Water Demand Prediction Research Articles

Forecast of regional water demand based on NSGAII-FORAGM

Abstract Regional water demand is an important basic data for regional engineering planning, design and management. Making full use of multi-source data and prior knowledge to quickly and economically obtain high-precision regional water demand is of great significance to the optimal allocation of regional water resources. In order to accurately predict the regional water demand, this study took Yulin City as a research area to predict the water demand of the city from 2017 to 2019. Aiming at the oscillating characteristics of the regional water demand sequence and the over-fitting problem of traditional prediction models, this study proposed the non-dominated sorting genetic algorithm II-fractional order reverse accumulative grey model (NSGAII-FORAGM). The regional water demand oscillation sequence was transformed into a monotonically decreasing non-negative sequence. Based on the transformation sequence, an optimization model was constructed according to the two objective functions of ‘maximum (or minimum) order’ and ‘best fit to historical data’, and the NSGAII method was adopted to solve the model. The three model structures of ‘fractional order’, ‘reverse accumulation’ and ‘obtaining order through multi-objective optimization model ‘ were tested based on the water use sequence of the three sectors (industry, tertiary industry and domestic) in Yulin City, and the performance of the method is compared with NSGAII-IORAGM, NSGAII-FOFAGM and SOGA-FORAGM. The results showed that the average relative error of the model established in this study for the simulation of industry, tertiary industry (the tertiary industry is a technical name for the service sector of the economy, which encompasses a wide range of businesses), and domestic was 15.54%, 11.20%, 9.98% respectively. The average relative error of the model established in this study for the prediction of industry, tertiary industry and domestic was 9.46%, 7.9%, and 1.8%, respectively. For the simulation of water demand sequences in three sections, the simulation average relative errors of the other three models were not absolutely dominant except for the SOGA-FORAGM model. The average relative predicted error by the model in this study was the smallest (the relative errors of the three sequence predictions for industry, tertiary industry and domestic were lower than the relative errors of the optimal results of the comparison model, which were 0.97%, 0.72% and 4.5%, respectively), indicating that the model had certain applicability for the water demand prediction of various sectors (industry, tertiary industry and domestic) in the region compared with other models, and can improve the accuracy of the prediction results.

Multi-Model Coupling Water Demand Prediction Optimization Method for Megacities Based on Time Series Decomposition

The water supply in megacities can be affected by the living habits and population mobility, so the fluctuation degree of daily water supply data is acute, which presents a great challenge to the water demand prediction. This is because that non-stationarity of daily data can have a large influence on the generalization ability of models. In this study, the Hodrick-Prescott (HP) and wavelet transform (WT) methods were used to carry out decomposition of daily data to solve the non-stationarity problem. The bidirectional long short term memory (BLSTM), seasonal autoregressive integrated moving average (SARIMA) and Gaussian radial basis function neural network (GRBFNN) were developed to carry out prediction of different subseries. The ensemble learning was introduced to improve the generalization ability of models, and prediction interval was generated based on student's t-test to cope with the variation of water supply laws. This study method was applied to the daily water demand prediction in Shenzhen and cross-validation was performed. The results show that WT is superior to HP decomposition method, but maximum decomposition level of WT should not be set too high, otherwise the trend characteristics of subseries will be weakened. Although the corona virus disease 2019 (COVID-19) outbreak caused a variation in water supply laws, this variation is still within the prediction interval. The WT and coupling models accurately predict water demand and provide the optimal mean square error (0.17%), Nash-Sutcliffe efficiency (97.21%), mean relative error (0.1), mean absolute error (3.32%), and correlation coefficient (0.99).

Sensitivity analysis and prediction of water supply and demand in Shenzhen based on an ELRF algorithm and a self-adaptive regression coupling model

AbstractGiven that sensitive feature recognition plays an important role in the prediction and analysis of water supply and demand, how to conduct effective sensitive feature recognition has become a critical problem. The current algorithms and recognition models are easily affected by multicollinearity between features. Moreover, these algorithms include only a single learning machine, which exposes large limitations in the process of sensitive feature recognition. In this study, an ensemble learning random forest (ELRF) algorithm, including multiple learning machines, was proposed to recognize sensitive features. A self-adaptive regression coupling model was developed to predict water supply and demand in Shenzhen in the next ten years. Results validate that the ELRF algorithm can effectively recognize sensitive features compared with decision tree and regular random forest algorithms. The model used in this study shows a strong self-adaptive ability in the modeling process of multiple regression. The water demand in Shenzhen will reach 2.2 billion m3 in 2025 and 2.35 billion m3 in 2030, which will exceeded the water supply ability of Shenzhen. Furthermore, three scenarios are designed in terms of water supply security and economic operation, and a comparative analysis is performed to obtain an optimal scenario.

Application of edge computing and GIS in ecological water requirement prediction and optimal allocation of water resources in irrigation area.

The purposes are to use water resources efficiently and ensure the sustainable development of social water resources. The edge computing technology and GIS (Geographic Information Science) image data are combined from the perspective of sustainable development. A prediction model for the water resources in the irrigation area is constructed. With the goal of maximizing comprehensive benefits, the optimal allocation of water quality and quantity of water resources is determined. Finally, the actual effect of the model is verified through specific instance data in a province. Results demonstrate that the proposed irrigation area ecological prediction model based on edge computing and GIS images can provide better performance than other state of the art models on water resources prediction. Specifically, the accuracy can remain above 90%. The proposed model for ecological water demand prediction in the irrigation area and optimal allocation of water resources is based on the principle of quality water supply. The optimal allocation of water resources reveals the sustainable development ideas and the requirements of the optimal allocation model, which is very reasonable. The improvement of the system is effective and feasible, and the optimal allocation results are reasonable. This allocation model aims at the water quality and quantity conditions, water conservancy project conditions, and specific water demand requirements in the study area. The calculation results have great practicability and a strong guiding significance for the sustainable utilization and management of the irrigation area.

Model Prediksi Kebutuhan Air Berbasis Sistem Dinamik di Kabupaten Mojokerto

Mojokerto is one of the leading regions in East Java Province This of course has consequences for growth in all fields, especially industry and housing. So that it will have an influence either directly or indirectly on the growth of other supporting facilities growth in various fields leads to increased demand for water due to climate change, several springs in Mojokerto district have decreased by 60% from their original condition. This study aims to predict future water needs with the influence of changes in population, the effect of increasing the number of industries and facilities, both commercial and non-commercial using a dynamic system so that it can be used as a basis for water resource management decisions. In analyzing a complex system that works with real conditions, it is very risky and costly, therefore we need a model that can represent the conditions of the existing system. The stages in the research are as follows: 1. Secondary data collection in the study area which includes: Population data, data on the number of public facilities, data on the number of hotel rooms, data on the number of hospital rooms, data on rice fields, data on the number of livestock, data on the area of tourism and data on the number of markets; (2) Data Analysis; (3) System Dynamic Analysis;(4)Simulation of water demand prediction in the study area;(5) Model validation with the structure validation test and the AVE and AME validation tests. The simulation model for the prediction of water demand in Mojekerto Regency based on a dynamic system is declared valid because it has fulfilled the structure test and validation test both AVE and AME From the results of the scenario simulation applied, it was able to save water by 30% for domestic water needs and 92% for non-domestic water needs. It is necessary to make a model to determine the availability of existing water resources so that a model of water resources balance in Mojokerto Regency is compiled.

Estimating Future Peak Water Demand with a Regression Model Considering Climate Indices

Although Austria is a water-rich country, impacts of climate change on water supply are already noticeable. Some regions were affected by water scarcity in recent years. Due to climate change, an increase in peak water demand is expected in the future. Therefore, water demand prediction models that include climate indices are of interest. In this paper, we present a general multiple linear regression (GMLR) model that can be applied to selected study sites. We compared the performance of the GMLR model with different modeling approaches, i.e., stepwise multiple linear regression, support vector regression, random forest regression and a neural network approach. All models were trained with water demand and weather data reaching back several years and tested with the last available observation year. The applied modeling approaches achieved a similar performance. As a second step, the GMLR model was used to estimate the peak water demands for the time period 2025–2050. For the future water demand estimate, 16 different climate projections were used. These climate projections represent the worst-case climate change scenario (RCP 8.5). The expected increase in peak water demand could be confirmed with the modeling approach. An increase in peak water demand by 3.5% compared to the reference period was estimated.

Research on Forecast of Water Demand in Jinzhou Based on Grey Model

According to the related requirements of water demand prediction in Jinzhou, and based on the current situation of water resource carrying capacity in Jinzhou from 2011 to 2017 and the grey model, the water demand from 2021 to 2025 was predicted, and the grey GM(1,1) model of ecological water demand, virtual aquatic yield, residential water consumption and urban public water consumption was established. The results showed that the regional water demand showed a trend of slow increase, among which the urban public water demand had a larger growth space, while the ecological water demand and virtual aquatic production had a smaller growth range. The accuracy of the grey forecast model is more than 90%, It can provide a new way to forecast the main water use and total water demand.

Research on regional water demand prediction of the upper and middle reaches of The Pearl River Basin based on system dynamics

The prediction and calculation of water resource demand play a crucial role in the balance of water resource supply and demand in the research area. Only when the water demand of the research area is accurately and reasonably grasped, it is possible to make corresponding management and adjustment according to the actual situation in the process of water resource utilization, so as to ensure the sustainable development and utilization of water resource. In this paper, the upper and middle reaches of the Pearl River basin is taken as the research area. According to the relationship between supply and demand and the social and economic situation of the Pearl River basin, VENSIM PLE is used to establish the dynamic model of the water demand forecasting system of the Pearl River Basin. The model involves four subsystems of population, industry, agriculture and water resource, and the water resource demand of the three provinces is simulated. Compared with the results of the pearl River Basin planning report, the prediction error is small, which proves that the system dynamics model is feasible to predict the water demand of the Pearl River Basin.

An Ensemble-Learning-Based Method for Short-Term Water Demand Forecasting

Short-term water demand forecasting has always been a hot research topic in the field of water distribution systems, and many researchers have developed a wide variety of methods based on different prediction periodicities. However, few studies have paid attention to using ensemble learning methods for short-term water demand forecasting. In this study, an ensemble-learning-based method was developed to forecast short-term water demand. The proposed method consists of two models: an equal-dimension and new-information model and an ensemble learning model. The purpose of the equal-dimension and new-information model is to update data for modelling periodically, while the ensemble learning model is used for water demand forecasting. To evaluate the forecasting performance, the proposed method was applied to a data set obtained from a real-world water distribution system and compared with the single back-propagation neural network (BPNN) model and the seasonal autoregressive integrated moving average (SARIMA) model. The results show that the proposed method improves both the accuracy and stability of water demand prediction. The proposed method has the potential to provide a promising alternative for short-term water demand forecasting.

Burst Detection by Water Demand Nowcasting Based on Exogenous Sensors

Bursts of drinking water pipes not only cause loss of drinking water, but also damage below and above ground infrastructure. Short-term water demand forecasting is a valuable tool in burst detection, as deviations between the forecast and actual water demand may indicate a new burst. Many of burst detection methods struggle with false positives due to non-seasonal water consumption as a result of e.g. environmental, economic or demographic exogenous influences, such as weather, holidays, festivities or pandemics. Finding a robust alternative that reduces the false positive rate of burst detection and does not rely on data from exogenous processes is essential. We present such a burst detection method, based on Bayesian ridge regression and Random Sample Consensus. Our exogenous nowcasting method relies on signals of all nearby flow and pressure sensors in the distribution net with the aim to reduce the false positive rate. The method requires neither data of exogenous processes, nor extensive historical data, but only requires one week of historical data per flow/pressure sensor. The exogenous nowcasting method is compared with a common water demand forecasting method for burst detection and shows sufficiently higher Nash-Sutcliffe model efficiencies of 82.7% - 90.6% compared to 57.9% - 77.7%, respectively. These efficiency ranges indicate a more accurate water demand prediction, resulting in more precise burst detection.

Application of LSTM Networks for Water Demand Prediction in Optimal Pump Control

Every morning, water suppliers need to define their pump schedules for the next 24 h for drinking water production. Plans must be designed in such a way that drinking water is always available and the amount of unused drinking water pumped into the network is reduced. Therefore, operators must accurately estimate the next day’s water consumption profile. In real-life applications with standard consumption profiles, some expert system or vector autoregressive models are used. Still, in recent years, significant improvements for time series prediction have been achieved through special deep learning algorithms called long short-term memory (LSTM) networks. This paper investigates the applicability of LSTM models for water demand prediction and optimal pump control and compares LSTMs against other methods currently used by water suppliers. It is shown that LSTMs outperform other methods since they can easily integrate additional information like the day of the week or national holidays. Furthermore, the online- and transfer-learning capabilities of the LSTMs are investigated. It is shown that LSTMs only need a couple of days of training data to achieve reasonable results. As the focus of the paper is on the real-world application of LSTMs, data from two different water distribution plants are used for benchmarking. Finally, it is shown that the LSTMs significantly outperform the system currently in operation.

Short-term water demand predictions coupling an artificial neural network model and a genetic algorithm

Abstract The application of artificial neural network (ANN) models for short-term (15 min) urban water demand predictions is evaluated. Optimization of the ANN model's hyperparameters with a genetic algorithm (GA) and use of a growing window approach for training the model are also evaluated. The results are compared to those of commonly used time series models, namely the Autoregressive Integrated Moving Average (ARIMA) model and a pattern-based model. The evaluations are based on data sets from two Canadian cities, providing 15 min water consumption records over respectively 5 years and 23 months, with a respective mean water demand of 14,560 and 887 m3/d. The GA optimized ANN model performed better than the other models, with Nash–Sutcliffe Efficiencies of 0.91 and 0.83, and relative root mean square errors of 6 and 16% for City 1 and City 2, respectively. The results of this study indicate that the optimization of the hyperparameters of an ANN model can lead to better 15 min urban water demand predictions, which are useful for many real-time control applications, such as dynamic pressure control.

A Water Evaluation and Planning-based framework for the long-term prediction of urban water demand and supply

Increased usage and non-efficient management of limited resources has created the risk of water resource scarcity. Due to climate change, urbanization, and lack of effective water resource management, countries like Pakistan are facing difficulties coping with the increasing water demand. Rapid urbanization and non-resilient infrastructures are the key barriers in sustainable urban water resource management. Therefore, there is an urgent need to address the challenges of urban water management through effective means. We propose a workflow for the modeling and simulation of sustainable urban water resource management and develop an integrated framework for the evaluation and planning of water resources in a typical urban setting. The proposed framework uses the Water Evaluation and Planning system to evaluate current and future water demand and the supply gap. Our simulation scenarios demonstrate that the demand–supply gap can effectively be dealt with by dynamic resource allocation, in the presence of assumptions, for example, those related to population and demand variation with the change of weather, and thus work as a tool for informed decisions for supply management. In the first scenario, 23% yearly water demand is reduced, while in the second scenario, no unmet demand is observed due to the 21% increase in supply delivered. Similarly, the overall demand is fulfilled through 23% decrease in water demand using water conservation. Demand-side management not only reduces the water usage in demand sites but also helps to save money, and preserve the environment. Our framework coupled with a visualization dashboard deployed in the water resource management department of a metropolitan area can assist in water planning and effective governance.

Water Demand Prediction Using Machine Learning Methods: A Case Study of the Beijing–Tianjin–Hebei Region in China

Predicting water demand helps decision-makers allocate regional water resources efficiently, thereby preventing water waste and shortage. The aim of this study is to predict water demand in the Beijing–Tianjin–Hebei region of North China. The explanatory variables associated with economy, community, water use, and resource availability were identified. Eleven statistical and machine learning models were built, which used data covering the 2004–2019 period. Interpolation and extrapolation scenarios were conducted to find the most suitable predictive model. The results suggest that the gradient boosting decision tree (GBDT) model demonstrates the best prediction performance in the two scenarios. The model was further tested for three other regions in China, and its robustness was validated. The water demand in 2020–2021 was provided. The results show that the identified explanatory variables were effective in water demand prediction. The machine learning models outperformed the statistical models, with the ensemble models being superior to the single predictor models. The best predictive model can also be applied to other regions to help forecast water demand to ensure sustainable water resource management.

A data-driven model to evaluate the medium-term effect of contingent pricing policies on residential water demand

The relative scarcity of water resources has encouraged cities to create mechanisms to control water demand and avoid water stress. In the decision-making process, water companies need to assess the price influence on water demand predictions to design better policies. The aim of this study is to estimate the medium-term effectiveness of the implementation of a contingent tariff and its consequences for water demand elasticity to price. A novel model that requires only secondary data is proposed, that can be useful for guiding the drought planning process. The methodology consists of a framework that provides monthly predictions of water demand at the household level, considering price, seasonality, and previous water use. The results indicated that the contingent tariff promoted a reduction of 11–17% in water demand, but at a higher cost for low income households. Also, reduction in water demand was found to be inelastic to price increase. Using google search hits as a proxy for public interest, we found that water cost has a higher influence on users’ decision to save water than drought awareness.

Deep learning with long short-term memory neural networks combining wavelet transform and principal component analysis for daily urban water demand forecasting

A reliable and accurate urban water demand forecasting plays a significant role in building intelligent water supplying system and smart city. Due to the high frequency noise and complicated relationships in water demand series, forecasting the urban water demand is not an easy task. In order to improve the model’s abilities in handling the complex patterns and catching the peaks in time series, we propose a hybrid long short-term memory model combining with discrete wavelet transform (DWT) and principal component analysis (PCA) pre-processing techniques for water demand forecasting, i.e., DWT-PCA-LSTM. First, the outliers of water demand series are identified and smoothed by 3σ criterion and weighted average method, respectively. Then, the noise component of water demand series is eliminated by DWT method and the principal components (PCs) among influencing factors of water demand are selected by PCA method. In addition, two LSTM networks are built to yield the daily urban water demand predictions using the results of DWT and PCA techniques. At last, the superiorities of the proposed model are demonstrated by comparing with the other benchmark predictive models. The water demand from 2016 to 2020 of a waterworks located in Suzhou, China is used for the experiment. The predictive performance of the experiments are evaluated by the mean absolute percentage error (MAPE), mean absolute percentage errors of peaks (pMAPE), explain variance score (EVS) and correlation coefficient (R). The results show that the DWT-PCA-LSTM model outperforms the other models and has satisfactory performance both in catching the peaks and the average prediction accuracy.

Intelligent Control of Agricultural Irrigation through Water Demand Prediction Based on Artificial Neural Network.

In irrigated areas, the intelligent management and scientific decision-making of agricultural irrigation are premised on the accurate estimation of the ecological water demand for different crops under different spatiotemporal conditions. However, the existing estimation methods are blind, slow, or inaccurate, compared with the index values of the water demand collected in real time from irrigated areas. To solve the problem, this paper innovatively introduces the spatiotemporal features of ecological water demand to the forecast of future water demand by integrating an artificial neural network (ANN) for water demand prediction with the prediction indices of water demand. Firstly, the ecological water demand for agricultural irrigation of crops was calculated, and a radial basis function neural network (RBFNN) was constructed for predicting the water demand of agricultural irrigation. On this basis, an intelligent control strategy was presented for agricultural irrigation based on water demand prediction. The structure of the intelligent control system was fully clarified, and the main program was designed in detail. The proposed model was proved effective through experiments.

Assessment of Water Supply and Demand in the Lower Mahi Sub-basin using WEAP Model

This study deals with the future water demand assessment in Lower Mahi Sub-basin, Bihar, India using Water Evaluation and Planning Model (WEAP). The river Mahi is flowing through the Mahi sub-basin, which is a tributary to the Ganga river. Understanding and analyzing the past and present water supply and demand scenarios and prediction of future scenarios are indispensable for this area. In the WEAP model development phase, two scenarios are considered: past scenario (2004-2013), present and future scenarios (2016-2021) together. The water supply nodes used are Mahi river, groundwater and canals from Gandak river. The results from the trend of supply and demand shows that the sub-basin is water deficit for most part of year except during July to September. The water deficit is also observed during Rabi season when there is no rainfall or availability of surface water is nil for the crops, especially wheat. This results in more and more usage of groundwater and hence leads to groundwater depletion. By analyzing the results of the WEAP Model for the present and future scenario, a depletion in groundwater storage at a faster rate is observed with the existing climatic conditions and increased population, livestock and industry.

Prediction model of ecological environmental water demand based on big data analysis

The existing prediction model of eco-environmental water demand has the problem of large prediction error. In order to solve the above problems, the prediction model of eco-environmental water demand is constructed based on big data analysis. In order to reduce the prediction error of the ecological environment water demand prediction model, the framework of the ecological environment water demand prediction model is built. On this basis, the principal component analysis method is used to select the auxiliary variables of the model. Based on the selected auxiliary variables, the minimum monthly average flow method is used to analyze the basic water demand of the ecological environment, the leakage water demand and the water surface evaporation ecological environment water demand, so as to analyze based on the results, the water demand of ecological environment is predicted by big data analysis technology, and the prediction of water demand of ecological environment is realized. The experimental results show that compared with the existing ecological environment water demand prediction model, the prediction error of the model is within 19.3, which fully shows that the constructed ecological environment water demand prediction model has better prediction effect and can provide a certain reference value for the actual use of water resources. • The framework of water demand prediction model reduces the prediction error. • Using the 3 σ rule to eliminate the abnormal data. • The minimum monthly average flow method is used to obtain the basic water demand.

WATER CONSUMPTION PREDICTION USING FUZZY TIME SERIES - A CASE STUDY IN PRIVATE COMPANY OF TANGERANG DISTRICT INDONESIA

Consumption of water in the Tangerang Regency continuously increases from year to year due to the increasing population and birth rates an average increase of 3% every year. So, the water demand prediction to be important to meet customer or community needs. The private water utility company needs to use a new method for predicting future monthly water consumption values and improves accuracy when forecasting time series using a visibility graph and presents to make more accurate predictions. In this study, we aim to measure the trend analysis volume of water consumption prediction by Fuzzy Time Series versus actual usage volume. Fuzzy Time Series (FTS) is a concept plan method that uses fuzzy logic that is able to provide predictions (estimates) of time series data analysis for the next several periods. Mean Absolute Percentage Error (MAPE) is obtained for different configurations of the input sets and of the FTS model structure. From the results of the average value error accuracy was only 4.5% using FTS Chen Method and included in the low category and water consumption actual versus prediction with the FTS Chen method shown related stable.