Water Distribution Network Control With Robust Demand Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Water distribution networks operate conservatively to ensure pressure and demand satisfaction with high energy costs due to large safety margins, and real-time re-optimization with varied demand data is computationally expensive and risky.
Innovation Solution
A method for controlling water distribution networks using continuous optimization with robust optimization techniques, incorporating uncertainty ranges in demand forecasts and tank levels, and reformulating discrete decisions into continuous complementarity constraints for fast convergence.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conservative operation mode is used to ensure pressure targets and demand satisfaction, then reliability is improved, but energy consumption increases
Solution Approach 1:
The system performs preliminary actions by maintaining adequate tank levels before demand deviations occur. The control method proactively manages water tank levels to ensure they can absorb demand fluctuations, preventing pressure drops before they happen rather than reacting after problems occur.
Solution Approach 2:
The invention applies beforehand cushioning by using water tanks as buffer elements that absorb demand uncertainties. The control method ensures tanks maintain sufficient levels to cushion against unexpected demand variations, reducing the need for conservative pump operations and lowering energy consumption while maintaining reliability.
2Use of energy by moving object
If optimization based on demand forecasts is applied to reduce energy cost, then energy consumption is reduced, but reliability deteriorates due to pressure drops and empty water supply nodes
Solution Approach 1:
The control method incorporates feedback by continuously monitoring water tank levels and using this information to adjust pump operations. The system uses measured tank levels as feedback to determine whether to operate pumps conservatively or follow optimization recommendations, ensuring pressure targets are maintained while reducing energy consumption.
Solution Approach 2:
The invention uses water tank levels as an intermediary variable to mediate between demand forecasts and pump control decisions. The control method translates forecast information into tank level management strategies, which then guide pump operations to achieve both energy efficiency and pressure target satisfaction.
3Adaptability or versatility
If re-optimization with varied demand data is performed to handle uncertainty, then adaptability is improved, but computational time increases to minutes or hours
Solution Approach 1:
The control method extracts and focuses on the most critical factor for handling demand uncertainty - water tank levels - rather than performing full re-optimization with varied demand data. By isolating tank level management as the key control mechanism, the system achieves adaptability to demand uncertainties with minimal computational effort.
Solution Approach 2:
The invention changes the control parameter from complex demand forecast re-optimization to simple tank level-based pump control decisions. This parameter change allows the system to respond adaptively to demand uncertainties by monitoring and managing tank levels, achieving the same adaptability goal with dramatically reduced computational time suitable for real-time operation.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
A method (300) for controlling a water distribution network (100) is described, including: providing a hydraulic model for characterizing the water distribution network (100); providing a start sequence of control variable values for each of at least one actuator (110a, 110b, 130a, 130b, 130c) of the water distribution network (100); and providing an initial hydraulic head value (115) for each of at least one water supply node (140a, 140b, 140c) of the water distribution network (100); providing an objective for optimized controlling the water distribution network (100); providing a forecast (340) of a sequence of nominal values of a water demand for each at least one water demand node (120a, 120b, 120c, 120d, 120e) of the water distribution network (100) within a time horizon; and providing an associated uncertainty range for each nominal value of the water demand; determining a sequence of control variable values for each of the at least one actuator (110a, 110b, 130a, 130b, 130c) for controlling the water distribution network (100) based on a continuous optimization problem for controlling the water distribution network (100) with respect to the objective, wherein a solver module for the continuous optimization problem is configured for being started based on the provided initial hydraulic head value for each of the at least one water supply node (140a, 140b, 140c) and/or based on the provided start sequence of control variable values; and based on the provided forecast; and wherein the values of the control variables are optimized under consideration of the provided water demand uncertainty range for each at least one water demand node (120a, 120b, 120c, 120d, 120e); and wherein the sequence of control variable values (372) are determined to be provided to the at least one actuator (110a, 110b, 130a, 130b, 130c) for controlling the water distribution network (100) within the time horizon.