TwinCAT projects with AI-assisted engineering

Beckhoff has developed TwinCAT Chat for the TwinCAT XAE engineering environment. With TwinCAT Chat, large language models such as ChatGPT from OpenAI can be conveniently used to develop a TwinCAT project. This increases productivity in control programming.

Profitable use of LLMs

Large language models (LLMs) are created based on a neural network and trained with a large number of texts. LLMs have become widespread in recent years and are used for a whole host of tasks, including as the basis for chatbots or language translation tools. They also offer a number of advantages for programming and have the potential to revolutionize the development process by automatically generating and completing code.

TwinCAT Chat harnesses the new possibilities that LLMs provide: AI-assisted engineering with automated creation of – or addition to – code or code optimization, code documentation, and restructuring.

How does TwinCAT Chat work?

The global chat window in TwinCAT XAE connects to the host cloud of the respective LLM – i.e., Microsoft Azure^TM for ChatGPT. It provides a user interface and enables TwinCAT projects to be modified via the Automation Interface, among other elements.

The LLM integration is specially optimized for TwinCAT 3 users – i.e., the knowledge base has been extensively supplemented with TwinCAT-specific content. This means that specific questions can be asked directly without needing to inform the LLM that TwinCAT is being used and that the code examples are expected in structured text (ST). The generated code can then be transferred directly. This saves time and minimizes the potential for errors, which is high for manual transfer. The LLM is deeply integrated and specialized in engineering, which simplifies the development process considerably and offers a clear advantage over simply using ChatGPT in the web browser, for example, as communication and code exchange are seamlessly interwoven.

Using TwinCAT Chat functions

New practical TwinCAT Chat functions are currently under development. In addition to a chatbot interface to the extensive Beckhoff documentation system, there will also be the possibility to create and set up I/O configurations directly from the chat. There is even the option to create TwinCAT HMI controls automatically. The aim is for users to simply be able to state how they want their HMI to be structured. TwinCAT Chat then places the HMI control in the background and establishes the link to the PLC. This means that an HMI page can be generated far more quickly and easily.