Dr Dongda Zhang Cantab, DIC, AMIChemE, AMRSC

Lecturer in Process Systems Engineering and Machine Learning

Full contact details
View graph of relations

I am always looking for high quality PhD students to undertake research in broad areas of bioprocess systems engineering, chemical reaction engineering, machine learning, systems biology, and process analytical technology. The main ongoing activities in my group are divided into three categories.

Topic 1: Development of online digitalisation technology for industrial bioprocess monitoring, prediction, optimisation, and decision-making.

Online digitalisation technology

 Online Bioprocess Digitalisation

This topic aims to explore and integrate a range of mechanistic and machine learning based modelling strategies to develop a multilevel digital framework (digital twin) for industrial bioprocess dynamic simulation and prediction, online optimisation (online learning), and product quality control. The framework will effectively rectify and covert low quality industrial plant data (e.g. high noise, missing information) into a series of accurate predictive system models for process real-time digitalisation and state estimation. Cutting-edge online optimisation technologies will be also embedded into this framework to further support decision-making during an ongoing process.

 

Topic 2: Multiscale visualisation technology for biological knowledge discovery and physical laws generation.

Multiscale visualisation technology

Multiscale Visualisation Technology

This topic aims to take advantage of frontier modelling techniques in systems biology and bioinformatics, bioprocess reaction engineering and systems engineering, and computational fluid dynamics to construct a multiscale dynamic modelling framework which can visualise activities of metabolic reaction network under industrial fermentation and photo-bioreaction systems. The framework will directly reveal primary metabolic constraints and scale-up challenges limiting the synthesis of targeted bioproduct at large scale upstream operation systems, thus providing a clear direction for industrially-desired strains development at an early research/commercial stage. Moreover, the large amount of data generated from this multiscale visualisation framework will be exploited systematically to construct new constitutive relations (empirical physical laws) for the design of high efficiency bioreaction systems and unit operations. 

 

Topic 3: Data-driven based ‘whole process’ approach for (bio)chemical manufacturing system design and integration.

Integrated process design

Manufacturing Process Design and Integration 

Designing self-sustained chemical and biochemical manufacturing systems is one of the grand challenges prioritised by the current process industry. To facilitate development of the next generation (bio)chemical processes, it is essential to refine existing plants and design new manufacturing systems via a ‘whole process’ approach: identifying process interactions to simulate and optimise not only a single unit operation but the whole process sequence. This requires the use of both conceptual process structure identification methodologies and rigorous process simulation techniques such as superstructure construction, first-principle mathematical modelling, life cycle assessment, transfer learning, and surrogate (machine learning based) modelling and optimisation to quantify complex relations between feedstock pre-processing, upstream reaction, and downstream separation, and to determine optimal reaction-separation pathways and operation strategies for process integration and intensification.

 

Topic 4: Soft-sening, online monitoring, and predictive maintenance technology for industrial batch/continuous processes 

Process monitoring

Process soft-sensing and predictive monitoring

Smart manufacturing is one of the novel concepts arising from the era of the 4th Industry Revolution. At present, the downtime of equipment and processes is costing UK manufacturers more than £180 billion per year. Given development of disruptive artificial intelligence technology and large amount of data accumulated in the process industry, applying machine learning to construct data-driven predictive models offers a paradigm change in the way that maintenance is planned and executed in manufacturing. An approach to predictive maintenance will be developed that combines different monitoring techniques using machine learning to target when maintenance will be required in the future. The predictive technique to be developed will also allow analysis of how manufacturing processes can be operated in order to extend operating campaigns in life extension. This will allow manufacturing production to be planned in such a way that operation of the production (capacity and operating conditions) can be combined with knowledge of the maintenance implications of the operation to maximise asset productivity. 

 

More detailed descriptions of these projects are available upon request. Students are also welcome to propose their own project and I am open for further dicussion.

 

PhD opportunities