INDUSTRY 4.0 PROCESS OPTIMIZATION THROUGH MACHINE LEARNING AND MULTIVARIATE STATISTICAL TECHNIQUES (INDOPT4.0)

About Technical University of Valencia (UPV) The Technical University of Valencia, Spain (UPV) features in the most important World University Rankings, has Spain’s highest revenue from both public research and knowledge transfer, and is a national leader in patent license income and start-up creation.

About Multivariate Statistical Research group (GIEM)
GIEM is devoted to research, development and innovation activities in the area of multivariate statistical techniques for quality and productivity improvement and mega-database analysis. GIEM treasures a depth knowledge in latent variables based-multivariate statistical methods for data analytics related to digital image analysis, missing data imputation, outlier detection, data fusion, process understanding, monitoring, fault detection and diagnosis, predictive maintenance, process improvement and optimization of industrial continuous and batch processes. Prof. Alberto Ferrer is the group leader.
(https://scholar.google.es/citations?user=T2X1PQQAAAAJ&hl=es&oi=sra)

About the position
GIEM-UPV seeks a dedicated and enthusiastic candidate for a position in integration of machine learning and latent variable based multivariate statistical techniques for process improvement and optimization in Industry 4.0. The position is funded by the Valencian regional government research grant:AICO/2021/111, INDOPT4.0 where the main aim is to develop data analytics methodologies that allow causal models to be obtained from historical production data, which can be used to optimize processes in industry 4.0.

Methodological objectives

1. Develop techniques that simultaneously allow the detection of outliers and the imputation of missing values.

2. Develop methodologies that allow the theory of design of experiments to be applied retrospectively to historical databases.

3. Develop algorithms for selecting information blocks and variables within blocks in PLS models.

4. Develop new methods of process optimization from causal models based on latent variables, using historical data.

5. Develop methodologies that allow the concepts of sequential design to be generalized in the subspaces of latent variables.

6. Develop methodologies that allow the joint definition (multivariate) of the specifications that must be imposed on the properties of the raw materials so that they guarantee with a certain probability (fixed in advance) that the manufactured product will have the desired quality.

7. Adapt the optimization techniques developed to the particularities of batch processes.

8. Develop friendly process optimization software with historical data.

The candidate will develop general methodology, but with strong emphasis on industrial relevance. The work will be based on real production data from several companies supporting the research project.

The position is located in Valencia (Spain).

Required qualifications:
– Master’s degree in engineering, chemistry, data science, applied statistics or mathematics.
– PhD, although preferable, is not required.
– Good knowledge in multivariate data analysis, statistics and machine learning.
– Good knowledge in scientific programming (R, Python or Matlab).
– Good written and oral communication skills in English.
– Ability to work well both independently and as part of an interdisciplinary team.

Additional information
The start date is as soon as possible, and the duration is till December 2023.

GIEM-UPV offers
GIEM-UPV offers a strong academic community and an excellent working environment. The salary is based on standard rates from the Spanish research grants.

Questions
For further details on the position, please contact:
Prof. Alberto Ferrer, e-mail: aferrer@eio.upv.es

CA EN ES