1 / 2

Analytics in Manufacturing | Tibil Solutions

Tibil's analytics in manufacturing uses all available data within your company that helps you to increase production, reduce cost and optimize supply chains. Our tools in analytics for manufacturing allow you to identify machine breakdowns, and predict the issues and performance of the machines. Tibil's analytics in manufacturing gives you real-time insights that allow you to take action and make business decisions wisely.<br>Website: https://tibilsolutions.com/industry-solutions/manufacturing-data-analytics/

tibil
Download Presentation

Analytics in Manufacturing | Tibil Solutions

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. ARCHITECTING INFORMED DECISIONS Impactful Partnerships Manufacturing Sector India Background Tibil has assisted an Indian conglomerate, a leading metals flagship company, in overcoming the difficulties they encountered in forecasting the trend of Copper Matte from the raw copper ore The client utilizes a smelter to treat raw copper ore in its plant This results in Copper Matte (Cu₂S) and several byproducts including SO₂ and O₂ The difficulty lies in not being able to predict if the Copper Matte value would drop below an acceptable threshold, therefore unable to act in advance Client Requirements Forecast the production trend of Copper Matte from the raw copper ore They carry out a chemistry-based experiment on the specimen to gauge the output's quality. They are unaware in advance if the Copper Matte value is dropping below a safe threshold, hence unable to act in advance ML model to predict the trend of Copper Matte was the devised solution Business Challenges So₂ and Cu₂S are the byproducts of the same reaction. The extent in which SO2 is produced is highly correlated to the Copper Matte production SO₂ analyzer trials were conducted in the FSF settler and zone 2 as shown in the figure The trials were unsuccessful because of the extreme temperatures and dust load in zone 1 as shown in the figure After placing the sensor at zone 2 though the temperature was low but readings were highly noisy since the dust was mixed with the off gas of the reaction

  2. Probe installed at WHB-31 convection inlet Probe completely jammed and melted 1 2 Data Preparation The independent variables include IOT sensor values and Dryer outlet Composition The dependent variable is Matte grade readings which comes at random intervals The data preparation includes: Step 1 – Integrating the data coming with IOT sensor and Dryer outlet Composition, and the matte grade readings Step 2 – Normalizing the temporal part of the data Step 3 – Perform Data Cleaning and Consistency checks Key Challenges and Solutions Data for the most significant characteristic (SO₂) are quite noisy Solution: is instead of relying on SO₂ which is a byproduct we can look for the reaction environment to the make the predictions Not enough data points Solution: is Linear Imputation The majority of trained models cannot capture the breadth of variance in y. Generalization (high bias) is to blame for this, although performance declines noticeably as variation is raised. Solution: is to Increase the amount of data on the model with high variance and experiment with modeling after Feature Engineering TALK TO enquiries@tibilsolutions.com www.tibilsolutions.com +91 83748 44335 US

More Related