uhr, Aula der HSU, Prittwitzstr.10, Ulm

I finished the last chapter about the conclusion and outlook a week before the final submission date. Though I knew there is still plenty of work left to polish it up, I already felt relieved and happy. As my supervisor now moved to Dubai as well, we met in a hotel nearby to discuss his comments. The chapter about the experiments and results needed major restructuring and clearer table formatting to compare the simulation runs. A few statements were left that need further references. Because we cannot celebrate on the day of submission, I took one day off to have a barbecue and Salsa party with my friends. The last night I read through the whole thesis at once to correct all the small grammar and formatting errors, but it took till 7 am in the morning. I postponed a meeting with my supervisor by 2 hours to get a bit of sleep before driving all the way to Abu Dhabi. We added the final touches to my thesis and discussed my contract extension. Then I submitted my thesis by email to my university in Germany. The secretary confirmed that they’ve received it and the printed version I can submit in a few days when I am back in Germany. I guess everyone who wrote a thesis can imagine how relieved and happy I was that day. It was so much hard work and I really felt burned-out sometimes, so that I am really looking forward to my holidays now. Though my supervisor here is not grading my thesis, he told me that I did an excellent job. As he is from MIT and demands the same standards at MIST, I felt proud about my achievements, the efforts were worth it.

With the simulation model verified and calibrated, experiments were designed to test the anticipated system performance and optimize its design. The most important performance parameters and their metrics were picked and seven scenarios chosen to represent a repertoire of testing options and to test a wide range of system parameters and their impact. We tested:

1. Level of service for different PRT fleet sizes to determine the optimal number of PRT vehicles in the network.
2. Vehicle allocation algorithms, which offer the overall best performance energy and service level wise.
3. Different battery charging strategies, such as the recharging in depots only, charging facilities at station berths and additionally with supercapacitors aboard.
4. Change in occupancy rates to measure service level improvements and energy conservation.
5. The impact of 2000 additional passengers in case of a conference or trade fair at the Convention Center as well as increasing the freight demand locally if the hotel is outsourcing its laundry cleaning to a service provider outside of Masdar.
6. General system capacity stress-testing by incrementally increasing overall passenger numbers of the basic demand model.
7. The temporary closure of tracks due to emergency or maintenance and to observe the traffic flow when vehicles need to take alternative routes in such a limited network.

Running the experiments is time consuming, one run takes about 20 minutes and every scenario is repeated several times with different parameters. Furthermore I used a screenshot tool to record a video of the running simulation and various views of the stations and the network. With Adobe Premiere Pro I did the cutting, blending as well as adding of titles and background music (by 5tarbuck). The result is a two minutes video about the simulation model.

Abstract, Acknowledgements, Non-Disclosure Agreement, Contents
1 Introduction
1.1 Masdar City
1.2 Project Objectives
2 Personal Rapid Transit Systems
2.1 What are PRT Systems?
2.2 Technical Configurations
2.3 Advantages and Disadvantages of PRT Systems
2.4 Masdar City PRT Specifications
3 Simulation as a tool for Systems Engineering
3.1 Simulation in General
3.2 Discrete-Event and Agent-Based Simulation
3.3 Methodology
4 Implementation of the PRT Simulation Model
4.1 Simulation Software
4.2 Model Hierarchy and Structure
4.3 Functionality
4.4 Verification, Calibration and Validation of the PRT System Model
5 Scenario-based Performance Analysis
5.1 Design of Experiments
5.2 The Baseline-Model
5.3 Allocation Strategies
5.4 Battery Charging Strategies
5.5 Occupancy Rate
5.6 Convention Center and Hotel Logistics
5.7 Stress-Testing
5.8 Track Closures
6 Analysis of Results
6.1 System Costs
6.2 Conclusions
6.3 Outlook
Attachments: Lists of Abbreviations, Figures and Tables

I finished the programming and included some more functionalities than originally planned. Now the most difficult part is the verification and validation of the simulation model. Verification is a parallel process to programming, whenever a new model is finished or a function added, the model needs to be tested and debugged. The larger and more complex the model becomes, the more difficult the verfication process. There are several ways to find programming errors. First the Anylogic compiler checks the syntax and highlights typing mistakes. Second when running the simulation Java runtime exceptions might occur when variables are not properly initialised or there is something wrong with the SQL queries. Third, one can look at the animation and local variables of the objects. Furthermore it is possible to create output into the console to check wether functions and if-cases are processed correctly.

For this phase of the study we had to alot enough time, as we need to assure that the model is working correctly before analysing the output. The model and all its parameters need to be set correctly and the model behaviour should be monitored to find a sensible combination of parameters to obtain reasonable output.

As I was well aware of the cost of living in the UAE, I applied for a scholarship from the German Academic Exchange Service – DAAD, who offers special scholarships for writing a Master’s thesis abroad.
The application has to be submitted latest three months prior to departure and comprises the application form, CV, transcript of records, Master’s thesis outline, time plan, exploratory work, list of literature, approval of the topic by university, comprehensive assessment by supervisor, invitation letter of university abroad etc. It took me several weeks to get all the paperwork done and I had planned well ahead to submit a complete and carefully worded application on time. As scholars will receive several thousand Euros, it is worth to put great effort in writing the application documents. Furthermore this helps to prepare for the thesis, making it more likely to become a successful research project.
The scholarship includes up to 6 months living allowance, travel expenses (fixed amount) and insurance. The amount they pay depends on the country and is adjusted from time to time, for UAE I will receive 850 EUR/month plus once 435 EUR for travel expenses. As the Masdar Institute of Science and Technology is offering me a room in their students accomodation, I do not need to rent an expensive apartment in Abu Dhabi.
Just two days before I left Germany I received the letter that I was granted the scholarship, which was a big relief. In return I have to write a report on my thesis and experiences during my stay abroad to give advice to other applicants. The DAAD has an Information Centre in Abu Dhabi and upon my arrival I received an email from the director inviting me for a meeting to discuss my project objectives.