Ingest and analyze equipment data in the cloud

To remain competitive, an industrial company in the process manufacturing space needs real-time visibility on how their plants are operating and performing. This blog describes how an industrial company created a dashboard solution to optimize their manufacturing process by ingesting and analyzing equipment data from a distillery plant to AWS securely.

The need for data-driven visibility

A sugar manufacturer with multiple plants across India use molasses as the key raw material to produce Extra Neutral Alcohol (ENA) through a 4-step process; 1/ Fermentation, 2/ Distillation, 3/ Evaporation, and 4/ Purification. This company needed better visibility into their production data to make better decisions, and ultimately improve overall equipment effectiveness (OEE).

Some of the challenges this company struggled with included:

1. No historical data: The customer is using a SCADA system locally deployed in each plant to visualize the data from PLC or DCS. The SCADA system has a constraint of retaining data beyond 60 days – which disables the customers from referring to data beyond 60 days for any analysis regarding plant performance.
2. No centralized view: There is no mechanism to present the data from SCADA system in the OT network to the leadership team in a clear, easy to follow format. Hence the customer did not have a centralized view of their production process and manufacturing efficiencies across all plants.
3. No analytics and reporting: The plant operators from each plant created reports manually, on a monthly basis, to update management on plant performance in terms of downtime, consumption, output, etc. The management has to trust and depend on plant teams for any report customizations. The major concern is that there is no way to analyze the past data to derive insights, such as: replicating best practices and learnings from one plant to another, optimizing the consumption of steam, molasses, and energy, and no clear cause of downtime issues.
4.  No notification: The customer didn’t have a near real-time notification process to reach the plant team or management in case of downtime in plants, or stock updates of raw materials like Molasses, Coal, chemicals used, etc.

Due to these challenges, the customer wanted a Smart Manufacturing solution to ingest data from the various PLC & DCS systems and build a dashboard that provides near real-time visibility on how the plant is operating and performing to help optimize the process and cost based on that data.

AWS worked closely with the customer to build a solution that supported their Smart Manufacturing vision by providing: 1/ a mechanism to ingest data from PLC and DCS systems, 2/ support to securely ingest the data into the AWS Cloud, 3/ ability to analyze the OT data, and 4/ a dashboard for centralized real-time visibility into their production operations to aid in decision making.

The AWS Cloud solution

Building a scalable smart factory solution for 8 plants can be difficult in an on-premises setup. A data lake in the cloud handles terabytes of machine data. Adding future new use cases around the data lake is instantly scalable in the cloud, without any on-premises infrastructure investment. Using the cloud enabled the customer to build a scalable, secure solution with minimal risk. The plant network is securely extended to the AWS Cloud via redundant Direct Connect connections which ensures consistent bandwidth and reliability.

The Smart Manufacturing solution has four parts:

1.  Mechanism to pull the data from PLC and DCS systems

Extracting data from PLCs and DCS systems in a cost-effective way was a critical blocker for the customer. The native way to pull data from PLC and DCS systems requires additional license costs, which are not affordable. Multiple options were evaluated to extract data from PLC and DCS systems, including:

• AWS IoT Greengrass – This required the team to build drivers and the required customization.
• External Module – 3rd party module deployed in the OT network with additional security and management overhead.
• Litmus Edge – A partner-developed framework that contains 275+ pre-built industrial drivers to connect and collect data from industrial assets securely through the firewalls by deploying the edge gateway in the IT network. The Litmus Edge is an OS running directly on a physical server/virtual machine/docker and contains pre-built industrial drivers to connect and collect data from PLC and DCS systems via the local network, which speeds up the deployment time. The team decided to pull the data from the PLC and DCS systems using Litmus Edge.

2.  Securely ingest the data into the AWS Cloud

Data security, data protection, and a secure channel for data transfer were the top priorities for the customer. The customer used AWS Direct Connect with redundancy and auto failover to ensure that the data ingestion occurred on a real-time basis without any interruption and a consistent network performance.

AWS PrivateLink provided private connectivity between virtual private clouds (VPCs), supported AWS services, and the customer’s on-premises networks without exposing their traffic to the public internet. AWS PrivateLink helps connect AWS services like AWS IoT Core from customer VPC and on-premises hardware, and transfers critical data in a private, secure and scalable manner through AWS Direct Connect. The Litmus Edge continuously collects and ingests the data from PLC and DCS systems in the plant to the AWS IoT Core via AWS Direct Connect and AWS Private Link.

3.  Analyze the PLC and DCS data

AWS IoT Core can connect billions of IoT devices and route trillions of messages to AWS services without the customer needing to manage infrastructure. Data from the plant machines was securely injected into AWS IoT Core, and with defined rules, the data moves to AWS IoT SiteWise and storage layers. AWS IoT SiteWise helps to contextualize the data to harness it effectively and to efficiently take the business actions. The customer also used the Monitor options on the server to check day-to-day transactions without any custom development. The Industrial Data Lake stores the raw data from plant equipment (PLC & DCS) in Amazon Simple Storage Service (Amazon S3), and the data analysis such as ETL and KPI Calculations were carried out using AWS Glue.

4.  Build a Dashboard to optimize the manufacturing process

With the machine data ingested into AWS on a real-time basis, we built the dashboard for use cases like: 1/ Monitoring KPIs, 2/ Steam consumption and correlation with alcohol production, 3/ Distillation Plant Performance, and 4/ Statistical Process Control (SPC) for effective control of the column.  Grafana was used to create dashboards, which is an open-source analytics and monitoring solution as the use case required near real-time monitoring and Statistical Process Control (SPC). The dashboards were created using the processed data, but also included KPI calculations in the PostgreSQL database and AWS IoT SiteWise data.

Use Cases

Use Case 1: Monitoring Key Performance Indicators (KPIs)

Real-time visibility of critical Production KPIs such as Total Molasses, Total Alcohol production and Spent wash on an hourly basis.

Use case 2: Steam consumption and visual correlation with alcohol production 

Total steam consumption as well as the visual correlation with alcohol production in near real time including the ratio between them.

Use case 3: Distillation plant performance

Dashboard showing the overall performance of the distillation plant by capturing the various column pressure, temperature, and outputs such as Extra Neutral Alcohol (ENA) and Fermented Wash Flow.

 Use Case – 4: Statistical process control for effective control of the column 

Statistical process control (SPC) monitors manufacturing processes with technology that measures and controls quality. SPC triggers various machines and instruments to provide quality data from product measurements and process readings. Once collected, the data is evaluated and monitored to control that process. This dashboard shows the SPC to measure and control the manufacturing process such as Flow Meter, as well as Column metrics like Pressure and Temperature.

Results

This manufacturing dashboard solution is solving the following customer challenges: 1/ building the historical data for plant performance analysis, 2/ centralized visibility of production process KPIs and manufacturing efficiencies, 3/ receiving near real-time analytics and reporting, 4/ Opportunity to optimize the cost on Steam, Molasses and Steam consumptions based on the data, 5/ near real-time notification on critical incidents, and 6/ flexibility to add new use cases.

Summary

This sugar manufacturer was able to optimize their machine processes and improve productivity by using AWS cloud technologies and services. They could ingest, analyze, store, and manage their data to receive data-driven insights on their performance dashboards. It’s one of many use cases that demonstrate the importance and value of using an industrial data lake in the cloud.