Digital twins for sustainable paper production

In the paper industry, optimising energy consumption is a central factor in making production processes more sustainable beyond established paths. Digital twins are a valuable and powerful tool for these efforts. However, their potential is still far from being fully explored in the paper industry. Partly because there is still no unified understanding of the concepts, applications and benefits of this industry 4.0 technology.

Digital twins are new solutions for the development and operation of products and product systems, of machines and entire production systems. However, they are more than just a digital copy of their physical object and offer a wide range of advantages: from the provision and analysis of current operating conditions to the prediction of future behaviour and the real-time optimisation of processes or virtual rescheduling during operation.

The potential of this tool is not yet fully exploited. As a 2020 study by the Fraunhofer Institute for Production Systems and Design Technology (IPK) shows, digital twins are mostly used for data provision or for validation and error analysis. However, they are still rarely used as a basis for automated value-added services or for designing autonomous or adaptive systems. One reason for this is that the term is a catch-all for various concepts at different stages of maturity.

What is a digital twin?

Digital twins are virtual representations, often described as a digital image of a real product or system. They are used to identify optimisation potential in complex production processes or entire systems, but also to test potential changes before they are transferred to the real world.

In practice, they come in a wide variety of forms, depending on the physical object under consideration and the specific application. The reference architecture model RAMI 4.0, a kind of 3D map that systematises the technology approaches, provides orientation. According to this, digital twins can be distinguished at the hierarchy level, the application level and the product life cycle.

• Hierarchy level: Digital twins can be developed at the level of individual components as well as representing complex multi-systems. In the case of a paper machine, the digital twin can refer to individual pumps or paper rolls or map control loops of machine sections, right up to modelling entire paper machines and supply chains.
• Life cycle of an entity: The life cycle of the physical asset under consideration also plays a role in the design of the digital twin. The entity is characterised by certain states and parameters throughout its entire ‘life’. These can be operating, status or process data, which are recorded by sensors, for example. Thus, the digital twin provides a wide range of added values from product development to operation and decommissioning.
• Application level: How the digital twin ultimately works and how it is constructed is defined by a third dimension – the specific application. A distinction is made between the ‘digital shadow’ and the ‘digital twin’. If the state of a real object or system is recorded and visualised during the various phases of its life, it is usually referred to as a digital shadow. If the aim is to predict the behaviour, function or other properties of products, machines or processes by integrating simulation models, the Fraunhofer IPK refers to this as a digital master. The full potential of a digital twin is realised through the intelligent linking of master and shadow. The exchange of data between the real object and the digital copy, as well as the use of additional tools such as exploratory data analysis or optimisation algorithms, create real added value.

The above classification of digital twins not only highlights the wide-ranging applications of the technology, it also illustrates its complexity. Whether product, machine, plant or process twin – the forms and characteristics are manifold. The market research institute IOT Analytics has identified over 250 different types of digital twins.

How can a digital twin support the paper industry’s sustainability goals?

The potential of digital twins for the paper industry 4.0 is as diverse as the range of possible applications is broad. At the lower hierarchy level, simulation models help optimise sub-processes, while digital product passports can visualise CO₂ emissions. On the other end of the range, a detailed digital twin of the entire papermaking process helps process engineers and machine operators to better understand the complex interdependencies in the process, thus driving holistic optimisation. These benefits may be tapped at different levels with different objectives.

• Developing and evaluating innovative processes

If the paper industry is to achieve its energy-saving targets, innovative new processes are needed in paper production. However, a comprehensive evaluation of these processes is a must before they can be introduced. With the help of a virtual model, process innovations can be simulated and evaluated in detail. When these submodels are added to the holistic digital copy of a paper machine, the system behaviour as a whole can be predicted. This makes it possible to make statements about potential energy savings and to evaluate developments, like those being researched as part of the FOMOP research project, faster and more accurately. The advantage: the introduction of new processes into industrial paper production is thus accelerated.

The development of new process solutions itself is also accelerated: if they are first tested in virtual models before being tried out in laboratory and pilot tests, this not only saves time and money. It also significantly expands the potential search space for new ideas.

• Virtual commissioning and optimised production start-up

Commissioning a new unit or using a new raw material is a complex process. It often takes weeks before the new process is up and running optimally. If the commissioning is carried out virtually for the most part using a digital twin, downtimes and the start-up phase, during which the process is not energy and resource efficient, are reduced enormously. New technologies can be introduced faster at lower cost, reducing energy and resource consumption and accelerating return on investment.

• Visualising CO₂ emissions in production

The European Union wants to use the digital product passport to promote the industrial circular economy. The first product groups are to be equipped with it from 2027. The digital product passport records and collects data throughout the entire product life cycle. Created as a digital twin, it can be used by the paper industry to store not only general data and typical product information but also detailed data on a product’s CO₂ footprint. This means that for each roll of kitchen paper produced, it is possible to quantify how much CO₂ was emitted for raw materials, paper production and transport.

• Automated root cause failure analysis and optimisation in operation

A paper machine is a complex system in which even small errors and defects can have a major impact on product quality and efficiency. Finding them is sometimes a major challenge. Machine and plant twins map the current state during production. The automated comparison of the actual state with the simulated target state facilitates and accelerates the search for the causes of errors. This makes it possible to sustainably reduce waste and energy consumption.

Digital twins also improve the control of plants in terms of energy efficiency and CO₂ emissions. In their daily work, machine operators usually have little leeway to test process optimisations. Detailed process models of the entire system make it possible to virtually vary the operation of paper machines and identify optimal operating modes using explorative analytics. In the digital copy, the current data of the real machine – for example, the current raw material mix, the condition of screens and felts, or the availability and costs of sustainable electricity – can be used as the starting point for the simulation. This makes it possible to observe small changes in operation and derive effects on the entire system.

• Anomaly detection and predictive maintenance

Predictive maintenance promises faster, needs-based maintenance, especially in paper production. Defective system components can be detected more quickly by comparing the simulated target and actual states of the paper machine. This saves costs and reduces downtime compared to routine or time-based preventive maintenance.

Furthermore, resources for maintenance work can be better planned. With the help of the digital twin, the condition of plant components such as gearboxes and pumps, but also of consumables such as felts, can be predicted. Machine operators can thus calculate how the declining condition will affect process performance and plan the optimal time for replacement. This increases plant availability while improving spare parts handling.

The goals of the FOREST project

Industrial paper production is a highly automated process in which closely coordinated manufacturing stages are intertwined. The possible applications of digital twins are extremely diverse, especially here. With the ambitious goal of reducing the energy required for paper production by 80 per cent in mind, the FOREST research project is developing a scalable framework for digital twins that will enable many of the described applications in a modular fashion. This is necessary because the requirements and possible designs of digital twins vary greatly from company to company, depending on the plant, the age of the paper machine and the type of paper.

For the first time, virtual models are being created for the sub-processes of paper production, from stock preparation to winding and the energy supply systems. These models virtually map the processes and energy supply of a real paper mill and record the energy and material flows as well as the CO₂ footprints down to the sub-process and sub-product level. In this way, the processes of a real paper machine can be digitally mapped in an overall system and the transformation spurred on.