Process integration and electrification with digital twins

Abstract

The decarbonisation of industrial process heat is one of the most pressing challenges in the global energy transition. In New Zealand, fossil fuels remain the dominant source of process heat, despite having over 80% renewable electricity generation. Milk powder production is a major consumer of process heat, with evaporation and drying processes relying on large amounts of coal- and gas-fired steam. Electrification technologies such as industrial heat pumps and mechanical vapour recompression (MVR) have the potential to significantly reduce emissions, yet widespread adoption has been limited because of the complex interactions between heat and power, in addition to uncertainties around practicality. Conventional process integration (PI) techniques were designed for fossil-fuelled utilities and are poorly aligned with the work requirements and integration constraints of electrification. Meanwhile, legacy simulation tools are ill-suited to the complex fluids and system interactions of food and dairy processes.

This thesis addresses these gaps by developing a generalisable Process Integration and Electrification (PI&E) methodology that combines exergy-based targeting, retrofit strategies, and techno-economic evaluation coupled with an iterative design-centric digital twin framework. The thesis is structured in two parts. Part A develops the digitalisation foundations, including the preparation of a milk evaporation case study, the creation of advanced thermophysical property packages for complex fluids (milk, refrigerants, humid air), and the construction of a design digital twin using both commercial and open-source platforms. Part B applies the digital twin to PI&E, integrating operational optimisation, Exergy Pinch Analysis, and systematic evaluation of electrification technologies in both greenfield and retrofit contexts.

For greenfield design, the research extends Pinch Analysis principles to heat pump integration by utilising heat pockets to create multiple Pinch points, enabling systematic minimisation of temperature lift and improved integration opportunities. Building on this, an iterative PI&E design workflow was developed to guide technology placement and evaluate electrification pathways. This culminated in the design of a novel fully electric milk evaporator system that achieved a specific electricity consumption of 120 kWh per tonne of milk powder, compared with 159 kWh/tp for a simpler single heat pump design, demonstrating higher efficiency.

For retrofit applications, the thesis advances PI&E by extending heat pump bridge analysis to explicitly include process unit heat flows, allowing process modifications to be considered alongside heat exchanger reconfiguration. This innovation addresses a key gap identified in previous literature, enabling more retrofit strategies. The method was demonstrated through multiple related case studies of milk evaporator plants, producing a set of common retrofit solutions. These include replacing thermal vapour recompression (TVR) and/or direct steam injection with MVR systems, which were shown to deliver lower levelised costs of heat compared with reference boiler-based designs.

The culmination of the research is a unified PI&E methodology that combines digital twins, rigorous thermodynamic analysis, and practical integration strategies. The results show that electrification of milk evaporation systems can be achieved in both new and existing plants with significant efficiency gains and competitive economics. PI&E has been tested across multiple platforms: Aspen HYSYS, DWSIM and the Ahuora Digital Twin Platform, powered by IDAES – proving to be a platform-agnostic, yet digitalisation-centred, methodology. Although developed and applied in the context of New Zealand’s dairy sector, the methods and insights are broadly transferable to other low- to medium-temperature process industries, offering a robust and scalable pathway to accelerate industrial decarbonisation.