Determining accurate, representative water consumption and the spatial distribution of consumption throughout the network model is a key element of water distribution modeling. Consumption data can be derived from many sources including meter readings and billing records, telemetered system flows, and etc. These data are assigned as demand values at individual junction nodes within the network model, where each junction node represents a distinct demand (service) area polygon coverage. Consumption data must be properly evaluated and processed to accurately categorize customers, forecast demand, and load the network model based on demand location and variation.
Today, InfoWater Pro Demand Analyst brings you unprecedented speed, accuracy, and flexibility for calculating, distributing, and managing consumption data in your network model. It fully automates the process of extracting demand pattern from consumption data stored in various relational database and text file by associating geographical allocation of junction node demand to ensure the development and simulation of credible hydraulic network models.
An indispensable master planning tool, InfoWater Pro DemandAnalyst gives you five highly advanced and efficient geospatial and graph-theoretical (derived from computational geometry) methods for associating meter locations with junctions geometrically:
1. Polygon Containment – Spatial summation of consumption areas.
2. Closest Junction – Assign meter to nearest junction node.
3. Closest Pipe – Assign to the nearest junction of the closest pipe.
4. Meter Junction Assignment – User-defined assignment of meters to junctions.
5. Meter Pipe Assignment – User-defined assignment of meters to pipes.
Once usage patterns and meters are identified, Demand Analyst calculates node base demand and the demand pattern values over 24 hours using usage pattern and meter data. The tool reads the water meter data collected at multiple nodes of which demands are modeled with the same demand pattern.
Water demand is modeled using a base demand and a demand pattern.
Dij = BiPj
Dij = Demand of node i at time interval j
Bi = Base demand of node i
Pj = Demand pattern value at time interval j
Demand pattern is specified for certain period of time, usually one day, with a certain pattern time step (1hour, 0.5 hour, etc). In water demand modeling, it is common to apply one demand pattern to multiple nodes.
The analysis includes the following:
Bi is calculated simply as the average of water demand at node i.
Least square method is used to find the estimate of Pj that minimizes the total squared modeling error, which is the sum of (Dij-BiPj)*(Dij-BiPj) over all nodes and sampling time.
As typical in linear regression, uncertainty of the pattern values estimated from the meter data is analyzed. A confidence value is provided for each estimated pattern value for the calculation of the 95% confidence interval. If the estimate for Pj is Pj’ and the confidence value is Cj, the probability that the true pattern value is between [Pj’-Cj, Pj’+Cj] is 0.95. The larger the value of Cj, the higher uncertain the pattern value estimate.
Average demand modeling error is reported in the analysis summary, which is calculated as the square root of the average value of (Dij-BiPj’)*(Dij-BiPj’)/(Dij*Dij) over all the demand data used in analysis. It is a ratio value and unitless. Under perfect condition (exactly same pattern for every nodes of interest and such pattern repeats exactly day over day), pattern values that lead to zero modeling error can be found. But in reality, there are always modeling errors.
The Demand Analyst uses the Simulation Time Option from the host application. It uses Pattern Time Step, Pattern Start Time, and Start Clock Time from the Simulation Time Options to set the pattern parameters.