Custom IIOT APP for production monitoring
Empowered the client to leverage IIOT for real time productivity monitoring and extract insights for improvements with customization of existing platform
Challenges & Objective
- Existing IIOT platform with standard features is insufficient to support all requirements
- Actionable insights not available for analyzing gaps; therefore can not improvise
- Performance issues with data availability over longer durations
Solution
- Customization to handle operational aspects – shift timing, Job IDs, engineering cycle times etc
- Dashboards to display all relevant KPIs
- Data snapshots for quick data analysis over months
AI ML Model Development & Management Platform
"No Code” Web based Machine Learning models development and management platform will enable domain experts to embrace AI & ML quickly, without getting into details of model development is very important to tackle all these issues
Challenges & Objective
- Domain experts require python knowledge to develop the ML models or have python developers in the team
- Maintaining the models and refining over time is difficult
- Data preparation tasks are time consuming
Solution
- Easy GUI interface for data preparation & model development
- Develop multiple models, compare different algorithms and publish the most suitable ML model
- “No Code” workflow and web deployment enables wider adoption
IIOT & ML based Weld Quality Prediction in Cold Rolling Plant
Unsafe and error-prone hammer test for weld inspection is replaced with ultrasonic & AI based technique
Challenges & Objective
- Hammer test is performed to check the quality of the weld
- Error-prone as well as involves risk of injury or human life
- Any weld failure in the process causes complete shut down of the mill resulting into heavy losses
Solution
- Hammer test is replaced with ultrasonic head scanning the complete weld
- Real time analysis of ultrasonic scan data to assess the metal thickness variation along the length of the weld
- Machine learning of the processed scan data for prediction of weld quality
GUI and Cloud Deployment for Engineering Solver OEM
Easy and wider access to the simulation solver is enabled by developing API interface to solver functionalities, Web GUI and cloud enablement
Challenges & Objective
- OEM has developed Multiphysics solver
- Almost impossible for other users to access in the absence of GUI
- Very difficult to provide access to the customers who do not have large compute facility
Solution
- Web interface is developed including 3D visualization and post processor
- Cloud deployment is handled
- Working as joint team along with customer
Edge Computing and Embedded Device – Advanced Materials industry
Portable edge computing device will simplify field inspection using multiple sensors – thermocouples, Acoustic, Vibrations, Vision
Challenges & Objective
- Deploying sensors and data acquisition in the field is difficult due to variety of sensor protocols
- Managing and transferring data for analysis is tedious and time consuming
- Real time insights are not available
Solution
- Multi-sensor adapter is developed for handling various protocols
- Edge computing is performed so that data is analysedreal time
Optimal Cycle Design for Multi-transporter Chemical Treatment OEM
Application with optimization features allows the process expert to configure multi-transporter program of chemical treatment plant for maximum loads per cycle
Challenges & Objective
- Multi-transporter system has multiple constraints due to various treatment tanks and soaking times
- Coming with optimal program using manual methods take up few days; resulting into losing the contract against competition
Solution
- Mathematical model is developed which solves design space covering all process ranges
- It allows the process expert to perform fine tuning and arrive at optimal solution in minutes against days