How cloud computing works for industrial processes
Cloud computing is just one of the components for Big Data implementations in industrial process and facilities. While it can be particularly effective for remote monitoring applications, the data must first be collected and pushed to the cloud, then stored and analyzed to create usable information.
This usually involves several steps, starting at the edge device and ending with information delivery to end users. Modern human-machine interfaces (HMIs) play a crucial role, making connections to smart edge devices and controllers, and filtering the data before pushing it to the cloud.
Cloud computing defined
Cloud computing can be defined as the delivery of computing services-such as servers, storage, databases, networking, software, analytics, and more over the internet. There are many cloud service providers, and most typically charge for cloud computing services based on usage, like a monthly utility bill.
With little or no information technology (IT) management effort required, cloud computing provides network access to a wide variety of computing resources. Cloud computing is a little more than a decade old and includes common technology for tasks such as hosting websites and blogs, storing data, and streaming audio or video. The software applications can be a simple as email, calendar, and office tools. The cloud also can be used to deliver software on-demand, and to provide computing capability to analyze data.
For example, data can be pushed to the cloud, and the resulting patterns can be used to make predictions. With the appropriate data in the cloud, Software as a Service (SaaS) and Infrastructure as a Service (IaaS) often are used to provide cloud computing services to store and analyze data. The returned information or analysis results then can be used by the HMI on the plant floor, or by any device with an internet connection, such as a laptop, smartphone, or tablet.
Collecting and filtering data
While analytics, historian, enterprise resource planning (ERP), and other systems may be hosted in the cloud, these applications need plant floor data. This data starts at the edge, supplied by either a sensor, or a smart edge device such as a smart instrument, a power meter, a variable-frequency drive, etc. These components are connected to a controller, to a PC-based HMI, or other HMI. A new trend is embedding HMI functionality in the smart edge device, allowing data to be processed at the source.
Wherever the HMI is based, it can filter data before pushing it to the cloud or other storage area. It is not necessary to collect all the data every second, or even every minute, for some applications. The data may need to be saved only when it exceeds certain values or moves outside a defined range. Some modern HMIs can be configured, using faceplates, to filter and consolidate the data before pushing it to the cloud (Figure 1).
An HMI provides most of the communication with controllers and other edge devices using built-in drivers. Not only can it communicate with hundreds of devices, it also can communicate with the cloud, ERP systems, and local or remote databases. In addition to filtering and communicating, the HMI can buffer the data if a network or cloud connection is unavailable.
Hosting the HMI
A modern HMI can be hosted on many platforms. It can be installed on a desktop PC, a flat-panel industrial PC, a thin client, an embedded computer running Linux or some other operating system, or a smart edge device. This portability from one platform to the next is vital because it allows the HMI to be deployed on the best platform for the application.
There are low-cost platforms, such as microcontrollers, for hosting an HMI used for gathering data and pushing it to the cloud. A low-cost HMI, with or without a display, can collect data from a variety of devices, filter the data, and push it to the cloud. A smart edge device running an embedded HMI can push data to the cloud without connecting through a controller or a full-featured HMI.
Once an HMI pushes data to the cloud, it can be accessed by any internet-connected device, but first it often makes sense to analyze this data and turn it into actionable information.
The cloud provides reliable and secure storage of data and can be used to provide SaaS in the form of data analytics and other applications. For example, an online historian can be used to provide data analytics. Once analyzed, the data becomes information that can be accessed by web-enabled HMI software, and then displayed to users.
Data analytics can reveal patterns, and cloud computing enables artificial intelligence, robots and other Big Data tools to be used. The scalable processing power on the cloud is useful for performing these tasks, which often require considerable computing resources. By using a plant floor HMI as a gateway, access to information from the cloud is available in many forms, including bar charts, graphs, trends, tables, etc. (Figure 2).
Cloud computing analytics and an HMI help to not only visualize data, but provide a more cognitive view. These advanced data analytics systems have built-in agents to analyze and find patterns in the data. When the data is no longer following those patterns, the system can automatically indicate the result or finding to a user through an HMI and other methods.
An example of this would be an HMI application that is collecting and pushing production data to the cloud. One of the pieces of data is part counts with an average of 120 parts per hour. If the count is a little slower, the analytics will detect the pattern change. The software then can issue a warning, even if this condition wasn’t previously defined as an alarm or an event.
Cognitive computing can detect anomalies and report these findings, as the example shows, even for conditions not previously defined or anticipated. This saves time and prevents users from having to anticipate every possible abnormal condition, and program appropriate alarms and events.
Cloud computing is already in widespread use for commercial applications, and is starting to be applied in industrial settings. It’s much quicker to deploy, costs less, and is easier to scale than traditional on-premises solutions. These advantages will fuel increased use, and the right HMI makes it much easier to deploy cloud-based industrial Big Data solutions.
Fabio Terezinho is a profit and loss unit leader at InduSoft, a Schneider Electric Software brand. Edited by Emily Guenther, associate content manager, Control Engineering, CFE Media, firstname.lastname@example.org.
- Cloud computing for industrial processes
- The benefits of cloud computing
- How to host an HMI to filter and collect data.
What are the best practices for cloud computing for industrial processes?
Related products for industrial cloud applications
Products related to industrial cloud applications include the following.
- Wonderware Online is an online historian that can be used to provide data analytics.
- Web-enabled HMI software, like InduSoft Web studio, can display analyzed data to users.
- Microsoft Azure is a cloud service provider.
- Low-cost microcomputer platforms, such as a Raspberry Pi computer, can host an embedded HMI.
- InduSoft IoTView, hosted on a microcomputer, can gather data and push it to the cloud.