Professor Xue Z. Wang
- Position: Professor
- Areas of expertise: pharmaceutical engineering and process control; nanoprocessing; process modelling; process optimisation; data mining; artificial intelligence; nanotoxicology; crystallisation; nanomilling; granulation
- Email: X.Z.Wang@leeds.ac.uk
- Phone: +44(0)113 343 2427
- Location: 1.43a Engineering Building
- Website: Googlescholar
Profile
Intelligent Measurement, Control and Analytics of Particulate Processes
Sensor and PAT Data Mining (decision tree generation from data, neural networks, wavelets, principal and independent component analysis, 6s statistical quality control, multidimensional visualisation); On-line Process Analytics (Imaging, NIR, z-potential, acoustics, laser diffraction) for On-line Crystal Shape and Size Measurement and Control, and Characterisation of High Concentration Nano Particulate Systems; Multi-scale Modelling (integration of population balance with morphology modelling) for Shape Control; High-Throughput Formulation of Functional Organic and Inorganic Nano-materials; Quantitative Structure – Activity Relationships (QSAR) for Prediction of Eco-toxicity of Organic Chemicals and Mixtures.
Sensor and PAT Data Mining
The work involves development of:
(i) techniques for effective extraction of key features from sensor signals to enhance measurement capabilities.
(ii) a genetic programming technique for automatic generation of causal models e.g. decision trees that capture the complex relationships between process operating conditions and product quality.
(iv) data mining techniques for historical data analysis for continuous process performance improvement.
(v) data mining and experimental design to support high-throughput formulation of functional organic and inorganic nano-materials.
(vi) an integrated process data mining system integrating various tools including sensor data pre-processing, neural networks, causal networks, statistical and visualization tools.
(EPSRC EP/D038391, GR/L61774)
Measurement and Control of the Shapes of Growing Crystals
On-line video imaging and multi-scale image analysis are investigated to obtain real-time information of crystal growth rates, growth kinetics and population balance which are all characterised based upon individual crystal facets instead of as a spherical shape. The information is used for developing techniques for controlling crystal growth of individual facets. (Shape project, EPSRC/EP/C009541; Vision project, Malvern Instruments).
Multidimensional Population Balance and Crystal Morphology Modelling
Multidimensional population balance models based upon individual crystal facets are developed through the integration with crystal morphology prediction, which are combined with shape measurement to achieve model predictive control for morphology.
(Shape project, EPSRC/EP/C009541)
In-line sensing and quality control of high concentration nanoparticles
Size measurement in high solid concentration for nano-particles poses great challenges because dilution will cause the change of sample conditions. The work aims at developing an integrated solution by combining z-potential, acoustic attenuation spectroscopy, impedance process tomography and data mining.
(ZAPT project, DTI TP2/SC/6/I/10097)
Simultaneous Measurement of Solid and Liquid Phase Properties Using NIR
Our investigation on the information content of NIR spectra has shown that the spectra contain information of both solid and liquid phases including solid and liquid concentration, size, and polymorphs. Signal separation techniques are being investigated for simultaneous prediction of the multiple properties. (NIRIICA PilotPlant project, EPSRC EP/C001788).
QSAR Mixture Eco-toxicity Prediction
Quantitative structure – activity relationships (QSAR) for toxicity prediction of chemicals have been used only for pure compounds. This work aims at developing a new QSAR technique based on molecular descriptors, fuzzy membership function and inductive data mining for prediction of mixture toxicity.
(MixQSAR project, EPSRC Crystal Faraday Green Technology 01306000)
Research interests
Intelligent Measurement, Control and Analytics of Particulate Processes
Sensor and PAT Data Mining
The research involves the development of:
(i) techniques for effective extraction of key features from sensor signals to enhance measurement capabilities.
(ii) a genetic programming technique for automatic generation of causal models e.g. decision trees that capture the complex relationships between process operating conditions and product quality.
(iii) data mining techniques for historical data analysis for continuous process performance improvement.
(iv) data mining and experimental design to support highthroughput formulation of functional organic and inorganic nanomaterials.(v) an integrated process data mining system integrating various tools including sensor data pre-processing, neural networks, causal networks, statistical and visualization tools. (EPSRC EP/D0 8 91, GR/L6177; EPSRC High Throughput Project EP/H008853/1;).
Measurement and Control of the Shapes of Growing Crystals
On-line video imaging and multi-scale image analysis are investigated to obtain real-time information of crystal growth rates, growth kinetics and population balance which are all characterised based upon individual crystal facets instead of as a spherical shape. The information is used for developing techniques for controlling crystal growth of individual facets. (EPSRC SHAPE Follow-on project EP/H008012/1; EPSRC StereoVision Project EP/E045707/1; EPSRC SHAPE Project EP/C009541/1); TSB HMV CGM project; Vision project, Malvern Instruments).
Morphological Population Balance Modelling of Crystal Growth
Multidimensional population balance models based upon individual crystal facets are developed through the integration with crystal morphology prediction, which are combined with shape measurement to achieve model predictive control for morphology. (EPSRC StereoVision Project EP/E045707/1; EPSRC SHAPE Project EP/C009541/1).
Engineering Nanomaterials
Size measurement in high solid concentration for nanoparticles poses great challenges because dilution will cause the change of sample conditions. The research aims to develop an integrated solution by combining z-potential, acoustic attenuation spectroscopy, impedance process tomography and data mining. Research is also being conducted to combine modeling (population balance and computational fluid dynamics) with on-line sening to develop scale-up and control strategies for a hydrothermal process for nanomaterial synthesis (EPSRC EngineeringNano project /EP/E040624/1; EPSRC High Throughput Project EP/H008853/1; ZAPT project, DTI TP2/SC/6/I/10097).
Simultaneous Measurement of Solid and Liquid Phase Properties Using NIR
Our investigation on the information content of NIR spectra has shown that the spectra contain information of both solid and liquid phases including solid and liquid concentration, size, and polymorphs. Signal separation techniques are being investigated for simultaneous prediction of the multiple properties. (NIRIICA PilotPlant project, EPSRC EP/C001788).
QSAR Mixture Eco-toxicity Prediction
Quantitative structure – activity relationships (QSAR) for toxicity prediction of chemicals have been used only for pure compounds. This work aims at developing a new QSAR technique based on molecular descriptors, fuzzy membership function and inductive data mining for prediction of mixture toxicity. (MixQSAR project, EPSRC Crystal Faraday Green Technology 01 06000; NERC NanoSar Project).
Professional memberships
- CENG
- CSci
- MIChemE