Power electronics (PE) is a branch of electrical engineering that focuses on processing, controlling, and efficiently converting electricity from one form to another. Unlike electronic systems that process signals and data, power electronics control large amounts of electrical energy. Basically, our lives revolve around an infinite array of energy converters, motors, etc., which are the object of study in power electronics.
The goal of building a sustainable and green economy to protect our environment and improve our living standards amplify the role of power electronics as a major agent of large-scale change aimed at transforming our society and creating new business models. However, other technological advancements are also facilitating this transformation process. Artificial intelligence (AI) is currently at a major inflection point, opening up huge opportunities in many areas of human activity, so what benefits can AI bring to power electronics?
AI learning and problem solving
AI systems are trained on large amounts of data, allowing them to recognize patterns and make decisions. They perform human-like tasks like virtual assistants or identifying appropriate content on social media platforms, among other things. Generative AI is a type of AI that is capable of creating specific content based on user requests.
It uses advanced machine learning models based on neural networks, known as deep learning models, to generate text, images, and videos. Very popular chatbots, such as Microsoft's Copilot, have been embedded in browsers to conduct text conversations that look like humans have made. Other apps, such as Midjourney, can generate images from text descriptions.
The growing importance of new technologies
In order to achieve very ambitious targets, such as vehicle electrification and digitalization, renewable energy generation, etc., the following aspects must be taken into account:
Sustainability. From the early adoption phase of basic materials to the intelligent management of repairs, recycling, and reuse.
High efficiency must be maintained not only during operation, but also in all phases of the circular economy.
The product lifecycle management process is a universal concept and should therefore be ensured throughout the product life cycle.
Incorporate additional features to make all steps from building construction to plant operation and maintenance more flexible.
Deploy power equipment to support infrastructure such as smart grids.
The point of contact between power electronics and artificial intelligence
Bischoff and Hellinger propose a 3D diagram for identifying data flows and opportunities for artificial intelligence in power electronics:
The full life cycle, from design to recovery and recycling.
System operation, including the operation of semiconductor components and related hardware in applications such as solar power plants, charging stations, electric vehicles, or wind farms.
Data use cases, including design space exploration, operator training, and digital product passports (DPPs). The DPP is a tool that carries information on product sustainability. By scanning a product QR code, information about the materials used, carbon footprint, repair instructions, recycling recommendations, and manufacturing process can be gathered. Therefore, DPP plays an important role in promoting circular economy practices.
An example: inverter certification and production
Formal certification of producing a given product is a time-consuming process that involves many steps, whether testing or evaluation. A good performance-cost trade-off solution is to use digital twin-based simulation techniques.
Siemens and UL Solutions have developed a program that does not require extensive physical testing and can take advantage of Siemens' digital twin technology. UL Solutions is a leading global provider of certification, consulting and testing services, as well as audit and software solutions. By redefining the boundaries of testing, digital simulation can provide complete knowledge of the product with unmatched fidelity and speed. As a result, certification not only becomes more flexible, but also accelerates time-to-market and enables faster innovation by ensuring quality, safety, and performance requirements.
Thanks to this approach, the metaverse is no longer an abstract concept, but a new space in which digital representations of objects are combined with physical reality with remarkable results. Siemens' plant in Erlangen (GWE) is an example of how the production site can adapt to the new manufacturing model. In fact, in addition to producing inverters and drives, the factory is also equipped with an in-house test site that allows for comprehensive testing and commissioning of new "metaverse" technologies before making them available to customers.
Artificial intelligence applications in production lines
Wave soldering is a fully automated process used to solder through-hole components onto a printed circuit board (PCB). Basically, once the components are pin-mounted on the PCB, the board is placed in a wave soldering machine. The system uses a conveyor belt to move the board through several stages and has a slot to store the molten solder and a pump to lift the solder to the underside of the PCB, creating a "wave" of molten solder. When the PCB reaches the molten solder wave, heat, flux, and capillary action create a precise and reliable solder joint that connects the pins of the component to the PCB.
At GWE, an AI-controlled inspection system allows for the reduction of false alarms in through-hole wave soldering lines. Using standard inspection methods, 5% to 25% of PCBs must be screened manually, while 98% are false positives. The AI-powered inspection system reduces the number of necessary inspections to just 2.5% to 12.5%. At the end of the wave soldering line, an automated optical inspection system (AOI) analyzes the PCB camera images to determine if there is a solder bridge or open solder joint, and thus determines whether the PCB can be accepted or rejected. By combining AI technology with AOI, it helps to improve overall efficiency and reduce tedious manual intervention.
Design of Experiments
Generative AI looks very similar to the input data it is trained on, and with some creativity, it produces almost the same output as a human-generated output. An important emerging application is the various options for scanning a particular design.
Basically, generative AI is instructed to create a set of possible designs and evaluate them through simulations. The results are then sent back to generative AI to further fine-tune and refine the design. In short, generative AI can be effective in enhancing experimental design.
Data exchange
Modern manufacturing requires devices to share and communicate operational data through interfaces. This flow of information is mediated by AI services, big data analytics, the Internet of Things, and connectivity.
A fleet of large industrial pumps, fans, and compressors are driven by low-power motors that benefit from condition monitoring. Condition monitoring is a maintenance technology that predicts the health of a machine by combining machine sensor data (which provides real-time information on vibration, temperature, etc.) and the latest machine monitoring software. The analysis here is based on a comparison of the operational data with the digital twin of the motor extracted from the electrical and mechanical models of the motor.
Not only for operators and service providers, but also for OEMs, some of which are:
Optimal maintenance based on actual conditions extends product life by up to 30%.
Regular monitoring and condition analysis reduce maintenance costs by up to 30%. As a result, downtime can be reduced and plant productivity can be increased by up to 10%.
Artificial intelligence and data analysis help digital motors identify critical processes that would otherwise generate inefficiencies and greater CO2 emissions. Appropriate countermeasures lead to lower costs and a reduced carbon footprint.
Service providers can intervene quickly and remotely with access to the latest status status, reducing downtime with minimal effort. By reporting defects and deficiencies immediately, remedial action can be taken quickly to avoid damage and safety risks. This proactive approach to maintenance enhances operational reliability and safety in industrial environments.
The intersection of artificial intelligence and power electronics holds great promise for improving the efficiency, reliability, and performance of electrical systems and manufacturing. Due to its ability to efficiently process large amounts of data, the role of AI has become critical from design to operation and maintenance, while extending lifecycle performance.
END
About Us: Power Devices_Semiconductor Components_Electronic Components_Silicon Carbide Tubes_Solution Providers_Faust Technology