Asset Condition Monitoring
Understanding how the machines and equipment performs in the field or on the factory floor is the asset condition monitoring role. Optimal equipment usage were tracked through the data such as temperature, vibration, and error codes but capturing it manually is hard since technicians needs to inspect the machines physically. Through IoT all data can be capture and performance is easy to monitor. The utilization of the assets can be maximize and exploit your investment with an increase of visibility.
The following architecture is for doing Asset Condition Monitoring in the industrial environments.
These are the role of each IoT Services:
- IoT Greengrass brings local compute, messaging, data caching, sync, and ML inference capabilities to edge devices. In this specific use case, IoT Greengrass provides you:
- With IoT Greengrass Connectors and long-running lambda functions to integrate with any existing industrial protocols and devices. Besides that, with IoT Greengrass you have also access to the local resources of the gateway where it is running, enabling IoT Greengrass to receive sensor data and manage device via GPIO, serial ports or any other interface. IoT Greengrass can connect with higher-level systems like SCADA or even MES to enrich the information coming from the industrial devices and also to feed data back from the shop floor to the MES bus. In the following picture you can see how IoT Greengrass can communicate with existing legacy devices.
- You can also operate offline. IoT Greengrass lets connected devices operate even with intermittent connectivity to the cloud. Once the device reconnects, IoT Greengrass synchronizes the data on the device with IoT Core, providing seamless functionality regardless of connectivity.
- Helps you reduce the cost of running IoT applications. You can get rich insights at a lower cost by programming your device to filter data locally (and even doing machine learning inference at the edge) and only transmit the data you need for your applications to the cloud. This reduces the amount of raw data transmitted to the cloud, minimizing cost and increasing the quality of the data you send to the cloud. You could even have the ETL paradigm (Extract-Transform-Load) at the edge, where you extract the data from the factory machines doing protocol conversion, you transform the data into the right format and then load (i.e. send) the data into IoT Core.
- IoT Core is a managed cloud service that lets connected devices easily and securely interact with cloud applications and other devices. IoT Core can support billions of devices and trillions of messages, and can process and route those messages to endpoints and to other devices reliably and securely. In this specific use case,
- IoT Core can filter, transform, and act upon device data on the fly, based on business rules you define. You can use the IoT Rules to be able to detect in real-time malfunctioning in equipment and redirect this information to the right service. Besides that, all the information is sent to IoT Analytics for further processing and analyzing of the data.
- IoT Analytics is a fully-managed service that makes it easy to run and operationalize sophisticated analytics on massive volumes of IoT data without having to worry about the cost and complexity typically required to build an IoT analytics platform. In this specific use case,
- IoT Analytics can enrich the IoT data received from the industrial equipment with information located in other sources, can fill the gaps if data is missing, can eliminate false readings and can perform mathematical operations in case sensors are not right calibrated.
- IoT Analytics can prepare the data to be visualized directly with Amazon QuickSight and to be analyzed with machine learning using Amazon SageMaker.
The state of industrial equipment were captured by the predictive maintenance analytics which will identify the potential breakdowns. Monitoring and infer equipment status, health, and performance in order to detect issues in real-time can be done continuously with IoT. Predictive maintenance analytics are used to make equipment to last longer, increase the workers safety, and supply chain can be optimized.
Here are some of the extra functionalities that will enables you to anticipate the equipment failure
- Amazon SageMaker is a fully-managed service that covers the entire machine learning workflow to label and prepare your data, choose an algorithm, train the model, tune and optimize it for deployment, make predictions, and act. In this specific architecture,
- Amazon SageMaker can directly apply already existing or any custom-built algorithm on top of the clean data processed by IoT Analytics. You can do statistical classification through a method called logistic regression. You can also use Long-Short-Term Memory (LSTM), which is a powerful neural network technique for predicting the output or state of a process that varies over time. The pre-built notebook templates also support the K-means clustering algorithm for device segmentation, which clusters your devices into cohorts of like devices. These templates are typically used to profile device health and device state such as HVAC units in a factory or wear and tear of blades on a wind turbine.
- IoT Greengrass makes it easy to perform machine learning inference locally on devices, using models that are created, trained, and optimized in the cloud. IoT Greengrass gives you the flexibility to use machine learning models in Amazon SageMaker or to bring your own pre-trained model stored in Amazon S3. In this architecture,
- Once the predictive model is trained in Cloud, the model can be deployed in IoT Greengrass and perform machine learning inference locally. In such way, you can run immediate corrective actions on the edge, locally, if the predicted model anticipates malfunctioning behavior, then your factory will run always on the safe side.
Extracting insights from the industrial data sources such as manufacturing equipment, environmental conditions, and human observations is what the predictive quality analytics does. Determining the actions of adjusting machine setting or using different sources of raw materials is the goal of predictive quality analytics which could help in improving the quality of the factory output. Predictive quality don’t only monitor the state of the industrial equipment and predicts failures but it also monitors the quality of the manufactured product during the procedure of the production line through adding the computer vision and machine learning.
This is the elements of the new architecture:
- Computer Vision to capture via images and/or videos the product in each of the phases. In this architecture,
- Thanks to IoT Greengrass you can connect to any simple camera, perform the required protocol translation, and transform that camera in a smart camera by running machine learning inference at edge.
- Initially, enough images and/or videos have been uploaded to the cloud and stored in S3, to be able to train a vision machine learning model appropriate to your product. This model will do the detection of faulty products automatically.
- Once the machine learning model is trained, we can deploy this model in IoT Greengrass, and run machine learning inference locally, so even if you lose connectivity to internet, you will be still able to do the inference and asses the quality locally.
Sources for content & feature image: AWS Amazon Blog on IoT & ScienceSoft
Contact us for further information