AWS Partner Network (APN) Blog
Unlocking the Power of Machine Data with SmartInsights from TensorIoT
By Nicholas Burden, Sr. Technical Evangelist – TensorIoT
By Shahan Krakirian, Sr. Software Engineer – TensorIoT
By John Traynor, VP and GM, Products and Solutions – TensorIoT
By Yang Chen, Roajer Gilbert, and Harjot Kalra, Sr. Solutions Architect – AWS
 |
| TensorIoT |
 |
From early use of the steam engine to the advent of robotics, industrial manufacturing is one of the industries that most consistently embraces process improvement and benefits from technological advances.
Now, with the application of Industrial Internet of Things (IIoT), which refers to the use of IoT technology in industrial settings, manufacturers can connect sensor outputs and device controls to cloud technologies for data capture and analysis.
Armed with new insights, manufacturers can make additional process improvements and achieve specific business goals. For example, a manufacturer may want to maintain output while reducing resources, or produce more with fixed resources. Likewise, an industrial company may want to reduce downtime by predicting equipment failure and scheduling or dispatching maintenance to avoid unplanned downtime.
Although IoT connectivity unlocks new opportunities for businesses to collect data and use it to improve operational efficiency, there are often hurdles to implementing IoT, including concerns about cybersecurity and a lack of knowledge and skill within IT organizations.
In particular, one of the challenges manufacturing companies face is figuring out how to correctly collect, use, and interpret data so that, for example, raw vibration data from a piece of equipment can be used to correctly infer when that equipment is doing value-adding work in an optimal fashion.
While further details are beyond the scope of this post, there are typically three approaches to collecting data: via a modern protocol such as Open Platform Communications Unified Architecture (OPC UA), via a direct interface to programmable logic controller (PLC) and other control equipment, or by adding secondary sensors to equipment.
Then, selecting how to use and interpret the data involves selecting among various visualization and analytics options. In doing so, the primary objective should be to provide actionable insight in a way that is aligned with business goals, is easy to interpret, and which provides value to workers and managers alike.
TensorIoT SmartInsights
TensorIoT, an AWS Advanced Tier Services Partner with the IoT Competency, recognized the need for solutions that help companies collect and interpret equipment and process data. TensorIoT built SmartInsights as an AWS-based customizable end-to-end solution that industrial and commercial customers use to rapidly connect and derive actionable insights from operational systems.
Unlike software-as-a-service (SaaS) solutions, SmartInsights is deployed in a customer’s own AWS account, so that each customer owns and controls their data and has complete flexibility to use the solution as-is, or customize it and integrate with other enterprise systems and applications.
By eliminating the complexity of stitching together multiple AWS services, and by presenting everything in a single pane of glass, SmartInsights reduces the level of technical know-how and skill needed to implement a functional solution. The solution accelerates time to value for manufacturers and other industrial customers.
The Garland Company and Project Inception
In this post, we’ll look at how TensorIoT SmartInsights helped The Garland Company rapidly realize a 15% energy savings and 17% throughput increase in one of its manufacturing processes in a few short weeks after SmartInsights was first deployed.
The Garland Company has been a leading provider of commercial roofing and building envelope solutions for over 125 years, producing materials used in commercial, industrial, and public buildings in multiple countries across the world. Given its reputation for leading performance within the industry, Garland continually pursues innovation to develop the highest quality products and improve internal processes, which is why it chose to pursue an IIoT solution with AWS and TensorIoT.
One product line Garland manufactures is bituminous roofing membranes, and one of those product manufacturing lines was identified as a candidate for process improvement based on suspected process variation.
To improve Garland’s operations cost and manufacturing efficiency, the goal for TensorIoT was to make the wealth of telemetry data coming from the roofing membrane manufacturing line more accessible and actionable through SmartInsights.
In order to better understand the process and facilitate root cause analysis, telemetry readings from across the line needed to be extracted from the underlying hardware and organized in a central dashboard, making SmartInsights the perfect solution for visualization and analysis.
Deployment and Solution Details
Let’s take a deeper look at how TensorIoT uses AWS services to process data to extract and visualize meaningful insights by walking through the major steps in the Garland deployment.
During discovery workshops, Garland shared that it had previously obtained an eATM tManager ControlLogix (Softing) module to enable data collection and aggregation on site. The tManager module is installed directly into the chassis of a PLC, upon which the in-chassis module automatically enumerates PLC and database tags (a unique stream of data) that can be sent to a variety of destinations, including MySQL, Microsoft SQL, and Oracle.
More relevant to using TensorIoT SmartInsights, tManager also has the capability to natively connect and send machine tag data to cloud platforms such as AWS IoT Core via the MQTT protocol with minimal configuration, from where SmartInsights can propagate the data to other services.
Though Garland decided to bring data into SmartInsights via MQTT, the platform can be agnostic as to which industrial protocol is used for data transfer. This allows companies to use other common protocols such as OPC UA, Ethernet/IP, and Modbus to benefit from SmartInsights.
With a data ingestion plan, the next step in the solution was to deploy SmartInsights into Garland’s AWS account. This is done using SmartInsights Deployment Wizard, a custom-built and React-based application customers use to choose which AWS account and region to deploy the solution into, as well as other important deployment parameters.
The deployment parameters are then passed to AWS CloudFormation and subsequently executed, generating all of the required AWS services that constitute the solution.
Figure 1 – SmartInsights Deployment Wizard application.
SmartInsights data propagation starts with a single AWS service: AWS IoT SiteWise. For Garland specifically, incoming MQTT tag data lands in AWS IoT Core and is forwarded into AWS IoT SiteWise via an AWS IoT rule.
However data lands in SiteWise, it is batched and stored in Amazon Simple Storage Service (Amazon S3) and Amazon OpenSearch Service via an extract, transform, load (ETL) process facilitated by Amazon Kinesis.
Figure 2 – SmartInsights architecture diagram.
Once data propagates through SmartInsights, multiple types of visualizations become available to end users such as operations engineers and floor managers. Custom React-built graphs give end users layer data from multiple sources over one another, providing a side-by-side view that otherwise isn’t possible on many manufacturing floors.
Figure 3 – SmartInsights custom React graphing capabilities.
Separately, end users can build OpenSearch dashboards which have the capability of combining data from multiple assets and displaying them with multiple types of graphs. These are embedded into the SmartInsights platform on the “Extras” tab, which provides data in a single pane of glass.
Centralized data visualization proved to be a valuable tool for Garland to identify anomalous data from the mixer and the winder.
Figure 4 – SmartInsights OpenSearch embedded data visualization.
Using SmartInsights, Garland was able to ingest and observe important data points for the winder and mixer such as the following:
- Mixer: Current recipe, temperature, amount filled, and other related values.
- Winder: Speed, roll tension, roll length, roll acceleration, roll deceleration, mandrel feed speed, mandrel eject position, mandrel orientation, and more.
Customer Impact and Initial Results
By deploying SmartInsights, Garland’s operations staff gained immediate visibility into any unexpected variations that were occurring during the production process. With all key process data visible in a single pane of glass, production team members could identify unexpected anomalies and take immediate action to adjust equipment and optimize their production process.
The quality assurance lead at Garland’s Cleveland, Ohio plant works with counterparts at other manufacturing sites. With a multitude of manufacturing sites and processes, a successful SmartInsights launch offered Garland the opportunity to bring similar improvements across its locations.
After adding SmartInsights to the production line, Garland uses SmartInsights as a way to visualize baseline data to check for changes to various tags (mixer speed, for example) and identify causes of downtime or improve uptime. SmartInsights also provides an intuitive plant overview to allow leadership comprehensive visibility of their entire manufacturing processes.
By examining plant processes using SmartInsights, a Garland engineer was able to identify an anomaly in one mixing tank against baseline performance. While the anomaly did not impact product quality directly, it did indicate an opportunity for improvement in cycle efficiency.
The maintenance team tuned the malfunctioning mixer which led to an immediate reduction of about 17% in processing time per batch and a 15% energy savings. Now, Garland can mix more bituminous material in a day with less energy use and less waste, increasing throughput while also reducing the carbon footprint of its production operations.
Conclusion
TensorIoT SmartInsights, which is available on AWS Marketplace, helps manufacturing companies harness the power of their data.
If you’re looking for an innovative solution that will help your manufacturing company gain the maximum benefit from your production data and improve overall operational efficiency, contact TensorIoT to schedule a demo.
With SmartInsights and other TensorIoT solutions, you can increase situational awareness, gain actionable insight, and make more confident operational decisions in your production environment.
.
.
TensorIoT – AWS Partner Spotlight
TensorIoT is an AWS Partner specializing in cloud-native solutions for Internet of Things (IoT), artificial intelligence, analytics, serverless apps, and sustainability.