plc.myforum.ro

valz · Member 3 Joined: 26 Jul 2015 Posts: 771

SIMATIC S7-1500 TM NPU Application Example for Classification – Initial Setup with pretrained NN model
https://support.industry.siemens.com/cs/document/109826893/

Guide with sample code and projects for setting up your first Artificial Intelligence application with SIMATIC S7-1500 TM NPU on the shop floor.
The TM NPU application example enables you to initially setup the SIMATIC S7-1500 TM NPU with a visual Classification use case based on a pre-trained Neural network with your S7-1500 PLC and your third-party GigE Vision camera system. It can be used as a basis for different types of use cases e.g., visual quality inspection or the sorting of products and components. The application example includes not only the evaluation process, but also covers the aspects of saving captured images to an external FTP Server (e.g. for Quality Assurance or Re-Training purposes) and offers a method for updating the Neural Network on the TM NPU during runtime with one provided on the FTP server using the HMI.

The following scheme (Figure 2-1) illustrates how the hardware components have to be connected to use this application example with all available TM NPU features, incl. the FTP communication and Webserver functionality. The SIMATIC S7-1500 TM NPU is used here centrally behind a SIMATIC S7-1500 CPU.

Artificial Intelligence application with SIMATIC S7-1500 TM Artificial Intelligence application with SIMATIC S7-1500 TM

Artificial Intelligence application with SIMATIC S7-1500 TM Artificial Intelligence application with SIMATIC S7-1500 TM

The application example provides a pre-trained and converted neural network as part of a holistic SD-Card image which includes the required configuration files and a prepared Micropython script. Depending on the network and topology used, the network configurations (e.g. IP address, subnet mask, etc.) may have to adapted (in the network.conf file). The solution can be fully controlled directly via the TIA Portal or via the prepared HMI screen. The HMI displays the results of the performed image classification on the TM NPU and can also be used for giving the trigger signal to the TM NPU for capturing an image from the camera and evaluating it. Additionally, the HMI can be used for updating the neural network currently running on the TM NPU, by selecting a (converted) neural network from an FTP Server and deploying this to the module. A step-by-step instruction also leads through the steps for configuring the GigE Vision camera. The provided Neural network is only intended as an example and not to be used for productive purposes. In order to change the machine learning model (neural network) which is executed on the TM NPU the operator must perform specific changes in the configuration files on the TM NPU, which are described in detail in the setup guide. The application can be fully tailored to fit the specific use case requirements by adapting the Micropython script of the module. The application example is prepared to also be used as a basis for Object Detection use cases, which would require adaptions to the configuration files and TIA Portal project.

Documentation and Downloads
https://support.industry.siemens.com/cs/attachments/109826893/109826893_S7_1500_TMNPU_InitalSetup_DOC_V10_en.pdf
https://support.industry.siemens.com/cs/attachments/109826893/109826893_S7_1500_TMNPU_InitalSetup_PROJ_V10.zip
https://support.industry.siemens.com/cs/attachments/109826893/109826893_SDCard_image_V10.zip
https://support.industry.siemens.com/cs/attachments/109826893/109826893_FTPServerFolderStructure_V10.zip
https://support.industry.siemens.com/cs/attachments/109826893/109826893_Multilingual_ReadMe_OSS.html

[URL=https://www.file-upload.net/download-15251316/109826893_FTPServerFolderStructure_V10.zip.html]109826893_FTPServerFolderStructure_V10.zip[/URL]
[URL=https://www.file-upload.net/download-15251317/109826893_SDCard_image_V10.zip.html]109826893_SDCard_image_V10.zip[/URL]
[URL=https://www.file-upload.net/download-15251318/109826893_S7_1500_TMNPU_InitalSetup_PROJ_V10.zip.html]109826893_S7_1500_TMNPU_InitalSetup_PROJ_V10.zip[/URL]
[URL=https://www.file-upload.net/download-15251319/109826893_S7_1500_TMNPU_InitalSetup_DOC_V10_en.pdf.html]109826893_S7_1500_TMNPU_InitalSetup_DOC_V10_en.pdf[/URL]