Read Aloud the Text Content

This audio was created by Woord's Text to Speech service by content creators from all around the world.

Text Content or SSML code:

It’s typical for Plant Managers and Shift-Supervisors across industries, to be accountable for multiple products that are created using Continuous Manufacturing Processes. For each product, they must manage numerous raw materials that pass through multiple production stages. There may be tens of steps in each of these stages that can run for many hours or even for days. There are literally hundreds of process parameters like Temperature, Pressure, Flow etc that are involved across these stages to get the final yield. Due to this complexity, they struggle with problems such as Yield fluctuation across batches, deviations in Quality KPIs leading to rejection of batches and ultimately fail to meet production targets. This leads to Plant Managers having to contend with ‘acceptable deviations’ on daily production runs. Whenever there are major yield deviations for a set of days, it is a common trend to rely on the tribal knowledge of SMEs and try to improve the yield. This is usually done as a one-off activity and fails to provide a long-term solution with real time visibility for operators. With the advent of digital transformation programs, the solution to the Yield problem is to leverage it for ‘Yield Optimization and Golden Batch Analytics’. The current solutions in the market do not have the ability to include Human Inputs with Sensor Data and combine advanced optimization algorithms with real time AI recommendations. Introducing the UnifyTwin Yield Optimizer App that solves these challenges by providing real time AI recommendations to optimize the yield. It accomplishes this using a multi-step data driven approach. Data Contextualization: UnifyTwin provides a ‘unified contextual data model’ that brings Human, Machine and Enterprise Data together in an Industrial Cloud. The Machine Data typically includes all process parameters from DCS/Historian, the Human element may include data like quality sample readings and Logbook entries and the Enterprise Data may include Cost of Raw Materials, fuels etc. for ROI calculations. Critical Feature Generation: The AI Engine utilizes the ‘contextual data model’ and intelligently generates the profile for Golden Batch using historical data. From the hundreds of process parameters, the AI model distills it down to a top-N critical parameters that directly impact the yield. It also factors in parameters that can be actually controlled by the Plant Operator to make recommendations. Outcome Validation: The list of critical parameters along with the importance score is made available to the Plant Process SMEs for review and this can be fine-tuned thru the App interface. Realtime Recommendations: Post the Validation, the App seamlessly combines the Optimization Algorithm with the Operator Recommendations and presents it real time in the App. The Operator has to just ensure that the critical parameters are within the prescribed ranges to achieve the Optimum Yield Objective. As a result of deploying this app, our customers have been able to achieve Optimum Yields consistently for each production run. Based on past implementations, customers have achieved Yield Improvement from 5 to 8%, ~10% improvement in profitability and 150 - 200 thousand dollars savings per annum on material usage (raw material, steam/fuel, solvents etc.). In addition to this, the App can be securely enabled across multiple plants with a scalable deployment model. To know about the Yield Optimizer app and UnifyTwin, contact us at