Most conversations about simulation begin with some version of this question, “Do we need low, medium, or high-fidelity models and what’s the difference?” Onyx 360 believes this is the wrong starting point. Fidelity discussions often create more confusion than clarity and impede progress toward solving your issues. Rather than treating model fidelity as fixed categories or ‘one size fits most’, it is more effective to view it as a continuous spectrum. We take a different approach: define the problem, align the simulation to the specific need, and move forward. Quite simply, we’ve ‘cancelled’ the fidelity discussion.
Whether your plant is currently being designed or has been operating for decades, the real challenge isn’t choosing a level of model rigor; it’s understanding the problems you’re trying to solve. Startup delays, operator errors, inefficient operations, and missed performance targets stem from gaps in how systems behave, how people interact with them, and how decisions are made. Dynamic simulation provides a way to explore and address these challenges in a safe and controlled environment and helps you better understand the connection between your technology, people, and processes. But to realize the full value of simulation, we must shift the focus from model categorization to defining the expected outcome and required results for each of your applications.
Onyx 360 focuses on translating your intended use case to model requirements that can evolve based on what you are trying to achieve. We apply simplifications or data-driven models where appropriate to design a solution that maximizes your return on investment. Some applications require simple, repeatable responses to validate control logic. Others require realistic system behavior to test procedures, train operators, or design improvements that will be deployed in production. The key is not selecting a predefined level of model accuracy or rigor, but aligning the simulation to the decisions, behaviors, and outcomes required for your specific use case. To follow are a few real-world examples where dynamic simulation was ‘right sized’ to solve tough challenges.
Improve Software FAT Quality and Efficiency
Software Factory Acceptance Test (SFAT) is a critical phase in a process control project where users verify that everything in their control process is configured as expected. Traditionally, SFAT requires forcing simulated values to test interlocks, sequences, and control strategies. This is a tedious and time-consuming process especially with modern, interconnected processes that require regression testing as changes are made to adjacent systems. Because the objective is to validate control logic and configuration rather than replicate the full process behavior, simple simulation responses are often sufficient.
Simple loop simulation supplemented with automated regression testing tools supplies additional value to FAT by reducing setup time and improving test repeatability. This level of simulation allows you to test simple interlocks, verify that loops are configured correctly, check pumps and discrete behaviour and fail states. Additionally, simulated loop responses improve testing quality by reducing human error and testing bias (i.e.: the team who develops the control logic then runs manual tests for accuracy by forcing values).
In other cases, simple loop responses may not be sufficient for your needs — even when supplemented with automated testing tools. More complex control strategies such as batch, sequences, or supervisory control require higher accuracy models that incorporate flow, pressure, and mass conservation to adequately test control logic. Testing a simple control sequence for recirculating tank contents with a pump also requires a more rigorous level of process simulation. For example, if the sequence starts the pump, opens a valve, but fails to set the pump speed or open a flow path, a more accurate model will show that no flow is returning. Finding and troubleshooting small but critical issues like these in a simulated environment before encountering it during startup saves costly engineering hours and helps ensure fewer delays.
Reduce Time to Startup with Virtual Commissioning & Start Up
“Connected Performance is not just about observing the current state of a physical asset — it is about engineering the confidence to predict and control its future.”
The final phase of any automation project is where the control system is turned on and exercised against the real process for the first time. This is when your risk level is at its highest. At this stage, issues are no longer theoretical. Every gap in logic, sequencing, or process understanding becomes an immediate problem to solve and often requires you to mobilize a full onsite team.
— ONYX360 Engineering Team
Traditionally, even after a successful software FAT, the control system sits largely untested against realistic process behavior until startup. As systems are brought online and sequences don’t behave as expected, teams work reactively to stabilize operations which often results in a series of troubleshooting “fire drills”. The associated cost of extended site presence, delayed startup, and lost production can be exponential.
Virtual commissioning and startup can solve this by moving the first true test of the control system earlier in the lifecycle. Instead of scheduling the control configuration after SFAT, the control system is connected to a simulation environment capable of exercising full operating procedures from initial energization through ramp-up, steady state, and shutdown. This allows you to shift your focus from validating control logic to validating how the system is behaving during real operating conditions.
In this example, the accuracy of your simulation models is critical. To execute startup procedures and validate sequences, the simulation must behave realistically across a broader operating range. This requires us to incorporate key process characteristics into the models such as:
- Transient responses as systems heat up, pressurize, or react
- Thermodynamic behaviors that impact phase changes, energy balance, or reaction rates
- Logical process constraints such as purge sequences, interdependencies, and material movement
The engineering tendency is to drive towards maximum theoretical behaviors, but this is a trap. Tuning model behavior to more closely match reality exponentially increases the necessary effort and cost. We’ve proven that the simulator needs to drive sufficient behavioral accurately enough to expose how the control system and process interact during real operating scenarios. This allows you to run startup procedures multiple times before ever reaching the field, identify potential control logic issues early in the game, and validate sequences under realistic conditions.
What would have been discovered during a high-pressure startup becomes a planned problem-solving exercise in a controlled environment. The result is a fundamentally different startup experience. Instead of reacting to problems in real time, teams arrive prepared having already exercised the process, aligned on procedures, and validated system behavior. Edge cases can be explored without consequence. Operators get practice reps with startup procedures. This translates to reduced time to startup, lower commissioning costs, and more efficient transition to stable, revenue-generating operations.
Drive Operator Adoption with a Platform Training System
While virtual commissioning and startup help you ensure your system is ready for production, it also creates an opportunity to prepare your operators. The same simulation environment that was used to validate control logic and procedures can be extended to train operators on how to interact with your unique system before they ever touch production.
As control systems become increasingly more complex, having a safe way to train operators is imperative. There are several limitations to traditional platform training approaches. First, platform training is rarely specific to your actual configuration which means you will need to overcome this knowledge gap. Second, training is typically a one-time event and often occurs months before startup which leaves ample time for knowledge decay. Third, new operators are hired and the cycle repeats with the same limitations.
Alternatively, a simulator with even simple control responses provides operators with an opportunity to learn your specific control platform including your graphics, alarms, interlocks, and control strategies. And allows them to interact with them in the exact way they will use them in your live plant. Training courses can also be designed for your simulator prior to deployment, creating a turnkey, repeatable training program tailored to your operation.
This same simulation environment can also act an offline platform to safely test control system changes and demonstrate to other operators without introducing risk to production.
Upskill Operators & Transfer Years of Site Experience
Most plants don’t realize they have a knowledge gap until it’s too late. An Operator Training System (OTS) that simulates the steps required in a specific control sequence allows you pretrain operators to handle infrequent events, including those that require a multitude of steps and years of practice. More rigorous models enable you to provide instruction for complex procedures like startup, help operators understand what causes a plant trip, and learn how to avoid or recover from a costly shutdown event.
To provide this level of training, the simulation must behave in a way where operators believe they are interacting with the real plant. Subtle process responses, timing of events, and cause-and-effect relationships are what help operators develop contextual understanding of how your system behaves. Realistic, hands-on practice also builds confidence in their decision-making abilities in stressful, real world operating conditions. In most cases, this type of training simulator requires first-principles process modeling including fluid dynamics, reaction kinetics, and thermodynamics. Not for the sake of theoretical accuracy, but to ensure the simulator responds realistically as operators make decisions and interact with the system.
Working with seasoned operators to develop a system that represents the reality of your plant is critical to the successful use of simulation. Their years of knowledge, including subtle indicators and decision-making patterns, help you create training scenarios that go beyond procedural steps and reflect how your plant realistically behaves. This approach enables organizations to capture and transfer knowledge that traditionally takes years of on-the-job experience to develop. It also provides you with a repeatable way to effectively train operators (or add new ones) for years to come.
Provide a Safe, Predictable Way to Optimize Performance
Simulation can also create a predictable way to optimize performance with zero impact to production. Unlike earlier use cases, the objective here is not just to validate behavior – it is to design and test strategies that will be deployed in a live environment. The level of model rigor required is dependent on the problem being solved and the behaviors that must be represented.
In all cases, the simulation must respond with a level of accuracy that ensures solutions developed offline translate directly to production. This may require incorporating system characteristics such as piping isometrics and friction losses and their impact on header and pump dynamics. For example, a dynamic simulation of a complex system such as utilities (including boilers, commercial Cogen systems, etc.) enables you to accurately replicate real system dynamics in a safe, offline environment. Using the simulator as a ‘sand box’, engineers can design and test improvements such as feed-forward loops that reject disturbances caused by load change.
This same approach can be extended across other systems such as distillation, reaction units, or blending operations where process interactions and disturbances must be fully understood before implementing changes. In each of these cases, simulation is designed to capture the behaviors that matter, ensuring that optimization efforts are both effective and predictable when deployed in production.
Conclusion: A properly designed simulator can help you get to startup faster, reduce risk during commissioning, improve operator readiness, and provide a safe path to optimize performance. The value is not determined by how “high fidelity” a model is, but by how well it supports the specific decisions, behaviors, and outcomes required for each use case.
Rather than debating levels of model rigor, the focus should be on clearly defining the problem, understanding the intended application, and translating those requirements into a right-sized simulation solution. This includes not only the appropriate modelling approach, but also how the simulator is deployed, used, and maintained over time to continue delivering value.
Onyx 360 eliminates non-productive fidelity discussions and instead focuses on your specific applications, challenges, and expected outcomes. By doing so, we help organizations align simulation capabilities to real operational challenges and implement solutions that deliver measurable impact today and can be scaled for future needs.