Development of an engineering simulation and techno-economic-environmental optimisation framework for industrial decarbonisation

Abstract

As policymakers and the private sector set out their targets to reach net zero in the coming decades, it is critical to assess how to reduce the greenhouse gas (GHG) emissions in a cost-effective way. Many global decarbonisation models such as integrated assessment models and individual company-level models have been developed. However, there is a relative absence of tools focusing on the industrial region/cluster level and the infrastructure that is required to enable such an ecosystem to decarbonise. To address this question, this thesis developed a simulation and optimisation methodological framework to explore cost-effective strategies to reduce both GHG emissions and costs of mitigation. This thesis focuses on industrial GHG emissions, in particular emissions related to natural gas pipeline systems and the opportunity to mitigate industrial cluster CO2 emissions by connecting and transporting liquid CO2 for storage. Besides CO2 direct emissions, CH4 emissions, from natural gas pipelines are amongst the hardest-to-abate and widespread given our legacy societal-economic reliance on gas pipeline transport. As gas pipelines are expected to be in operation for a very long time and at least until 2050, these are expected to be significant for the foreseeable future. On a positive angle, pipelines can play a critical role in industrial decarbonisation by supplying efficiently natural gas, a lower GHG emitting fossil fuel, removing captured CO2. Many industrial clusters, especially those close to CO2 storage sites, may rely on gas pipelines to help reach their decarbonisation goals. The designing and understanding of economic and GHG mitigation potential of such options is essential for decarbonisation accounting and reporting. This research developed a bottom-up engineering simulation inspired decarbonisation optimisation framework that comprises three distinct models: 1) GHG-Estimate, 2) CO2-Network-Design, and 3) GHG-Mitigation-Optimisation. GHG-Estimate is a simulation model that is designed to quantify cradle-to-gate CO2 and CH4 emissions from pipeline transport systems. It is based on thermodynamic modelling and verified using industry data from several international operations. CO2-Network-Design is a multi-period multi-objective optimisation model for the optimal design of pipeline networks, with a goal to design how to optimally connect many sources and sinks of variable size over an extended planning horizon. GHG-Mitigation-Optimisation is a multi-objective optimisation model designed to minimise the GHG mitigation costs considering alternative mitigation options and to maximise GHG emission reductions. There are several advantages the models developed in this study offer. Unlike most other decarbonisation planning models, which use linear programming to simplify the underlying engineering system, this framework features detailed techno-economic granularity using bottom-up engineering simulation, calibrated using data from industry. Thanks to the completeness of models, as well as the broad choice of user inputs, the tools developed can be applied to a wide range of uses cases. Moreover, the three underlying models can be used separately or used together. The individual models also explicitly account for and represent the key decisions relevant to decarbonisation. These include the mitigation options available, the timeframe of decarbonisation implementation, budget constraints, cost assumptions, carbon price assumptions, technology availability assumptions etc. The modelling framework developed uses a multi-objective approach, rather than the simpler least-cost objective found in most other models. The benefit of the multi-objective approach is that enables users to explore the Pareto curve balancing the trade-off between cost and mitigation goals. In doing so, the modelling tools developed avoid introducing arbitrary assumptions regarding monetising GHG emissions, which is especially important when an effective carbon trading market or carbon tax is not in place. The models developed in this work are validated through implementation using real-world data provided by industry partners and publicly available data, as well as benchmarking with independent literature, where possible. The most important conclusions from these are as follows: • Granular emission modelling (spatial, temporal, individual system components) offers great insights in understanding GHG emissions dynamics and mitigation strategies. Most of the emissions from pipeline systems come from compression related energy consumption CO2 emissions, while CH4 emissions come from a wide range of sources. The early and late years of a pipeline operation have higher emissions intensity than the years of normal full scale operation due to lower system efficiency. The cradle-to-gate emissions at markets located at a short distance from the start of a long distance pipeline can be less than half of the emissions at markets at the far end. • Appropriate clustering of emitters can effectively reduce unit costs and increase cost-effectiveness of CO2 transport networks. • Unlike the single objective setup used in most mainstream decarbonisation planning models, multi-objective optimisation can bring multiple benefits, including more options for users to flexibly choose from, avoidance of arbitrary assumptions on monetising emission objectives, providing dynamic marginal abatement curves. • The impact of key system considerations (technology preferences, carbon price, cost pass-through, policy subsidies) can be explored with the model using scenario analysis. In the implementation presented in this research, it is found that delaying the transition, preference of renewables, higher carbon prices, and a higher rate of cost passthrough would help reduce the unit cost of decarbonisation for an industrial cluster. • Group interest of the whole cluster and individual interest of each contributing emitter can be effectively respected at the same time through constraint setting in the optimisation. The three models developed are designed such as to support several decision-making processes for decarbonising industrial emissions. The GHG-Estimate model provides a reasonably accurate and fast way of estimating natural gas transmission pipeline GHG emissions even when detailed proprietary and costly monitoring data are not readily available. The CO2-Network-Design model provides a framework for the optimal design of large-scale gas pipeline networks, while minimising cost and GHG emissions at the same time. The GHG-Mitigation-Optimisation model provides a decision-making tool for developing optimal long-term decarbonisation strategies at both project level and industrial cluster level. The Python code for these models is open-source and available on GitHub to enable and encourage use by industrial players, investors, policymakers and to support future academic research in this direction.