Short-term water demand forecasting using hybrid supervised and unsupervised machine learning model

Regression Tree (RT) forecasting models are widely used in short-term demand forecasting. Likewise, Self-Organizing Maps (SOM) models are known for their ability to cluster and organize unlabeled big data. Herein, a combination of these two Machine Learning (ML) techniques is proposed and compared to a standalone RT and a Seasonal Autoregressive Integrated Moving Average (SARIMA) models, in forecasting the short-term water demand of a municipality. The inclusion of the Unsupervised Machine Learning clustering model has resulted in a significant improvement in the performance of the Supervised Machine Learning forecasting model. The results show that using the output of the SOM clustering model as an input for the RT forecasting model can, on average, double the accuracy of water demand forecasting. The Mean Absolute Percentage Error (MAPE) and the Normalized Root Mean Squared Error (NRMSE) were calculated for the proposed models forecasting 1 h, 8 h, 24 h, and 7 days ahead. The results show that the hybrid models outperformed the standalone RT model, and the broadly used SARIMA model. On average, hybrid models achieved double accuracy in all 4 forecast periodicities. The increase in forecasting accuracy afforded by this hybridized modeling approach is encouraging. In our application, it shows promises for more efficient energy and water management at the water utilities.

Analysis of river bed variation using SSARR and RMA-2 models

Future changes in river bed were predicted under the assumption that flow velocity of past and changes in flow rate at upstream river due to construction of large-scale artificial structures downstream occur in the future. Therefore, the long-term runoff volume from the downstream part of Hosan Stream was estimated using the SSARR (Stream Synthesis and Reservoir Regulation Model). Changes in the river bed were simulated using RMA-2 and SED-2D, which are hydraulic models. As a result, it was found that the river bed variation is significantly affected by the inclusion of sediment in flood flow at upstream. A comprehensive evaluation of above results revealed that the river width has significantly affected flow rate, and the inclusion of sediment in flood flow from the upstream has a huge effect on changes in the riverbed. In this regard, there is a need to devise measures to mitigate future flood damage to artificial structures by reflecting sedimentation trends downstream before the construction of large-scale artificial structures at downstream of river.

A design and implementation of low-power ultrasonic water meter

Smart water meter, which incorporates IoT (Internet of Things) technology, is receiving high attention due to recent development of information and communication technology. If traditional mechanical water meters are replaced by electronic ultrasonic water meters, micro flow rate can be measured and the measurement uncertainty can be improved due to the age of use. This enables smart metering such as AMR (Automatic Meter Reading) or AMI (Advanced Metering Infrastructure) as well as various water related services. In this paper, a low power ultrasonic water meter will be designed to operate with a battery for a long period of time. A water meter shall be designed to operate for at least 9 years, which is the requirement for type approval. In this paper, a low-power modeling is performed for battery-operated ultrasonic water meter to work for at least 10 years. The proposed low power embedded system model will be verified with actual test circuits.