Process AI
Easily optimize cost and line speed.
Process AI automatically predicts the offline quality of product while it’s still being produced. These quality predictions are streamed to operators and displayed together with recommendations on the optimal process settings for a given line and product.
This greatly reduces the risk of adjusting process settings while empowering operators to optimize their cost and line speed. The financial impact of changes are also predicted instantly to document the impact on material usage and energy costs.
Process AI is built to leverage the powerful Oden data architecture, this innovative solution represents years of experience applying artificial intelligence and machine learning in a manufacturing setting.
Manufacturers with operator-dependent processes, especially those with a high degree of process variability, stand to realize the greatest value by using Process AI.
Process AI de-risks cost and line speed optimization with turn-key predictive quality models to keep the end product in spec.
An early customer of Process AI saw a 5% improvement in related costs, while increasing line speed by 3% within 6 months.
Oden’s core data architecture contextualizes, cleans, and enriches manufacturing data in real time. This is what makes Process AI’s predictive quality, cost, and line speed optimization possible. The entire process is documented to measure financial impact automatically.
Process AI contextualizes both operational and enterprise data, analyzing millions of data points across systems. Historical production data, predictive quality results, and company set parameters are all used to identify which process settings will optimize material and energy costs for the current state of a given line and product. The predicted financial impact of these process settings are then instantly delivered to the user.
This workflow would otherwise take several systems, weeks of time, and collaboration across multiple teams. Using Process AI, prescriptive process settings can be delivered to operators within seconds.
First, the operator selects the line and product they want to improve or check the quality of.
Typically, this will be the product/line combination currently running.
Process AI will pull information about the current cost and speed of the line selected.
The model analyzes stable periods of production in historical data from the last 6 months.
Process AI also incorporates predicted quality results and company set parameters that align with current state. Using this information, optimal process settings are identified.
Next, Process AI delivers a prescriptive recipe for the operator’s process settings.
Process AI automatically streams several key data points:
Join Our Manufacturing AI Interest List
The predictive quality in Process AI looks at the current production conditions and predicts what the offline quality test would be. These predictions are enabled through automated analysis of historical data.
For more information about how predictive quality works, click here.
Predictive quality is used in Process AI to ensure that process setting recommendations keep product quality in spec.
The Artificial Intelligence (AI) in Process AI is intended to give intelligent recommendations for operators to reduce material costs, reduce energy costs, and to optimize line speed.
The modeling and recommendations in Process AI are an example of Machine Learning (ML), which is a category of artificial intelligence. A Machine Learning algorithm identifies patterns in a given set of input data to determine its correlation with output data. The ML modeling in Process AI is used to predict the quality of products and to identify historical segments of optimal process settings.
Process AI gives numerical recommendations, such as lowering the setting for material flow or increasing the temperature at a certain step of a manufacturing process.
Delivering value with Process AI involves teaching the software about a given manufacturing process. Process AI automatically does the heavy lifting to identify stable segments where there are optimal cost savings while maintaining quality spec. However, it’s still helpful to build trust in recommendations by exposing more controls to an experienced process engineer.
This iterative process of feedback, improvement, and validation is supported by a module in Process AI specifically designed for process engineers. Using this module, the process engineer has the option to experiment and adjust recommendations before pushing the recommended changes to operators.
Process AI provides insights and recommendations on segments that are stable for 15 minutes or more. So if a very long run had a mix of high and low performance, the user can understand the actual optimal performance within the run, not the average. This allows manufacturing teams to make the most of their process improvements, faster.
Process AI is built for manufacturers with continuous or batch-continuous processes. Operator-dependent, high variability processes stand to realize the greatest value by using Process AI.
Manufacturing industries that are a perfect fit for Process AI include: paper & pulp, wire & cable, packaging, plastic extrusion, building materials, non-wovens, compounding, laminates, ink, coatings, pipe, chemicals, and additives.
The accuracy and actionability of our recommended process settings are our top priority. The first phases of deploying Process AI center on the training, testing, and validation of recommended settings in various situations.
While Process AI does the heavy lifting to analyze and improve its own performance, there are critical tools for our customers to provide feedback and correction. Process Engineers have a module designed specifically for them to experiment and fine tune the recommended settings.
Manufacturers can also implement parameters to restrict the recommendations available in the Operator View. These can be based on cost savings, quality targets (found throughout the Oden solution), and/or speed limits. If you’re a current Process AI customer and you’d like to implement these, please contact your Customer Success Manager.
Process AI is focused on material usage and energy costs. By predicting the quality metrics during a manufacturing process, the operator can see when they are over processing or adding more material than they need to.
