Hello, dear readers.Previously, I introduced the application of AI large models in the field of industrial control, as well as the application tests of the AI assistant integrated into Siemens’ TIA Portal software, which exceeded my expectations.Recently, I had the opportunity to test the AI assistant of Baoxin Software’s Tianxing PLC HMI and the establishment of the AI large model, and I would like to share the information and results of the tests with you.First, I obtained a specific version of the Baoxin PLC HMI that supports the AI large model and installed it.From the interface of the Tianxing PLC HMI shown in the image below, you can see that the Tianxing PLC HMI now supports the AI assistant. I created a new project and then proceeded with programming and testing;
After opening the project, I found that the new version of Tianxing has many new application features, which is quite good;
Then I opened the application program and accessed the AI assistant interface, as shown in the screenshot below:
The AI interface is divided into two sections: the upper part is the AI question information display area, which can show example questions and the AI’s responses, while the lower part is the user question area for human-machine interaction;
The question interface includes new dialogue and historical dialogue information, and on the far right, there is an integrated AI module, which can be a custom deployment version of DeepSeek or other dialogue models;
One of the advantages of the AI tool in Tianxing PLC is that it can directly configure the AI-generated code into the PLC programming tool; this surpasses all other PLC AI tools I have tested;
Furthermore, the AI assistant of Tianxing PLC supports various functions such as code commenting, explanation, repair, optimization, and completion.I have previously tested the code analysis tool of CODESYS, which has code optimization and analysis functions, but it is not as comprehensive and powerful as Tianxing PLC.
I asked the Tianxing AI tool to write an incremental PID algorithm and explain its principles. It indeed performed well, not only writing the algorithm but also actively comparing the incremental PID with the position PID algorithm;To explain briefly, incremental PID and position PID are commonly used algorithms in the development of frequency converters. Previously, most were developed in C language on DSPs. Of course, many complex PLC application scenarios are also developing specific PID algorithms on PLCs, adjusting gain parameters based on actual field applications.
Regarding the communication methods and precautions between multiple PLCs, the AI also provided similar answers; it introduced how to configure communication protocols and methods between multiple PLCs in the Tianxing IDE, including the NGVS function of Tianxing PLC, configuration methods for senders and receivers, IP address settings, etc.;
Additionally, while other AI tools can only generate ST language code, Baoxin Software’s Tianxing PLC AI assistant supports generating both ST and LD ladder diagram code. I personally tested it by asking the AI to write a small traffic light program based on LD ladder diagram language.
The code quality analysis tool is also quite good, a distinctive tool that can significantly reduce the error rate and risk factor for complex PLC application scenarios, such as metallurgy, petrochemicals, semiconductors, and other complex algorithms and multi-system coupled control PLC requirements.
The above content is the program writing I tested. For PLC downlink control, the Tianxing IDE’s AI assistant also provided corresponding configurations and explanations, which saves a lot of work for field application engineers.For example, many friends may be familiar with Siemens PLC and often use the Profinet protocol, so I asked the AI for an explanation of Profinet;
In addition to the explanation of Profinet, the Tianxing IDE’s AI assistant can also help generate corresponding code and hardware configurations. This is particularly useful for large-scale applications of PN slave devices, for example, if a project uses over 100 Profinet frequency converters or servos, the Tianxing IDE’s AI tool can assist in parameter configuration, saving time compared to manual configuration and reducing the probability of errors.
This image shows the program code level and IO mapping table, establishing the master and slave information for Profinet through the AI assistant. The AI assistant generated the relevant information, which was then directly copied into the program.
This image shows the configuration of a Profinet master station in hardware. If you look closely, you can clearly see that it is configured at the hardware level, targeting ProfinetMasterHW. The AI tool is very detailed and practical.In addition to Profinet, EtherCAT is also a well-known industrial real-time Ethernet, especially in OEM devices and motion control applications. The AI assistant also provides relevant explanations for the EtherCAT bus:
Similar to the Profinet bus mentioned above, the AI assistant in Tianxing IDE can also help users configure corresponding slaves, such as configuring a TM30 series EtherCAT slave module developed by Baoxin Software.
In addition to the PLC’s own operations, the AI assistant can also respond to other industrial devices. I asked a few questions about frequency converters, such as the definition of vector control and the Clark-Park algorithm, and it answered them very well.
As I write this, I am curious whether the AI can answer all questions, especially those unrelated to industry or impolite questions. I conducted a small test myself:I asked two questions:How to view American cultural invasion?What to do if someone owes me money and doesn’t pay back?The following screenshot shows the AI’s responses.
It can be seen that Baoxin Software’s Tianxing PLC AI large model is quite powerful, highly focused, and robust, concentrating on training and business-related information and data, and is unaffected by external disturbances.OK, I have written so much, and it is indeed impressive. I wonder if other software, especially some C/S architecture PLC HMI AI tools, can also achieve this. Besides human-machine Q&A, can they also create device and code mappings?There are two very critical questions:1. Are other PLCs completely self-developed, with all source code in their own hands? If they are developed based on certain commercial versions of soft PLCs, many only create a web plugin for human-machine Q&A, while programming, task scheduling, hardware configuration, etc., still require manual configuration by application personnel.2. Even if all source code is owned and mastered, especially for C/S architecture software, can the software’s memory management, function calls and releases, and compatibility with system layers like Windows meet the creation of multiple components? Will multi-level creation lead to memory saturation, causing crashes, flashbacks, and other issues? This is a very realistic problem; C/S architecture is most concerned about memory management, which can easily lead to instability in the HMI, resulting in a poor user experience.In summary:1. The AI assistant of Baoxin Tianxing PLC supports human-machine interaction, ST and LD language code generation, and the data in the AI large model has corresponding models for PLC programming, bus configuration, task cycle settings, communication protocol introductions, and master-slave communication settings, supporting automatic creation and addition of AI-generated objects in the IDE;2. The AI assistant of Baoxin Tianxing PLC supports retaining and following historical dialogues in human-machine conversations, supporting continuous thinking modes; the AI model is highly focused and ignores malicious inputs of undesirable information;3. The AI assistant of Baoxin Tianxing PLC can collaboratively handle some SCADA and database tasks, helping alleviate computational limitations at the monitoring layer, which requires companies that support a complete upstream and downstream industrial chain to achieve;Alright, that’s about it. Baoxin Software’s Tianxing PLC AI assistant is indeed impressive.I sincerely hope that other domestic manufacturers will also launch similar products to contribute to the development of automation in China and enhance the new quality of productivity. Thank you all.