Coca-Cola Bottler Digitizes Manufacturing Processes with AWS

Coca-Cola İçecek (CCI) is the world’s sixth largest bottler of Coke products. Headquartered in Istanbul, Turkey, CCI brings some impressive stats: They manufacture 1.2 billion cases of drinks produced annually across 26 bottling plants with operations in 10 countries, including Turkey, Azerbaijan, Iraq, Jordan, Kazakhstan, Kyrgyzstan, Pakistan, Syria, Tajikistan, and Turkmenistan, and serve 400 million consumers.

CCI is modernizing its manufacturing facility by creating a digital plant replica—a digital twin—in the cloud. It hopes to unlock value with advanced analytics, artificial intelligence (AI), and real-time asset monitoring. In fact, CCI has produced a repeatable playbook so other Coca-Cola bottlers can deploy the same digital manufacturing solution in their facilities.

Digital Twin Explained

Connected, digital, and automated supply chains and manufacturing process are in huge demand as solutions across CPG in 2021. In fact, more than 80% of our global Consumer Products Food & Beverage customers are asking for some element of supply chain and manufacturing solutions from AWS.

Let me first share some key manufacturing terms:

Digital twin—A computerized replica of physical manufacturing assets that uses sensors to collect data from equipment. The cloud-based digital twin displays the real-time status of the paired physical asset. The solution allows mechanics working on the physical machines to understand the status of the machinery (for example, to detect damaged parts or performance deficiencies).
Connected factory—A manufacturing facility equipped with technologies to seamlessly share information between people, machines, and sensors. It can be within a single factory or across many facilities. Insights from the connected factory help workers make decisions to improve operations, quality, performance, and profitability.
AI and machine learning (ML) in manufacturing—Computer algorithms that simulate human intelligence to actively perceive the world, analyze and understand data, and make informed decisions. Examples of AI in manufacturing are dynamic and optimized production scheduling and predictive analytics to detect machine breakdowns before they occur.
Industrial IoT—Here, sensors on machines are connected to the internet in a manufacturing facility. These internet-based sensors are referred to as the Internet of Things (IoT). The sensors send binary data to an Industrial Internet of Things (IIoT) analytics platform that uses AI and ML to generate decision insights.
Industry 4.0—Also referred to as the “fourth industrial revolution,” the automation of traditional manufacturing and industrial practices with modern smart technologies. Companies are enhancing manufacturing equipment with machine-to-machine (M2M) communications systems and IoT sensors to automate processes and implement self-monitoring solutions so the smart machines can analyze and diagnose issues without human intervention.
Computer vision—An element of AI and ML that trains computers to interpret and understand digital images from cameras and videos. With AI and ML algorithms, machines can accurately identify and classify objects to provide insights and actions based on the images.
Supply chain control tower—A connected, personalized, and near-real-time dashboard of operational metrics and events across the supply chain. Often linked to supply chain visibility, it enables manufacturers to understand, prioritize, and resolve operational issues in real time.

Executing a Digital Vision

CCI has a vision to capitalize on the digital age by investing in data and technologies across four key pillars:

Customer experience
Asset optimization
People experience
Innovation for growth

I’ll focus on asset optimization. The CCI team worked with AWS to create a continuous improvement engine for the company’s manufacturing assets to reduce equipment downtime with predictive analytics, increase production line utilization, and prevent wasted resources, like water and electricity. With the digital twin solution, we automated the equipment and manufacturing processes in one of its plants. The solution will eventually be deployed across CCI’s 26 bottling facilities.

CIP Solution Overview

CCI has a new clean-in-place (CIP) solution that cleans the interior surfaces of production lines and equipment without disassembly. Because the CIP solution would streamline the daily cleaning process, CCI can save energy and water, better adhere to industrial hygiene regulations, and improve product quality.

Developing the CIP Solution

To create the CIP solution, AWS and CCI had to collect and process an enormous amount of industrial data and create dynamic data models—not an easy task with legacy monolithic technology and data warehouse solutions. The team deployed AWS IoT solutions, including AWS IoT SiteWise and AWS IoT Greengrass, to ingest data from equipment and processes and provide insights on intuitive, visual dashboards. The team also deployed Amazon DynamoDB, a fast and flexible NoSQL database service, to provide near real-time visibility into the CIP process with digital representations of machinery. With AWS IoT Analytics, AWS IoT Core, AWS Lambda, and AWS Glue, the team created processes to enrich the CIP data in the cloud. CCI stores the enriched data in an industrial data lake, and business users can create reports with Amazon Athena and Tableau.

Results of CIP Solution

CCI and AWS built the CIP digital twin in two months. Within four months of deployment, CCI realized these results:

Identified 20+ points of optimization in its equipment cleaning processes.
Reduced electricity usage by 20%. At an estimated annual savings of 5,200 kWatt for two systems, CCI could save 156,000 kWatt for 60 systems.
Cut water usage by 9%. At an estimated annual savings of 3,000 tons for two systems, CCI could save 90,000 tons for 60 systems.
Added 34 days to its production schedule, thanks to streamlined cleaning process.

After these impressive results, CCI deployed the digital twin CIP solution in four plants and expanded the solution to include filler-mixer machines in three plants to identify machine failures and improve production line performance. Meanwhile, CCI is working on solutions to automate all manufacturing processes to optimize assets, human-machine interactions, line utilization, and sustainability by building a complete digital twin solution in the cloud on AWS.

Bringing It to Life

There’s so much more to the Coca-Cola story than I could share in this blog post. To learn more about AWS solutions at Coca-Cola İçecek, read the AWS case study. You can also listen to this on-demand webinar to hear my interview with the CIO of CCI.