Client: Leading Manufacturer of Industrial Pipes
Industry: Manufacturing
Segment: Asset Maintenance
Solution: Predictive Analytics
Technology: .NET, SQL Server, SAP Integration
About the Client
The client is a large Indian Conglomerate with interests across a wide range of businesses – Textiles, Pipes and Plates, Infrastructure and Steel.
They are a fully integrated player within the Line Pipes and Home Textiles sectors, with a significant presence in Infrastructure development, spanning Roads, Water and Oil & Gas. As a globally recognised industry leader, the group has a worldwide reach panning more than 50 countries, and employing over 26,000 people across geographies, ethnicities and cultures.
Business Challenge
The client was looking for a solution for accurately and proactively pin pointing manufacturing defects during the pipe manufacturing process.
There are 12 major defects which gets detected at the end of the manufacturing line. The client was looking for a solution to predict the defects on a Pipe right after it is out of the Welding process (ID/OD machines) and before the start of IUT step. The solution draws inputs from the Welding process (e.g., variations in heat rate, voltage, current etc.) to predict the defects that would occur on the Pipe.
Solution
AiMetric implemented its innovative LARA predictive analytics platform for the client. The platform enabled the client with advanced inputs and intelligence.
In the first phase of this requirement LARA collects the process parameter from all machines for each pipe ID. This is done by collecting data from defect monitoring machines for each pipe ID.
The next step was to co-relate the sample defect data with backward validation with process parameter captured from machines and prepare the defect mapping with potential causes.
Reports were generated that will help to pre-determine the failure before it occurs, which directly contributes to improved plant efficiency and reduction in production costs. The following reports were generated:
- Defect Mapping report: This report gives detailed defect mapping with all possible causes based on the number of parts processed and defects reported
- Defect Analysis report: This is to understand the analysis done by application to finalize the mapping
- Defect Pareto report: It will have a detailed analysis of the top identified defects with causes that highly impact the production
- Unidentified defect report: This gives the list of defects and analysis of which of those do not follow the rules and pass/ fail parameters set as per the application algorithm
- System off-line reports: This gives detail information related to downtime of process machines, SAP connection, Defect monitoring machines
Welding Fault Detection – Defect Detection
Welding Fault Detection – Pass
Machine Learning Model Preparation
Technology
Database: MS SQL server 2005
MES application: Microsoft.NET
PLC: LSIS make
ERP system: SAP
Benefits
- 20% Reduced downtime and increased OEE
- 30% cost savings through better Solar PV power predictions
- 15% Reduced labor cost from physical monitoring of the rooftop solar PV plant
- Better cash management through the intelligent stocking of right maintenance parts at the right time
