Decentralized Disinfectant Dosing Control for Water Distribution Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Maintaining optimal disinfectant residual levels in large Water Distribution Networks (WDN) is challenging due to spatial and temporal variations, requiring effective management of disinfectant dosing to prevent microbial re-growth and minimize harmful Disinfection By-Products (DBP) formation, while conventional systems struggle with optimal control of dynamic parameters and multiple species interactions.
Innovation Solution
A system and method for local decision-making in disinfectant dosing using data-based partitioning of WDN into sub-systems, employing Principal Component Analysis (PCA), Partial Correlation Analysis, and Effective Relative Gain Array (ERGA) to map decision variables for optimized disinfectant dosage, ensuring adequate chlorine residuals and minimizing DBP formation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional systems engineering tools (modelling, optimization, and control) are used to manage water quality in large WDN, then water quality control can be achieved, but the system complexity and difficulty of detecting and measuring dynamic parameters increase significantly
Solution Approach 1:
The patent applies segmentation by dividing the large Water Distribution Network into multiple zones based on spatial variability of water quality parameters. Each zone is independently modeled and controlled, reducing the overall system complexity while maintaining comprehensive water quality management. This allows conventional control tools to be applied more effectively to smaller, manageable subsystems rather than the entire network at once.
Solution Approach 2:
The patent introduces data-driven methodologies as intermediaries between the complex physical-chemical processes in the network and the control decisions. Machine learning models and data analytics serve as mediators that process sensor data, predict water quality parameters, and provide actionable insights to operators, simplifying the detection and measurement of dynamic parameters without requiring direct complex modeling of all species interactions.
2Reliability
If multiple species interactions (disinfectant, microbes, NOM, DBP) are modeled to ensure water quality, then microbiological safety is improved, but the measurement and control difficulty increases
Solution Approach 1:
The patent implements feedback mechanisms using sensor networks that continuously monitor water quality parameters throughout the distribution network. The collected data is fed back into data-driven models that predict the behavior of multiple species (disinfectant, microbes, NOM, DBP) and provide real-time feedback to control systems. This allows indirect measurement of difficult-to-detect parameters through measurable proxies and predictive modeling, reducing direct measurement difficulty while maintaining microbiological safety.
Solution Approach 2:
The patent replaces direct mechanical/chemical measurement of multiple species interactions with data-driven predictive models. Instead of directly measuring complex interactions between disinfectant, microbes, natural organic matter, and disinfection by-products, the system uses sensor data combined with machine learning algorithms to predict these interactions and their outcomes, significantly reducing measurement difficulty while maintaining model accuracy.
3Reliability
If disinfectant dosage is increased to maintain residual levels and prevent microbial re-growth, then microbiological quality is improved, but harmful Disinfection By-Products formation increases
Solution Approach 1:
The patent applies dynamics by implementing real-time, adaptive disinfectant dosing control based on actual network conditions. Instead of static dosing schedules, the system continuously adjusts disinfectant dosage in response to changing water quality parameters, flow conditions, and demand patterns. This dynamic approach maintains sufficient residual levels for microbiological safety while minimizing excess disinfectant that would form harmful by-products, achieving a balance between the two conflicting objectives.
Solution Approach 2:
The patent changes operational parameters by optimizing disinfectant dosage levels based on predictive modeling and real-time sensor data. The system dynamically adjusts key parameters including disinfectant concentration, dosing timing, and distribution across different network zones. By continuously optimizing these parameters based on actual network conditions, the system maintains microbiological quality while minimizing disinfection by-product formation through precise, condition-based dosing rather than uniform high dosing.
Data Source
AI summary
Embodiments herein provide a system and a method facilitating a local decision making for disinfectant dosing in water as the water flows through a Water Distribution Networks (WDN). WDN data is collected and the WDN is partitioned into one or more sub-systems by using a data based partitioning methodology over the WDN data. One or more decision variables are mapped to the one or more sub-systems for coordinated decentralized control of water quality in large WDN. The decision variables comprise disinfectant dosing rate to be applied to water at the one or more booster stations.