Nov-2024

Quantify and reduce risk for acceleration of new projects

Understanding and reducing the uncertainty associated with new sustainability projects will close the gap between actual and targeted levels of investment.

Ana Khanlari and Ron Beck
Aspen Technology

Article Summary

In the fast-evolving landscape of sustainable chemical manufacturing, vetting projects and making final investment decisions can present difficult choices. Some of these complexities include the flurry of emerging technologies, conflicting objectives, uncertain future economics, and re-shoring momentum. A lack of delivered assets operating at scale to benchmark against can hinder the progress of any new project. Decisions made along the lifespan of the project without considering future changes might be counterproductive or even put the project in jeopardy. Having a complete, quantitative system view to show the trade-offs of every decision can bring a fact-based measure of risk for the project’s owners. Through such a holistic system model, owners can examine end-to-end dependencies, evaluate market and financial projections, and forecast future returns and performance without being caught off-guard.

Today’s decarbonisation and circularity imperatives are driving the industry towards new modes of operation. Electrification and utilisation of dedicated renewable power are decarbonisation steps that come with embedded stochastic and cyclical patterns of solar and wind.

Additionally, new value chains (such as plastics recycling and renewable-based chemical synthesis) introduce uncertainties like quantity and quality of raw material supply. Other uncertainties include weather, future material pricing, supply chain disruptions, and a dizzying array of optionalities in putting together new end-to-end systems. A probability-based system model can account for these uncertainties to predict future constraints and offer mitigating solutions.

Monte Carlo simulation is a statistical approach that uses repeated random sampling to predict a range of potential results. The method takes its name from the Monte Carlo Casino frequented by physicist Stanislaw Ulam’s (method inventor) uncle. This method was introduced effectively during World War II to improve decision-making under uncertain conditions. Since then, the application of the Monte Carlo method has expanded to all fields of science, finance, engineering, and project management. Conducting a Monte Carlo analysis on the most likely future pricing of material and labour is an established application in large capital project biddings. Augmenting this statistical approach with a model of the ‘connectivity’ of a system (understanding interrelationships like materials flow, and electricity) creates a sophisticated tool to rigorously evaluate a wide range of investment and operational alternatives across a system.

There are many opportunities for de-risking new projects using a Monte Carlo systems assessment. A system’s lifecycle cost can be minimised, production maximised, and availability predicted. For example, by considering equipment failure rates, maintenance costs, and spares availability, a ‘connected’ Monte Carlo model can improve the system’s reliability. The efficiency and productivity of the design can be improved by identifying optimal places to put redundancies, bypasses, and intermediate storage elements. Lifecycle emissions can be minimised by leveraging renewable energy sources while assigning emissions an opportunity cost in the model. Finally, Capex can be minimised by predicting performance and throughput changes, maintenance intervals, and weather patterns as well as other sources of uncertainty.

Aspen Fidelis is a commercial software system for de-risking capital projects, as described above. It explores multiple system configurations during conceptual design and improves the design by performing reliability, availability and maintenance (RAM) analysis. This could include alternative conversion technologies (such as CO₂ to chemicals), alternative feedstocks, power reliability options, storage options, and the like. Creating design flexibility supported by ‘nested’ models with different levels of detail is the strength of this system. Through this analysis, subsystems initially identified as the highest project risk can be modelled in higher granularity. Once the plant is operational, Fidelis can drive asset performance management to identify critical equipment and events. Figure 1 shows how the software helps to detangle capital projects.

Performance analysis to debottleneck sustainability projects
One of the most beneficial areas of risk analysis is to understand opportunities and challenges associated with sustainability projects. Understanding risk levels and how to mitigate them can bring confidence to project owners and lenders alike. In the following sections, we will review some examples of sustainability projects that can be de-risked using Aspen Fidelis.

Carbon capture projects risk analysis
Let us assume a rigorous model of a carbon capture process. Several flue gas streams (for example, from a boiler, gas turbine, and tertiary sources) are combined and sent to a carbon capture unit for amine absorption. We assume a nominal capacity of 419.6 t/y for the plant. If we assume 9.6t/y reduced capacity due to unplanned downtime, the system must deliver the remaining 410 kt/y performance with a probability of 85% or more. From a reliability and risk point of view, the question is whether the current design can meet the performance targets and if not, what can change to meet or exceed the expected performance?

For this hypothetical process, Aspen Fidelis simulated 100 different scenarios over a lifecycle period of 20 years. The main criterion was tons of captured CO₂ per year. The software reported a best-case scenario capture of 408 kt/y and identified a list of culpability. To redress the reduced capture rate, spare equipment (in this case pumps) needed to be considered. After adding spare pumps to the model and running the scenarios again, the system would capture at least 410 kton/y with a probability of 88% (identified with a red dot on Figure 3).

Here we have simplified the model and process flow diagram. However, there are other nuances involved in designing a new carbon capture plant, including pipe flows, storage tanks, dynamic behaviour of the system, and total cost. The Fidelis model incorporates all these uncertainties to predict the future performance of the system and what actions can be taken today to ensure meeting process targets.