5. Manufacturing Use Cases Addressed by the Sight Machine Platform
This article contains the following sections:
- Visibility Use Case
- Traceability Use Case
- Availability Improvement Analysis Use Case
- Quality Improvement Use Case
- Performance Improvement Analysis Use Case
Data fitness is the ability of the data to support a use case. This section provides a walk-through of a number of use cases from the data fitness framework that the Sight Machine platform has been used to address.
Visibility Use Case
Before moving on to using advanced analytics to solve complex business cases, a useful first step is to start by watermarking key performance indicators (KPIs) and comparing them across facilities or assets. This helps to determine opportunities for performance improvement and to accurately identify a potential problem and value opportunity.
Is one asset performing better than another in terms of quality, performance, or availability? What is your asset’s best performance? These are the key questions that you can answer by using the Sight Machine contextual data models as they relate to KPI watermarking. By standardizing key metrics across assets and facilities, you can compare similar assets across the customer network to watermark performance and determine areas for analysis/improvement. Often, this is the most valuable step to determine the necessary analysis for improvement.
In the example below, KPIs are monitored and compared in real time from two separate plants. Sight Machine can use this information to determine if there are any large differences between similar facilities, lines, or even assets. Each row in the example below represents a separate facility, with each column indicating a separate KPI.
Contextualized Data Visualization/Analysis
After you have access to your KPIs, it is important to trend these KPIs over time. Contextualized data visualization plots and analyzes any KPI or contextual data model to better understand the signal within the underlying data.
In the example below, the Sight Machine platform is used to visualize the average laser voltage, by day, over time, and compare that average voltage across 6 machines that are producing the same product in three separate facilities.
Traceability Use Case
Often, it is difficult to determine the specific root cause for a defective product, downtime event, etc. With the Sight Machine part/lot data model, you can track the flow of material through a process and attribute process readings along with outcomes to best inform and drive root cause analysis. Traceability has applications in furthering quality analysis as well as ensuring compliance, as in the pharmaceutical industry.
In the example below, the Sight Machine platform is used to determine the root cause for a recent spike in the quantity of defective engine blocks that are being manufactured. Additionally, the process engineer responsible for determining the root cause is unsure where in the process the defect was produced. With the Sight Machine part/lot data model, all process parameters and data associated with the production of the engine block are combined together with quality data to better determine the root cause. The depiction below shows the use of an out-of-the-box data visualization in which the distribution of casting injection pressure maximums for the last 7 days are stratified by the quality result of pass or fail. The chart shows that after a particular injection pressure maximum, the percentage of quality engine blocks produced drastically decreases.
Availability Improvement Analysis Use Case
A number of factors can impact availability, ranging from a change in material used within the process to a change in a process parameter. The Sight Machine platform can be used to identify the assets with lagging availability, determine the specific problem space (when did this start?, what products are impacted?, etc.), and drill down into the related process and material data prior to and during the contributing events. Availability analysis can contribute to a reduction in unplanned downtimes or even the prediction of unplanned downtimes for preventative resolutions.
In the example below, the Sight Machine platform is used to compare the availability of three laser cutting machines in the same plant, over time. The process engineer is quickly able to identify the problem space and realizes that Lasercut 2 is behind in terms of its availability. From here, the platform can be used to identify specific parameters or changes that may have contributed to the poor availability.
Quality Improvement Use Case
With a number of different processes, machines, and factors contributing to the production of a part/lot, it is difficult to filter through all of the noise and determine which signal is contributing to resulting quality tests. With the Sight Machine platform, all of that data—from process sensor data and line quality data to lab test results—is available within a single contextual data model that represents each piece (part) or collection (lot/batch) of material.
The Sight Machine platform allows for the following use case enablement:
- RCA: Drive root cause analysis and increase quality by identifying trends and correlations between contextualized process data and corresponding quality results.
- False Positive/False Negative Reduction Analysis: Catch defective material earlier and reduce the amount of defects that reach customers.
Performance Improvement Analysis Use Case
Performance Improvement Analysis is enabled out of the box with the Sight Machine platform. Performance analysis starts with tracking the key performance indicators (KPIs) and can be done with varying degrees of fidelity within the Sight Machine platform. Performance can be viewed and compared across facilities, lines, machines, assets, and shifts.
Asset In-Cycle Process Time
Explore the details of asset cycle times to better understand variables that contribute to differing cycle times, such as the type of product produced or operator comparisons.
In the example below, the starting place for performance improvement is comparing cycle count distributions by cycle time for two separate facilities with one type of asset. From here, further comparisons can be made to refine the problem space. For example: Is one plant producing a more complex product? Is one plant providing better training?
Line Bottleneck Analysis
With the Sight Machine platform, you can better understand where the bottleneck lies within a given line by comparing asset performance within the line for a given type of product.
In the example below, a single pick and place line is broken out by machine in the line and further stratified by product type. The analytic allows process engineers to quickly determine the bottleneck within the line, and is indicated using a heatmap.