Artificial Intelligence (AI) is becoming increasingly important in the production industry. Integrating AI into enterprise operations can lead to increased productivity, better product quality and improved resource management. However, many projects only result in proof-of-concept solutions and lack management over a longer period of time. Transforming prototypes into usable solutions poses challenges for many companies. Only a small percentage of companies have been successful in handling multiple use cases simultaneously.
A study from the International Center for Networked, Adaptive Production (ICNAP), (comprising three Fraunhofer Institutes) involved the development of an organisational process framework that classifies all AI processes occurring within a company in a structured manner. It contains roles, processes and best practices for AI within the company. Part of the framework is a technical reference architecture provides a modular integration of the use cases into the software and data infrastructure of companies, as well as a basis for the deployment and monitoring processes.
The framework and architecture supported a systematic implementation of model deployment and monitoring. They were both successfully validated based on two use cases that come from the areas of predictive maintenance and predictive quality.
SENSOR NETWORK SOLUTIONS
Wireless sensors are becoming more common due to advances in semiconductor, material science and networking technologies. They are traditionally powered by batteries, which limits their lifetime and implies a high cost and timing effort whenever they need to be charged or changed. Maintenance for huge wireless sensor networks seems to be unmanageable in the future.
A self-sustainable sensor was introduced to address this issue, as part of the study. These sensors can harvest small amounts of energy from the environment and convert it into electrical energy for their operation in a process called Energy Harvesting. The study investigated different power solutions in terms of performance in laboratory and industrial scenarios to showcase how wireless Industrial Internet of Things (IIoT) can be self-sustainable.
Several energy solutions designed to enable a self-sustaining IIoT in the future were tested in different scenarios. At the beginning, the concept for a demonstrator was developed and implemented. Here, the hardware components were determined in cooperation with the ICNAP community and combined into a system. Various energy harvesters and a wind turbine were tested as power generation solutions. During the tests, the largest amount of energy was generated with a vibration harvester. However, it is not possible to make a general statement about the efficiency of the individual solutions as this is dependent on the application.
PLUG AND PRODUCE
The plug-and-produce concept aims to simplify the integration of new equipment or systems into existing production processes. It involves the use of standardised interfaces and protocols, allowing different machines and devices to connect and communicate with each other. This plug-and-play approach enables manufacturers to add new components or upgrade existing ones without extensive reconfiguration. The goal is to enhance efficiency and interoperability in the manufacturing environment.
One of the primary challenges hindering the widespread adoption of plug-and-produce in the manufacturing industry is the lack of standardisation across manufacturing processes and equipment and the absence of standardised interfaces and protocols. Accordingly, each integration requires custom development and configuration, resulting in additional time and costs.
ICNAP’s study explored plug-and-produce in the context of the adaptable factory for the manufacturing industry. It introduced the concept of interoperability and distinguishes between its different levels. Furthermore, the analysis of use cases shows the potential of implementing plug-and-produce.
Additionally, the use case ‘plug-and-produce for control systems managing self-contained production resources’ was chosen for a deeper analysis. The implementation of plug-and-produce could be demonstrated for a bot and the control software COPE (corporate-owned, personally enabled) of the Fraunhofer IPT. The selected use case should be seen as a starting point and was chosen because existing standards already addressed most aspects of it. However, it does not encompass the entire concept of the adaptable factory, but it potentially provides reusable concepts for comparable use cases.
The benefits of having a physical twin of an application are numerous, but the costs are prohibitive. Therefore, the digital twin was introduced as a concept to create a digital representation of a physical object, machine or system in real-time.
However, integrating digital twins in the manufacturing sector depends on high quality and reliable data. If the data used to create and update a digital twin is incomplete or incorrect, this can affect the accuracy and reliability.
The report delved into the challenges, benefits, and practical aspects of integrating digital twins into industrial processes. The aim was to develop a digital twin demonstrator that sheds light on suitable technologies, their readiness levels, current limitations and potential future benefits.
The hardware selected for the demonstrator was the fischertechnik Learning Factory 4.0, which provides a cost-effective solution for demonstrating industrial processes with machine-product interactions. Automation of the machines in the demonstrator is carried out using the Siemens S7-1500 PLC, an industrial programming logic controller. On the software side, the digital twin components consist of a standardised data interface and technology components. The standardised interface, based on the Asset Administration Shell deployed on a BaSyx server, enables interaction and data exchange among various components. It offers features such as asset registration, management, provisioning, metadata management, and security/access control. The report highlights the benefits of the standardised interface.
In the digital twin demonstrator, these benefits are displayed by utilising self-contained software components that can interact with each other.
ENERGY MONITORING FRAMEWORKS
With the current global trends in climate change and carbon dioxide emissions, it is increasingly essential for any business to reduce energy usage and become self-sustainable. Manufacturing sector industries are facing pressure as they contribute to 30-40% of total energy consumed. With the increasing energy prices fuelled by the geo-political situation, many businesses are enduring heavy losses and going bankrupt. To tackle this, organisations must implement an energy management plan. Energy management is the proactive optimisation of an organisation’s energy consumption to conserve use and decrease energy costs.
As technology advances and awareness of the importance of energy conservation grows, the integration of energy monitoring systems will become increasingly commonplace.
The full report can be read at: www.tinyurl.com/msh8xmt5