Intelligent calibration method for full-scale parameters of online flow monitoring model

To address the disadvantages of single model form, low efficiency of manual parameter calibration, and insufficient adaptability to environment in traditional online flow monitoring (OFM) technology, and efficiently calibrate the parameters of the OFM model, this paper uses the cuckoo algorithm to intelligently divide flow intervals and calibrate the full-scale parameter set for high, medium, and low water levels of the OFM model. The observed flow results are used to compare and analyze the applicability and rationality of the optimal parameter set, with system error and random uncertainty as accuracy evaluation indicators. The Baihe station of the Han River was selected for case study. A full-scale parameter set of the OFM model was developed for Baihe station. The parameter fluctuations within each flow interval are less than 10%. The optimized OFM model based on this parameter set showed high accuracy in various flow scenarios, with system errors and random uncertainties within 1% and 10%, respectively. The system errors are less than 0.5% for flow intervals with more samples. Especially, it is helpful to improve the OFM capabilities of extreme flow by dividing the flow interval to calibrate parameters. The proposed method can achieve fast, real-time, and accurate parameter calibration of OFM model under complex water conditions, which has obvious accuracy improvement effect and reference value for hydrological stations that are significantly affected by environments such as backwater jacking and complex cross-sectional shapes. It can provide technical support for digital twin hydrological stations, digital water network construction, and water resource management.