Management of Change (MoC) in petrochemical operations requires engineers to trace cross-document dependencies manually across many interdependent document types per process unit. These dependencies are latent: shared equipment tags, instrument identifiers, and operating parameters are distributed across Word reports and Excel spreadsheets with no machine-readable links. To our knowledge, no existing system infers cross-document dependencies at the section level by reconstructing entity relationships from document text across heterogeneous engineering formats under cold-start conditions; commercial systems organise documents by metadata (and in some cases link documents at the identifier level) but do not perform content-level dependency inference, and NLP-based traceability research targets software requirements rather than multi-format engineering documentation.

This thesis constructs a knowledge graph from 9 document types under cold-start conditions, with no pre-existing structured metadata or asset registries (though the pipeline relies on domain conventions such as ISA-5.1 tag formats). Custom parsers and deterministic entity extractors preserve document-structural provenance; entity resolution links variant mentions to canonical representations; dependency classification and graph traversal identify impacted documents and sections. A scope expansion step using cause-and-effect matrix (CEM) relationships discovers instruments attached to scope equipment, enabling the system to reach instrument-related sections that the scope narrative does not explicitly name. A two-tier addition gap strategy handles entities introduced by the change that have no prior graph representation: same-prefix peer lookup within the process unit, followed by entity-class profiles computed across units. A topology-based review ordering derived from entity and parameter density produces consistent document priority across all six units (mean pairwise Kendall’s τ = 0.795) and matches the review workflow confirmed by Shell process engineers.

Evaluation covers 12 MoC cases from Shell spanning 6 process units, with 467 ground-truth change locations derived from before-and-after document diffs (capturing what engineers actually changed, which may undercount true dependencies). The system achieves document-level recall (macro-averaged) of 0.904, precision of 0.706, and F1 of 0.776. In 9 of 12 cases, all modified documents were correctly identified; of the three cases with a missed document, two involve enclosure-calculation documents connected through indirect chemistry-mediated dependencies, and one is a pure-addition case that missed a P&ID drawing. Addition-gap mitigation raises document recall from 0.756 to 0.904. Region-level localisation reaches recall of 0.625 (F1 0.507). Baseline comparisons against five non-graph approaches confirm that the KG pipeline outperforms the best non-graph baseline (dense embedding, F1 0.612) by 16.4 F1 percentage points; no baseline simultaneously matches the KG’s recall and precision. Results demonstrate robustness across 6 process units of a single operator, not cross-client generalisability. A structured evaluation session with Shell process engineers indicated that the impact scope predictions align with engineering expectations and that cross-document dependency visibility is the system’s strongest practical contribution. ... Read more

In the past few years, rapid strides have been made in the modelling of complex systems due to the advent of machine learning (ML) technologies. In particular, the transformer neural network (TNN) has gained a lot of attention due to its powerful “sequence-to-sequence” modelling for several tasks in science and engineering. Predictive ML models are trained to learn the relationship between a set of input and output data so that new input data can be mapped to their expected output values. This gives us the ability to understand the key factors impacting a system and to make predictions about its future behaviours. This research investigates how to use TNNs for decision making by optimising the output of a TNN. To this end, a non-convex mixed-integer non-liner programming (MINLP) formulation for a trained TNN is proposed. The proposed formulation facilitates solving problems with TNNs embedded
to global optimality. The effectiveness of the formulation is tested on three case studies consisting of an optimal trajectory problem, a verification problem, and a reactor optimization case study. Results show that optimisation over small TNNs can be achieved in under 3 minutes. However the tractability of the formulation quickly vanishes for larger models, highlighting the need for further research to refine the proposed formulation. ... Read more

The modern chemical industry faces many challenges, such as energy transition. However, energy transition alone will not provide enough improvements to the industry to maintain profitability and increase sustainability. To achieve these goals, chemical processes have to be appropriately optimised, and only the synergy of these two factors can improve existing processes. Epichlorohydrin production is an important industrial process, but it suffers from several drawbacks such as high energy consumption, significant wastewater production, and low atom efficiency. This is caused by chlorohydrin-based technology, which requires operations in very diluted solutions. In this thesis, a novel chlorohydrin-free technology for ECH production was investigated. This approach could allow operation in more concentrated solutions, but this route is in the early development stage. One of the most crucial design parameters for this process is proper solvent selection. On the one hand methanol appears to be the most suitable compound for this purpose. On the other hand, some papers reported the separation system for this case to be infeasible due to several azeotropes present in the post-rection mixture. However, with proper constraints and understanding of components' azeotropic behaviour, a separation system, which enables obtaining high purity ECH, was created and applied to the production process. To perform a comparison between HP route and chlorohydrin process an Aspen Plus simulation of both processes were created. For the hydrogen peroxide route with methanol as solvent a novel separation system which enables high purities of ECH were created. Furthermore, possibilities to optimise distillation in the given process were investigated because this unit operation requires significant expenses in terms of CAPEX and OPEX. A review of advanced distillation techniques concludes that Dividing Wall Column distillation is the most suitable technique for this purpose. This technology was then applied to replace two columns, which purifies the intermediate Allyl Chloride (ACH) from the process. Aspen Plus simulations of both processes with and without applied DWC distillation were created to evaluate the influence of these improvements. Moreover, to establish the impact of DWC distillation, an Aspen Plus model of this apparatus was created. Simulation results indicate that this novel epoxidation reaction produces 98% less wastewater than the traditional process. Additionally, the novel approach offers a 10% higher yield and a smaller amount of by-products than the chlorohydrin process. Energy consumption per unit of ECH is also lower for the novel route. Application of DWC distillation led to 3.5% decrease in OPEX, while the CAPEX was smaller by almost 5%. These results indicate that applying a novel epoxidation route and DWC may benefit a given plant. However, more research needs to be performed to implement a novel process in the industry. ... Read more