Industrial Safety and Impeller Defect Detection Setup Guide

Prerequisites

1. A development edge machine with Ubuntu/WSL2.
2. A deployment machine with
   • Windows 10/11 (Pro, Enterprise, IoT Enterprise). Build greater than or equal to 19044.
   • Minimum of Intel® Core™ i7 processor (with Intel ® Virtualization Technology and iGPU) with 6th Gen or greater.
3. Azure Account with active suscription. Follow this documentation for know more about Creating Azure Account.

View the video tutorials here

Youtube Tutorials

Setup

Setting up Development machine

Note: Follow this document for setting up Windows Subsystem For Linux (WSL2) on your Windows machine.

Complete the following steps to setup the deployment machine.

1. Install Docker

   Note: Add the non-root user to the docker group by following Manage Docker as a non-root user.

2. Install Visual Studio Code
3. Install Azure IoT Tools Extension for visual studio code.

Setting up Cloud Resources

1. Create Azure Resource Group

2. Create Azure IoT Hub

3. Create Azure IoT Edge Device using Register your device section.

4. Create Azure Container Registry.

   • Note: Get CONTAINER_REGISTRY_USERNAME and CONTAINER_REGISTRY_PASSWORD from the container registry created on step Create Azure Container Registry.

     • Navigate to Settings > Access Keys and enable Admin User.
     • use the Registry name and password to update the file.

5. Setting up Azure IoT Central

   1. Create Azure IoT Central Application.

   2. Create Azure IoT Central Device for the Azure IoT Central application created on the previous step.

      • Note: Click on the Connect option on the created to get IOTHUB_DEVICE_DPS_ID_SCOPE (ID Scope), IOTHUB_DEVICE_DPS_DEVICE_ID (Device ID) and IOTHUB_DEVICE_DPS_DEVICE_KEY (Primary Key) values.

   3. Create a Device Template by following the below steps

      • Open Device Template on the IoT Central application created.
      • Click on New and select IoT Device as type to select a custom device template.
      • Provide a name for the template (eg. smart port) and click on Review and then on Create.
      • Once created, open the Device template (smart port) and click on Import a Model.
      • Select the file iot_central/app.json as the model file.
      • Click on Publish to publish the template.

Setting up Deployment Machine

1. Setting up the EFLOW on your Windows deployment machine follow EFLOW-GPU documentation.

Deploying the Solution

Note: This step assumes that you have successfully completed installing EFLOW on your deployment Windows machine.

These steps needs to be done from your development machine not deployment machine.

Connect Visual Studio Code to the IoT Hub

1. Follow Obtain your IoT Hub connection string to copy your IoT Hub connection string.
2. Open the Explorer tab by naviagting to View > Explorer on the Visual Studio Code
3. Open Azure IOT HUB from the lower-left corner of your visual studio code on Explore Tab.
4. Click on the More Action to set the IoT Hub Connection string and paste the Primary Connection String copied on step 1 on the pop up input box shown and press Enter key.
5. After successfull setup, Azure IOT HUB from the lower-left corner of your visual studio code will list the IoT Edge Devices under your Azure IoT Hub.

Building and Pushing Docker images

1. Open the cloned repo in the Visual Studio Code using File > Open Folder

2. Expand the FACTORYSOLUTIONS folder.

3. Update the .env file with the following details

   Variable	Description	Default
   CONTAINER_REGISTRY_USERNAME	Azure Container Registry Username. Refer step Create Azure Container Registry
   CONTAINER_REGISTRY_PASSWORD	Azure Continer Registry Password. Refer step Create Azure Container Registry
   DEFECT_TARGET_HARDWARE	The Model Target Hardware. Avilable options: CPU/GPU	CPU
   IOTHUB_DEVICE_DPS_ENDPOINT	Azure IoT Device Provision Service endpoint	global.azure-devices-provisioning.net
   IOTHUB_DEVICE_DPS_ID_SCOPE	IoT Central Device ID Scope. Refer Creating Azure IoT Central Device
   IOTHUB_DEVICE_DPS_DEVICE_ID	IoT Central Device ID. Refer Creating Azure IoT Central Device
   IOTHUB_DEVICE_DPS_DEVICE_KEY	IoT Central Device Key.Refer Creating Azure IoT Central Device

4. Building and pushing the solutions

   • For CPU, right click on the deployment.template.json and then select Build and Push IoT Edge Solution.
   • For GPU, right click on the deployment.template.gpu.json and then select Build and Push IoT Edge Solution.

   The above step generates the deployment manifest file namely deployment.amd64.json or deployment.gpu.amd64.json based on whether CPU or GPU template was chosen above in the config folder.

Deploying the Solution

Continue to deploy the solution to your deployment machine once all the module are build and pushed.

1. Right click on the config/deployment.amd64.json (or config/deployment.amd64.gpu.json if you want to run on the GPU) file and select Generate Deployment for Single Device.
2. Select the IoT Edge Device ID from the pop box that appears.
3. An OUTPUT window will pop up confirming that the deployment has succeeded.

Verify the deployment

The deployment can be verified in multiple ways, here we will be verifying the deployment directly on deployment Windows machine.

1. Open a Power Shell on your deployment machine.

2. Run the following command to ssh to the EFLOW virtual machine from your windows machine.

   Connect-EflowVM

3. Run the following command to list all the deployed modules

   sudo iotedge list

4. Wait until the following modules appear on the list with status as running

   • edgeAgent
   • edgeHub
   • telegraf
   • grafana
   • mosquittoserver
   • rtspserver
   • opcua
   • influxdb
   • industrial-safety

   Run the command on step 2 to recheck the module status.

The same can be viewed from

• Visual Studio Code - Azure IOT HUB panel
• Azure IoT Hub portal
• Windows Admin Center dashboard