If you are from industry and interested in any of these projects please contact us by emailing aapscdt@bath.ac.uk

If you are student and are interested in any of these projects then please complete an expression of interest and state which project you are interested in.

The current available projects are listed below, followed by full project descriptions

  • Virtual Vehicle Validation, with Quick Release
  • Enabling Effective Transdisciplinary
  • Towards new statistical modelling techniques combining expert knowledge and experimental data for propulsion - systems, with AVL
  • Testbed. CONNECT: Intelligent approaches to improve the system reliability of advanced testing methods, with AVL
  • Testbed.CONNECT: Automated Configuration of Simulation Parameters, with AVL
  • Air handling system optimisation for Fuel cell applications, with Cummins Turbo Technologies

Project Descriptions

Virtual vehicle validation

  • Supervisors: Chris Brace
  • Industrial Partners: Quick Release
  • AAPS Research Theme: Digital Systems, Optimisation and Integration
  • Start: October 2021

The automotive industry, along with other high-tech sectors, needs to move away from their reliance on physical prototypes. Today these prototypes are used to validate the performance of the vehicle against the design requirements. This is an expensive and time-consuming process but there are still many areas of real-world operation that are not possible to validate using traditional methods.

The vision is to move to a future state where there will be an increased reliance on virtual validation. The need is for better, faster and cheaper techniques that harness the power of advanced simulation and data analysis. The stakes are high – leaving things as they are is not an option, the industry needs a better way of working. But if the new tools proved to be inadequate the risks are huge, product recalls resulting from errors in vehicle validation could bankrupt a car maker.

There are 3 main facets to this type of validation, of which the first 2 are already areas of investigation, but the third is an essential enabler and often overlooked:

  1. The ability to use technology to virtually test the components, systems and vehicles in a manner that replicates physical testing
  2. The correlation of the virtual testing to the physical experiments to the level of fidelity that will be required for engineering sign off
  3. The novel digital processes that underpin these two objectives. These processes need to describe the way we build, validate and use the vast datasets that all the other steps rely on; the way we design and build the digital models we use; and the complex tests that we perform using these models.

This PhD opportunity will explore these topics in partnership with industry and will result in new techniques with significant potential to change the way that the industry works.

Enabling Effective Transdisciplinary working with the automotive sector

  • Supervisors: Linda Newnes
  • AAPS Research Theme: Transport Policy and Economics
  • Start: October 2021

Over the past few decades transdisciplinary (TD) has been the subject of increased discourse in the context of large, complex, ill-defined, ‘wicked’ problems. However, there has been less consideration of the potential it offers within the practice of engineering. This research looks to create tools which enable effective TD working within the automotive sector. The Mobility Engineering 2030 FISTA White Paper identifies that changes within the sector mean that interdisciplinary working, involving groups formed from people working in similar disciplines, will not be sufficient. It recognises that in the future there will be a need for transdisciplinary working, which goes beyond the academic disciplines to understand the societal context. For example, legislation, standards, culture. However, achieving effective TD working within organisations is not simple. It requires the creation of tools (e.g. processes and methods) which enable clear communication and knowledge transfer within and beyond an organisation. This PhD will leverage input from the TREND (TRansdisciplinary ENgineering Designers) £1.8m platform grant (Dec 2017 – Dec 2022). The over-arching aim of TREND is to provide tools to assist engineers to work in a transdisciplinary manner and to identify the types of engineers that are transdisciplinary. Identifying what makes engineering teams in the automotive sector transdisciplinary and how to assess their current readiness level to be transdisciplinary is the focus of this PhD activity. The PhD will have a particular focus on ‘common’ characteristics and automotive design team behaviour within and across industry case studies. Mapping findings at various life cycle stages such as designer requirements, use of digital tools etc. for each case study/domain against the manufacturing life-cycle phases. This would be followed by cross case-study analysis. The analysis may use techniques such as input/output system modelling to map the designer requirements at each stage of the manufacturing life cycle, and/or socio-technical analysis could be used to classify and model the designer behaviour. In summary the PhD researcher will be required to create a structured framework to estimate the automotive sectors transdisciplinary readiness level. Specific objectives may include 1. Undertake a literature review to understand the state of TD working within the automotive sector. 2. Engage with stakeholders to gather information which informs the design of the TD readiness tool. 3. Create a TD readiness tool. 4. Validate the proof of concept tool within industry.

Towards new statistical modelling techniques combining expert knowledge and experimental data for propulsion systems

  • Supervisors: Nic Zhang
  • Industrial Partners: AVL
  • AAPS Research Theme: Digital Systems, Optimisation and Integration
  • Start: October 2021

Modern automotive powertrain labs create large amounts of data. The data include various key performance metrics, crank angle resolution cycle events and high frequency recordings of all channels in time traces. Historically, the experimental results in the form of lookup tables and scatter plots have not fully exploited the potential of the data and engineers are increasingly focusing on creating statistical models using the available dataset. High quality statistical models can replace some experimental work as the digital twin of physical systems for predictive analysis and can be embedded directly into automotive controllers for model-based control.

With the wide range of modelling tools available, automotive engineers would benefit from a framework of statistical modelling for specific powertrain systems in the form of an automated tool. This PhD will seek to create such a tool with a help of a large commercial database of experimental data. Familiarity with the physical models for individual components (batteries, motors, engines, fuel cells…) should be the starting point of the study. Open source machine learning libraries, such Keras, will then be used to explore the available dataset to investigate the predictive performance of statistical models, such as Neural Networks, compared to the physical models.

The technical know-how generated in this study is expected to provide the tool users with specific guidance such as:

  • whether important inputs are missing for specific technologies; 
  • how to reduce the number of input variables of the problem for faster model training.
  • how to run iterations of training to eliminate irrelevant areas of the problem space and instead focus on areas of special interest.
  • which statistical models are most suitable for the specific engineering problem.
  • whether physically inspired rules should be included in the ML structure to improve the model performance.

A likely deficiency of this approach in highly non-linear systems is that the density for experimental data needed to allow the training of a Machine Learning structure would be impractical. If this proves to be the case, an alternative approach should be considered that seeks to embody the engineer’s understanding of the physical causality that underlies the unit under test. This can be represented in the form of physically inspired ‘rules’ or ‘toy models’ that can then be calibrated to represent the unit using an iterative training approach. Such a model could allow a more sparse dataset to be used without sacrificing predictive power.

The outputs from this PhD would be expected to be integrated into a model factory engineering software tool that supports engineers in the creation of mathematical models.

Testbed. CONNECT: Intelligent approaches to improve the system reliability of advanced testing methods

  • Supervisors: Chris Brace
  • Industrial Partners: AVL
  • AAPS Research Theme: Digital Systems, Optimisation and Integration
  • Start: October 2021

Automotive propulsion system development processes have advanced greatly over the last years. State of the art techniques use a blend of physical experiments and high fidelity simulation to provide a highly realistic environment in which to study system behavior.

Practically, this means that the simulated parts of the system need to behave just as if they were physically present in the test room. The test system needs to allow the Unit Under Test (UUT) to interact with the simulated components just as they would in a vehicle. Both requirements lead to significant complexity when compared with traditional scenarios. Take the example of a hybrid vehicle transmission undergoing physical test. The simulation of the remaining parts of the system – engine, vehicle, battery etc. need to cater for conditions such as start, warm up, shut down and error states in a way that is rarely called for in a pure simulation environment.

The preparation of models and test rooms for these blended test scenarios is therefore complex and time consuming. Considerable expertise is needed to diagnose and rectify the many potential error states that prevent the test proceeding as intended. The two main areas that lead to errors are errors in the software implantation (bugs) and errors in the simulation behavior that causes the system to stray into unrealistic operating states. For example – the engine simulation could simply crash (die to a bug) or it could execute correctly but give dangerously large output predictions which then cause damage to the UUT. There are two PhD projects offered in this space, first to improve the reliability of the test methods and secondly to assist the test engineer to identify an optimum setup for high quality testing.

Intelligent approaches to improve the system reliability of advanced testing methods

The first project seeks to consider all appropriate techniques that can speed up and improve the setup and verification of such complex test scenarios. It is likely that some measure of expert knowledge or ‘big data’ approaches would be useful, along with some procedures to ensure that all possible test conditions are anticipated and verified before the test program is scheduled on the real test room.

Some likely research objectives are to:

  • Produce all possible errors on a running system
  • Identify and enumerate (cluster) all possible error states
  • Classify error (Error by software or simulation).
  • Identify a recommendation to solve the error.

The successful project will develop techniques that offer new scientific approaches in areas such as:

  • Test Mutation methods to generate scenarios of interest
  • Recommender Systems
  • Property based testing
  • Model based testing
  • Increasing Quality of testing
  • System Transparency for complex hybrid test systems

Testbed. CONNECT: Automated Configuration of Simulation Parameters,

  • Supervisors: Chris Brace
  • Industrial Partners: AVL
  • AAPS Research Theme: Digital Systems, Optimisation and Integration
  • Start: October 2021

Automotive propulsion system development processes have advanced greatly over the last years. State of the art techniques use a blend of physical experiments and high fidelity simulation to provide a highly realistic environment in which to study system behaviour.

Practically, this means that the simulated parts of the system need to behave just as if they were physically present in the test room. The test system needs to allow the Unit Under Test (UUT) to interact with the simulated components just as they would in a vehicle. Both requirements lead to significant complexity when compared with traditional scenarios. Take the example of a hybrid vehicle transmission undergoing physical test. The simulation of the remaining parts of the system – engine, vehicle, battery etc. need to cater for conditions such as start, warm up, shut down and error states in a way that is rarely called for in a pure simulation environment.

The preparation of models and test rooms for these blended test scenarios is therefore complex and time consuming. Considerable expertise is needed to diagnose and rectify the many potential error states that prevent the test proceeding as intended. The two main areas that lead to errors are errors in the software implantation (bugs) and errors in the simulation behaviour that causes the system to stray into unrealistic operating states. For example – the engine simulation could simply crash (die to a bug) or it could execute correctly but give dangerously large output predictions which then cause damage to the UUT. There are two PhD projects offered in this space, first to improve the reliability of the test methods and secondly to assist the test engineer to identify an optimum setup for high quality testing.

Automated Configuration of Simulation Parameters

The hybrid test field is a complex combination of physical hardware and software simulations of other aspects of the system. Both sides of the system are underpinned by complex computation and test automation platforms. All of these elements need to work together in harmony and make the UUT believe it is operating in a real vehicle in the real world. To achieve this, the simulated aspects of the system must be carefully calibrated to operate in the desired manner. This is a significant task and this PhD aims to develop techniques to speed up and simplify this process for the test engineer. Such a system will need to identify:

  • The best simulation parameters (in terms of Robustness and fewest simulation errors)
  • Identifying the minimum set of simulation parameters that meet this requirement to reduce complexity
  • Generate user friendly information to allow users to understand the system
  • Automatic and guided parameterization with minimum number of experimental runs (Preferably 0, all parameterisation is performed and verified ahead of the physical test programme)
  • Strategies to allow self healing in the case that an error state is encountered.

The successful project will develop techniques that offer new scientific approaches in areas such as:

  • Recommender System
  • Co-Simulation
  • Optimization Techniques
  • Usability in Simulation (e.g. Robust System Stability)

Air handling system optimisation for Fuel cell applications

  • Supervisors: Richard Burke
  • Industrial Partners: Cummins Turbo Technologies
  • AAPS Research Theme: Chemical Energy Converters
  • Start: October 2021

Delivering zero emission heavy duty automotive applications is expected to involve the extensive use of fuel cells. Like traditional combustion engines, fuel cells require an air handling system to provide a pressurised flow of oxygen to react with the hydrogen fuel. However, the challenges presented by a fuel cell differ to those presented by traditional Diesel engines. On the one hand, the problem is simplified because there are no longer pulsating flows to deal with. However, there are new challenges caused by the need for oil free air and demanding temperature constraints. The fuel cell has a relatively high inlet pressure requirement combined with a relatively low inlet temperature requirement, which puts a high emphasis on the isentropic efficiency of the compressor. The exhaust from the fuel cell has much lower temperature than a Diesel engine, meaning there is not enough energy to drive the compressor alone, meaning there is a need to electrify the air handling system. These challenges present a number of new design variables which are yet to be fully understood by turbomachinery manufacturers.

In this PhD, the aim will be to establish a deep understanding of the air handling requirements for fuel cells in heavy duty applications. You will start by undertaking a theoretical exercise into the fuel cell itself to fully understand its operating principles and air flow requirements and constraints. You will then look to match these constraints to air handling solutions, considering the breadth of possible configurations, including proposing your own novel configurations where appropriate.

In undertaking this PhD, you will join a team of researchers within the Institute for Advanced Automotive Propulsion Systems (IAAPS). As part of this team you will be working as part of a £20m Advanced Propulsion Centre/Innovate UK funded programme led by Cummins Turbo Technologies. You will also have direct links to a Fuel cell manufacturer.

This project would suit a student with a background in Engineering with a strong interest in mechanical, thermodynamic, electrical, and chemical systems.