Managing Complexity Using Model-Based System Analysis

Such methodologies can also be used to prevent these faults from ever occurring. Analogue modelling methods have historically been used for this purpose, for instance through the construction of a prototype in a laboratory setting where different faults can be ‘injected’ into it, allowing for the assessment of how the prototype system reacts. 

Yet with the advance of technology comes new ways to perform model-based fault analysis. Digital methods, including causal and acausal analyses, can be used by organisations to assess how their systems respond to faults introduced to it. 

What benefits does system modelling possess when carrying out fault analysis? And what checks does it provide on the safety of critical systems? 

Virtual modelling 

Some advantages of modelling faults in a system virtually are obvious. It allows the modeller to replicate real-life scenarios in a cost-effective and deterministic way, with no need to build physical prototypes within a lab environment. It also allows the ‘injection’ of fault scenarios into the system, which can be extremely difficult to replicate under real-life conditions, giving a better understanding of what defects might occur and how this would impact the system at large. But how exactly is this achieved? 

Tools such as MATLAB Simulink can be used to easily construct a replica system digitally which accurately reflects the physical system to be tested. Differential and algebraic equations can be used as simple and reusable building blocks with which complex models can be created with relative ease. The key aspect of MBSA is to capture not only nominal functionality, but also faulty behaviour and allow for appropriate reasoning on the physics of failure.

Professional modelling tools allows for precise control over how models are built, where faults are injected, and hence a more accurate picture of what the effect on the wider system would be, especially in terms of mapping procedural logic and software-related diagnostic functions.

Knowledge-efficiency

Critical systems requiring extensive modelling are naturally complex and interconnected with other equally vast and sprawling systems. With this in mind, modelling such systems and assessing their capacity for faults on a physical prototype, or even in one’s head, is subject to the risk of human error or knowledge drain. 

System modelling can overcome this challenge. Model-based tools can be used as data management centres, covering requirements, functions, fault-modes, diagnostic symptoms, technical solutions, components, interfaces, software code and more. 

Keeping data from these injections in a centralised way provides two main benefits to those performing fault analysis on critical systems. It ensures information regarding the effect of faults on systems is collated in one place. It also increases the longevity of knowledge by ensuring information remains within an organisation and doesn’t leave with employees who are responsible for the analyses. 

Design-efficiency 

Adopting system modelling can also help reduce any errors in the design of a system. Initial virtual system representations can be implemented during the design phase of a project and enhanced throughout the entire product lifecycle (PLC). This can help connect ‘missing links’ that occur between different system levels, functionalities and hardware units which form part of the system. 

This proactive approach towards modelling can reap a number of benefits, from reducing the risk of costly system failure later on to allowing those working on the system to have a greater understanding of what the system and its components does, as well as the relationship between those components. This is particularly useful when modelling is being used to verify a software development methodology, as opposed to simply assessing a system’s safety, with tools like Ansys SCADE enabling consistent representations of internal data flow and contributing towards the production of safe software coding. 

A model approach

At the heart of effective critical system modelling is the aim to build, as well as maintain, safe systems. System modelling forms a single source of truth for critical systems, meaning knowledge of how the system is constructed and its interrelated components remains constant and does not fluctuate with the comings and goings of an organisation’s employees. Through keeping a detailed visualisation of systems, they can be kept in compliance with relevant regulations and built with these regulations firmly in mind, with system modelling reducing the amount of effort needed for V&V if performed consistently throughout the development process. As a result, a system’s “digital twin” is born.

For models to be truly effective, you’ll likely need a partner with experience in maintaining critical systems in a range of industries where safety and reliability are paramount. Check out Critical’s white paper on model-based safety analysis for high integrity systems here.