Turning powder allocation into a data driven production decision

The challenge

A consumer goods manufacturer needed to allocate powder across multiple production lines as their needs changed throughout each run.

Operators played a central role in this process. They calculated and estimated powder requirements, verified production progress, monitored bin levels, and coordinated the movement of buggies between production stages. Powder density variability added another layer of complexity by affecting visual assessment of powder needs in each bin, while insufficient buggy availability or blockages could lead to production stoppages.

The information required to make these decisions was distributed across multiple systems. Operators relied on Excel to compile data, calculate how many buggies were needed, and track production progress. The buggy drivers carried a little piece of paper they were given that told them where to go next. As a result, many operational decisions depended heavily on individual experience and continuous manual monitoring.

The challenge was to transform this fragmented process into a connected decision system capable of combining production plans with actual shop floor conditions and anticipating powder requirements as production evolved.

The solution

LTPlabs developed an analytical application combining predictive modeling, digital twin simulation, real-time monitoring, and production visualization.

Historical product, production, and packaging attributes were used to estimate the relationship between ordered powder and final packaged powder weight. As new production information became available, actual performance could continuously update the forecast and support decisions on additional buggy requirements.

This predictive layer fed a digital twin of the production environment. The model represented the production process through operational inputs including powder container levels, buggy capacity, filling capacity, production orders, density samples, production stoppages, scrap, and overweight packages.

By simulating the interaction between these variables, the digital twin provided visibility into expected order completion, powder and bin levels, and the number of buggies required. Operators could also adjust model parameters to represent changes that were absent from historical data and assess their potential impact before making operational decisions.

The application connected this analytical capability with real-time production monitoring. A dashboard centralized indicators such as scrap and overweight packages, buggy distribution performance, bin powder levels, packaging line status, and order compliance. Alerts highlighted relevant production deviations and potential stoppage risks.

The resulting workflow automated part of the data compilation previously performed in Excel and made analytical recommendations and production status information directly accessible to operators.

Results

The project established a more data-driven operating model for powder distribution.

Operators gained a centralized view of production status and powder requirements, supported by automatically integrated data rather than manual compilation across several information sources. Analytical recommendations provided additional support for decisions on buggy requirements both before and during production.

The digital twin also introduced simulation into day-to-day decision-making. Teams could evaluate expected order completion and replenishment dynamics, identify potential powder shortages or line stoppages, and assess changes in production conditions through model parameters.

At process level, activities previously dependent on manual calculations and repeated communication became automated or directly accessible through the tracking application. The manufacturer also retained ownership of the developed application and analytical tools within its infrastructure.

The project was designed to contribute to lower scrap, fewer production stoppages, reduced data compilation effort, and greater production efficiency.