HMIs Accelerate Edge Intelligence
EP Editorial Staff | February 10, 2021
Modern edge-capable HMIs do far more than provide visualization. They are essential to connecting with data and processing it into useful information.
By Fabio Carvalho, ADISRA
Distributed intelligence at the edge is becoming the norm for today’s expanding industrial-digitization environment. Human-machine interfaces (HMIs) on embedded devices, located ideally near field equipment, play a key role in this data-driven, distributed, and collaborative intelligence model.
According to research firm IDC, Framingham, MA (idc.com), “By 2025, 25% of enterprises will have headless analytics embedded into the business, thereby improving real-time operational decision making and achieving process automation.” HMI software supports this trend by remotely connecting operators to machines, systems, and devices to bring real-time analytics intelligence closer to the field. (IDC FutureScape Worldwide Data and Analytics 2021 Predictions (Doc # US46920420) is available for purchase here.)
The latest edge-capable HMI generation supports legacy platforms—and integrates with existing devices and systems—while incorporating emerging platforms, including edge devices, 5G and 6G network technologies, remote visualization, and embedded analytics. The trend is moving away from centralized cloud-computing solutions and toward distributed intelligence, provided by edge technologies, that is delivered to operators with HMI devices.
Exponential Data Growth
Data volume, generated across increasing numbers of applications and devices, is growing exponentially. Pre-processing data locally, before sending it to the cloud, becomes an absolute necessity to enhance efficiencies and avoid network degradation.
Sending enormous amounts of data to a private or public cloud consumes significant network bandwidth. This issue is further exacerbated for industrial-automation applications because some sensors constantly transmit data. HMI software preprocessing improves the situation by presenting only data that helps operators make real-time, data-driven decisions.
Latency, a measure of network delay, is another consequence of excessive network communication. For some applications, such as temperature condition monitoring, latency to the cloud is not an issue. However, for high-speed production lines, where real-time decisions affect product quality, scrap volume, and other production losses, latency cannot be permitted. Smart, edge-located HMIs reduce latency through data-quality filtering and aggregation.
HMI software’s traditional role has been to visualize and monitor industrial systems, especially for human interaction. In today’s data-driven world, HMI software’s role is expanding to include computational and cognitive functionality. Its evolving responsibilities can be characterized by five Cs:
Ubiquitous and agnostic connectivity, provided by HMI software, is necessary for seamless integration with a wide array of field devices, legacy systems, edge devices, other HMIs, historians, and the cloud. Every device on the factory floor, throughout corporate back offices, and in the expansive supply-chain universe should be connected for an efficient industrial-automation environment.
New industrial-automation business opportunities are possible with the introduction of 5G and 6G internet innovations. These innovations promise to deliver improved response times, accompanied by frugal bandwidth utilization. Accessing real-time
remote data from every step of an operation, including from mobile vehicles, is invaluable for reducing costs and waste in many industry segments.
HMI software remains important for collecting raw field data and providing visualization. Today’s HMIs have a significantly improved ability to contextualize information by adding disparate and related data sources together to make the data meaningful. Data collected from multiple sources is used to create an HMI information superset that delivers a more-complete end-to-end view of operations.
Original HMI configurations performed basic functions such as issuing commands and setpoints to a machine controller. Evolved HMI software runs on improved edge-located hardware and can execute more computational and elaborate data processing. To enable better-informed decisions closer to the equipment, HMI software on the edge must embrace real-time analytics by incorporating contextualized data.
As technology’s upward trajectory continues, advances in HMI software cognitive capabilities will help operators quickly understand available data in near real time. The preceding abilities are effectively shifting the event horizon from simple data acquisition to actual knowledge creation, and edge-located HMIs are ideally located and suited to add intelligence to any system.
One approach uses rules systems, which are a set of deterministic if-then-else statements that can provide useful results—if the right real-time data and correct rules are available. However, it is important to understand that bad data and/or improper rules can result in false positives.
Another method for adding cognitive capabilities directly to HMI software involves incorporating machine learning (ML) algorithms. ML algorithms are trained with labeled data, so they can then be used to interpret real-time data and identify data that meet predetermined categories without knowing the exact rules.
Deploying HMI software with embedded analytics delivers intelligence to the edge and provides a distributed collaborative intelligence model among multiple edge devices. In an uncertain environment, plants require operators to function autonomously by making decisions based on available contextual data. Carefully developed smart HMI software embedded in their workflow will supplement these efforts.
Getting the Big Picture
HMI software function has historically been to deliver simple, easy-to-comprehend interfaces, allowing an operator to understand machine or process conditions at a glance. Smart HMI software is now called to do more by delivering faster time-to-insight and greater operator efficiencies, but only if the information can be effectively conveyed.
Viewing form factors have evolved from standard plant-floor industrial operator interface terminals and control room PCs into headless operations supporting remote visualization on mobile and wearable devices. Remote visualization has effectively become mandatory with the current COVID-19 pandemic’s effect on the plant’s environment.
Therefore, it becomes important for a single HMI software-development environment to display across an array of viewing-form factors: on single HMI panels, desktops, high-definition TVs, tablets, and mobile and wearable devices. Each viewing experience must result in easy interpretation by the operator.
The development environment must be easy-to-use and intuitive, using built-in HTML5, which has emerged as the preferred method to support easy deployment to mobile devices. Smart HMI software balances compatibility with existing legacy systems with delivery of new functionality to users.
Cyber hackers are finding innovative ways to steal data in this remote and connected industrial-digitization world. Smart HMIs must include built-in and robust user-authentication methods to secure the distributed collaborative intelligence model that also incorporates the cloud. Cyber security is paramount to any smart HMI on the edge as the distributed collaborative intelligence model becomes more popular.
COVID-19 has accelerated many plant-digitization projects. Smart HMI software enables digitization projects by performing data pre-processing, quality filtering, and aggregation at the edge—increasingly important as the connected devices population and associated data expands exponentially. In addition, edge-deployed HMI software generates value by optimizing information processing and, consequently, delivering actionable distributed collaborative intelligence. These steps will, in turn, lead to new value models for businesses. EP
Fabio Carvalho is Director of Application Development and a senior developer at ADISRA LLC, Austin, TX (adisra.com). Carvalho has more than 17 years of software-development experience, with more than 11 years in the automation industry working on HMI and SCADA products.