Home

First Principle Models to bolster the Digital Twin Panorama

Richard Porter, Principal Engineer, Simulation, Yokogawa Australia Pty Ltd

Richard Porter, Principal Engineer, Simulation, Yokogawa Australia Pty Ltd

We would all like to know the future. Knowledge of the future changes how you act in the here and now. The problem is, how do we do it, how do we see the future?

With the increasing requirement to achieve “more with less,” the need to predict the future and act accordingly is greater than ever before. With the advancement of technology, it is now possible to answer the question, “How will my business respond to inputs in the next few hours, for the rest of today, tomorrow, and beyond?”

Digital Twin technology has advanced to the point where it can now be used to answer these questions in a consistent and timely manner. Successful Digital Twin projects have shown that users can now have confidence in the technology and that the results are reliable. This is especially true for the process industry where Yokogawa has been providing products and solutions for over 100 years.

The core technology that powers Yokogawa’s successful Digital Twin projects is First Principles Mathematical modeling. A first principles model applies the laws of physics, such the movement and properties of materials, to calculate how a process behaves. A good example of first principles models can be seen in Car Driving computer games that took a great leap forward in realism when they began implementing physics engines to calculate the behaviour and response of vehicles.

First Principles modeling has been available to the process industry for many years. However, it has typically been applied to off-line applications. The great leap forward for the process industry is the ability to synchronize a model to a real plant. As a result, it is now possible to apply this technology on-line in a Digital Twin to gain significant insight into the plant operation. The insight gained can be used to derive improved performance and profitability, among other things.

WHERE BIG DATA METHODS MUST ANALYZE THE DATA TO DETERMINE THE RELATIONSHIPS BETWEEN THE MEASURED VARIABLES, A FIRST PRINCIPLES MODEL BY ITS VERY DESIGN INCLUDES THE INVIOLABLE RELATIONSHIPS BETWEEN ALL PARTS OF THE PROCESS

First principles-based mathematical models can be applied to a variety of problems. This versatility is seen in the range of solutions Yokogawa has developed that use mathematical models at their core. Along with Digital Twins, these solutions include operator training simulators, real time dynamic optimizers, and plant design software. Of note is that the same mathematical model, once developed, can be used in each of these solutions. As has been succinctly described, it is “One model, many uses”. Implementing these, other solutions can be achieved with less cost by reusing the model developed for the Digital Twin.

A Digital twin built on first principles models can be used throughout the lifecycle of a plant. Unlike Big Data methodologies, this type of Digital twin is not dependent on and constrained by the available data. Where Big Data methods must analyze the data to determine the relationships between the measured variables, a First Principles model by its very design includes the inviolable relationships between all parts of the process.

A key element in deriving benefits from a Digital Twin is visualization of the data in a manner that is meaningful for users. The vast amount of data generated by the calculations need to be distilled into actionable information. One such visualization method utilizes the digital twins’ ability to calculate the plant condition faster than real-time. By calculating ahead, it is possible to show a trend of how measured variables, such as product quality, will change over the next minutes and hours. Users can then respond appropriately.

Significant benefits for a business can be gained by comparing the digital twin performance, which can be run to show the ideal operating condition of the plant under any set of external conditions with the real plant performance. In doing this, underperforming assets and poor practices can be clearly identified and their effect quantified, once again informing and facilitating good decision-making.

Digital Twins built using First Principles Mathematical Models at their core have been shown to provide insight into the current and future behavior of a process plant. The insights gained by being able to visualize what is happening and forecast what will happen will help reduce the psychological burden of the operators. Situations can be handled well in advance with the margin provided by predictive alarms. Abnormalities can be found in the parts of the plant where there are no measurements, and it is possible to take early measures to avoid dangers and irregularities. It is conceivable that visualization and preventive operations will allow stable, safe, and environmentally and economically optimized operations of unprecedented levels.