With the release of WISE-PaaS 3.0, the three application development frameworks— Visualization, Asset Performance Management, and AI Framework Service—now form a complete PaaS application architecture.
To successfully integrate IT and OT and realize datadriven industrial IoT application models, Advantech launched the WISE-PaaS industrial IoT cloud platform in 2014. The platform has now advanced to version 3.0. The latest WISE-PaaS provides three application development frameworks—Visualization, Asset Performance Management (APM), and AI Framework Service (AFS)—to help users accelerate the development of AIoT services.
Advantech’s Software Product Manager, Jamie Su, describes WISE-PaaS 3.0 as, “The most comprehensive AIoT edge-to-cloud architecture possible.” The bottom layer features edge intelligence, which contains a variety of edge software such as SCADA, M2I (machineto-intelligence) device connection, remote device management, and video/image management that acquire data and upload it to the PaaS service layer. The PaaS service layer is the core of Advantech’s WISE-PaaS Industrial IoT cloud platform. It supports public networks such as AWS, Azure, and Alibaba Cloud, as well as Advantech’s WISE-STACK private cloud. WISE-PaaS provides services such as database storage and IoT Hub services, in a microservices framework.
Visualization services contain two major components: WISE-PaaS/Dashboard and WISE-PaaS/SaaS Composer. The WISE-PaaS/Dashboard is developed by Advantech based on Grafana open source technology. It supports more than 49 data sources and 51 visualization plug-ins and can dynamically switch the display of information according to the data source, markup, time range, and other variables and users can define alarm rules and notification paths. Through the dashboard, administrators can see at a glance the utilization rate and alarm reports of each area of the plant, production line, or equipment.
The WISE-PaaS/SaaS Composer is a cloud configuration tool with visible workflow that supports customized component plotting for simple and intuitive 3D modeling application and interaction; allowing users to monitor parameter changes of various on-site equipment. More importantly, the 3D model built by the WISE-PaaS/SaaS Composer can be linked with APM, SCADA, or RMM data sources. Its screens have a millisecond refresh rate and it can be directly embedded in the WISE-PaaS/ Dashboard to show real-time dynamic information in the 3D flowchart. It can also control the equipment remotely in WISE-PaaS/SaaS Composer screen.
Combine WISE-PaaS/APM Asset Models with WISEPaaS/AFS to Easily Implement AIoT for Equipment
WISE-PaaS/APM provides key elements such as asset models, management configuration, alarms and notifications, and asset data analysis, to help users reduce unplanned downtime, increase availability, reduce operating and maintenance costs, and reduce the risk of equipment failure. The WISE-PaaS/APM import can be completed quickly in three simple steps: connecting the equipment to WISE-PaaS, opening the APM Portal to configure the asset model and management configuration, and automatically generating the dashboard.
The setup procedure for the WISE-PaaS/APM asset model and management configuration is very simple. Users can set parameters such as the factory areas, production lines, and equipment through the GUI, and then configure the event rules (high/medium/low alarm conditions) and metrics (status values such as equipment run and idle), and define the management logic via a tree structure to then automatically generate the dashboard based on the topology configuration, which clearly shows the A/P/Q utilization rate of the entire plant, production lines, and equipment.
Specifically, WISE-PaaS/APM generates availability information through statistical computing. If users want to satisfy advanced analysis requirements such as automated flaw detection, they must use AI technology. For this, Advantech launched an AI unified development framework service—WISE-PaaS/AFS—which makes it more convenient for developers to accelerate the implementation of model training and deployment.
WISE-PaaS/AFS is compatible with different data sources such as image data, time series data, label data, and IT data. It provides computing resource integration (including Kubernetes, CloudFoundry, and GPU) and data source selection (including PostqreSQL, InfluxDB, MongoDB, Ceph, and WISE-PaaS/APM) on the platform, provides infrastructure services such as machine learning (ML) and deep learning (DL) development frameworks (including Scikit-learn, PyTorch, Keras, TensorFlow, ONNX), and includes an analytical development environment with Jupyter Notebook online development tools and multiple algorithm modules.
When users complete the development of the model, they can use the model lifecycle management service of WISE-PaaS/AFS to develop training task rules, configure dashboard models, complete OTA (over-the-air) remote deployment, and subscribe to pre-built algorithms or solutions to improve secondary development efficiency.
For example, if a user wants to establish an equipment abnormality monitoring system, the user must first collect equipment vibration information on the edge side, configure the settings of the asset model on the WISE-PaaS/APM in order to send the vibration information to the cloud, launch the Jupyter Notebook development algorithm in WISE-PaaS/AFS, use the APM data as training material, then enter the training task and add the data labels and labels through the dashboard.
After completing the model training, the user needs to use the inference engine to deploy to the site and then configure the event rules in the WISE-PaaS/APM asset model. Once signs of abnormality of the on-site equipment are detected, it will drive the WISE-PaaS/ Notification to notify the administrator by LINE and email, or notify the notification service system via the Restful API, linking all the elements together form a complete AIoT application scenario.
The Advantech WISE-PaaS provides a complete architecture from edge-to-cloud. Its three application services—Visualization, APM and AFS—help our industry partners extract the value of industrial data efficiently and accelerate the intelligentization of equipment and workflow management to jointly develop innovative industrial solutions based on data-driven AIoT cloud platforms to deliver results for the AIoT industry.