River flood routing computes changes in shape of a flood wave over time as it travels downstream along a river. Conventional flood routing models, especially hydrodynamic models require high quality and quantity of input data such as measured hydrologic time series, geometric data, hydraulic structures and hydrological parameters. Unlike physically based models, machine learning algorithms, which are data driven models, do not require much knowledge about underlying physical processes and can identify complex nonlinearity between inputs and outputs. Due to the higher performance, less complexity, and low computation cost, novel machine learning methods as a single application or hybrid application were introduced by researchers to achieve more accurate and efficient flood routing. This paper reviews the recent application of machine learning methods in river flood routing.

In South Korea, in line with the increasing need for a reliable water supply following the continuous increase in water demand, the Smart Water Grid Research Group (SWGRG) was officially launched in 2012. With the vision of providing water welfare at a national level, SWGRG incorporated information and communications technology in its water resource management, aiming at the development of core technologies for a Smart Water Grid consisting of intelligent microgrids and a demonstration and tests of the developed technologies through a field operation in a living lab. The living lab was built in Block 112 of YeongJong Island, Incheon, South Korea (area of 17.4 km2, population of 8,000), where Incheon International Airport, a hub of Northeast Asia, is located. In this location, water is supplied through a single submarine pipeline, making the place optimal for responses to water crises and the construction of a water supply system during emergencies. From 2017 to 2019, ultrasonic wave type smart water meters and IEEE 802.15.4g advanced metering infrastructure (AMI) networks were installed at 527 sites of 958 consumer areas in the living lab, and core element technologies (intelligent water source management and distribution system, smart water distribution network planning/control/operation strategy establishment, AMI network and device development, integrated management of bi-directional smart water information), and operation solutions (smart water statistics information, real-time demand-supply analysis, decision support system, real-time hydraulic pipeline network analysis, smart DB management, and water information mobile application) were developed through a field operation and testing.

It is crucial to forecast the water demand accurately for supplying water efficiently and stably in a water supply system. In particular, accurately forecasting short-term water demand helps in saving energy and reducing operating costs. With the introduction of the Smart Water Grid (SWG) in a water supply system, the amount of water consumption is obtained in real time through an advanced metering infrastructure (AMI) sensor, which can be used for forecasting the short-term water demand. The models widely used for water demand forecasting include the autoregressive integrated moving average, radial basis function-artificial neural network, quantitative multi-model predictor plus, and long short-term memory. However, there is a lack of research on assessing the performance of models and forecasting the short-term water demand by applying the data on the amount of water consumption by purpose and the pipe diameter of an end-use level of the SWG demonstration plant in each demand forecasting model. Therefore, in this study, the short-term water demand was forecasted for each model using the data collected from the AMI, and the performance of each model was assessed. The Smart Water Grid Research Group installed ultrasonic-wave-type AMI sensors in the block 112 located in YeongJong Island, Incheon, and the actual data used for operating the SWG demonstration plant were adopted. The performance of the model was assessed by using the residual, root mean square error (RMSE), normalized root mean square error (NRMSE), Nash–Sutcliffe efficiency (NSE), and Pearson correlation coefficient (PCC) as indices. The water demand forecast was slightly underestimated in models that employed the assessment results based on the RMSE and NRMSE. Furthermore, the forecasting accuracy was low for the NSE due to a large number of negative values; the correlation between the observed and forecasted values of the PCC was not high, and it was difficult to forecast the peak amount of water consumption. Therefore, as the short-term water demand forecasting models using only time and the amount of water consumption have limitations in reflecting the characteristics of consumers, a water supply system can be managed more precisely if other factors (weather, customer behavior, etc.) influencing the water demand are applied.