Objective: Creation of the digital twin of a water treatment plant to optimize the dosing of coagulants and reactants

Introduction

The digital twin is a new concept that consists of creating a virtual representation of the plant in order to understand, simulate and make better decisions thanks to artificial intelligence models that learn from the large amount of data that the plant generates.

In the water treatment process, coagulants and reactants are used to remove contaminants and impurities present in the water. The dosing of these chemical products is critical, since an insufficient amount would not be able to remove impurities, while an excessive amount could have negative effects on the quality of the treated water and increase costs.

In this project, the digital twin of a water treatment plant was created in collaboration with a water management company.

Project objective

To intelligently and automatically adjust the amount of coagulant and chemical reactants to eliminate impurities and contaminants from the river.

Expected benefits

• Detect anomalies in the process, such as leaks, loss of pressure at different points in the process, or other events that require studying the root cause
• Have a tool to simulate different scenarios at no cost thanks to artificial intelligence
• Provide recommendations for dosing amounts of coagulants and reactants in real time, so that operators performing these operations have AI support to make certain decisions
• Reduce process costs by 10%, mainly corresponding to pumping and other rotating machines
• Be able to anticipate plant failures and implement a predictive maintenance plan

Explanation of the project and its different phases:

The different phases of creating a digital twin consist of measuring, processing with AI, validating predictions, generating recommendations, and acting on the process. Once implemented, the digital twin learns more and more from the process, in a closed and automated cycle:

The methodology we followed at Intemic for creating the digital twin consisted of 5 main steps: