Turning Manufacturing Data Into Business Value: Sight Machine at Hannover Messe

Turning Manufacturing Data Into Business Value
Facebook
Twitter
LinkedIn

Table of Contents

Sight Machine to Showcase its Manufacturing Data Platform and Operational Digital Twins with Microsoft

Despite many millions of dollars in investment and years of work on their plant data, countless manufacturers are still struggling to understand their production systems. Why?

It isn’t just the complexity of the data, the frequent lack of data labeling, and the incompatible formats. And it certainly isn’t a lack of data.

It’s the lack of context. Manufacturers are often overloaded with data but have little context to show what the data means, how the machines and processes in a connected system interact, how product moves through a system, and how to improve production metrics. In an IDC study, only 29% of Global 2,000 companies reported having no significant barriers to extracting value from their operational data.

What if you could have a system that showed you how machines are performing and products being created at every step of the process, from raw material entering the first machine to the finished product being loaded onto a truck? A system that could identify optimal settings for all the machines on a line, advising on how to adjust settings during line changeovers and shifts in raw materials? A system that could help optimize across the enterprise for any of the key production objectives, including increased output and energy efficiency, and reduced defects, waste, and cost?

Turning Manufacturing Data Into Business Value

Increasing Yield and Quality, Reducing Cost and Scrap

Sight Machine’s Manufacturing Data Platform can do that and much more. Our customers have achieved outcomes including:

  • World’s largest magnetic wire provider for EV market reduces scrap by 45% and improves OEE by almost 15%
  • Large paint and coatings company cuts cycle time by 7% to yield a $7 million improvement in throughput
  • Global brewer reduces downtime and increases throughput by 16%
  • Large paper and packaging company combines energy and production data to reduce steam usage by 2%, cutting energy costs and lowering CO2 emissions

Data lakes, machine digital twins, and general-purpose IoT platforms can be useful tools, but they can’t address the core issue of manufacturing: the complex relationships between machines, work, and product. Understanding how materials are flowing through the machines is the key to understanding how to optimize production.

When a paper mill shifts the proportion of recycled paper in its raw materials mix, how does that affect the number of downtime-causing sheet breaks after the material has passed through more than a dozen other machines? How should operators adjust settings on the machines to minimize those sheet breaks?

If you don’t deeply understand how variation at each step of production influences the nature of the product, it becomes extremely difficult to answer basic questions: why is the scrap on line 7 higher than line 5? Why has our first pass yield decreased? What is the correlation between different grades of lime and silica and the quality of cement?

So how do you fix this problem? Digital twins of each machine? A digital twin of a machine may help identify when a machine needs realignment or a part needs replacing. But it won’t answer questions like how speed and other settings of each machine along the line drive the key metrics of production. And each digital twin becomes a data silo.

A deep type of contextualization is required. The first step is to identify what the data represents – which sensor or system it comes from on which machine and what it measures. Too much data over the past few years has been collected in data lakes in a way that lost crucial context. But that’s just the start. Where in the process of creating this product does that sensor interact with the material, what impact do changes in operation at that point have on the finished product, and why?

A crucial aspect of the problem is the ability to track material flow through the process, accounting for time offsets and out-of-order data, including certain types of data (e.g., quality data) that might arrive far after production is complete.

Building the Bedrock for Intelligent Manufacturing with Microsoft

Visit us in the Microsoft booth (Digital Ecosystems, Hall 17 | Stand G06) at Hannover Messe, April 17 – 21

Sight Machine’s Manufacturing Data Platform (“MDP”) captures the entire manufacturing process, incorporating all the machines on a line, all the lines in a plant and all the plants operated by a company. The MDP creates a common data foundation that contextualizes all plant data through Sight Machine’s standardized data schema, and an operational digital twin to provide enterprise visibility of the entire process. Sight Machine is optimized on Microsoft Azure using machine learning to identify and label data streams, and uses artificial intelligence to identify and continuously refine optimal settings. Together, we are helping manufacturers transform production with their plant data.

Register Now

Request a specific time to meet with us at Messe to see our demos, hear best practices and have your questions answered.

Picture of Sight Machine

Sight Machine

Making Plant Data Continuously Useful for Operations, IT, and Data Science

Curious about how we can help? 
Schedule a chat about your data and transformation needs.