Water environment mathematical model water quality parameter uncertainty and sensitivity analysis method

Abstract

The invention discloses a water environment mathematical model water quality parameter uncertainty and sensitivity analysis method, which comprises the following steps of: screening water quality parameters to be analyzed and prior distribution characteristics, and taking a value range; extracting a proper amount of parameter groups by using a Latin hypercube sampling method; calculating a water quality simulation result of the water quality parameter through the water environment mathematical model; selecting a specific confidence interval as an upper boundary and a lower boundary for judgingan uncertainty interval, defining an interval coverage rate, an uncertainty interval width and a relative uncertainty interval width, and evaluating the uncertainty of the water environment mathematical model; measuring the sensitivity of the water quality parameters through a sensitivity analysis chart; and finally, optimizing the value range of the sensitive water quality parameters through parameter posteriori distribution. Water quality parameter evaluation is more comprehensive, and the problem of same effect of different parameters is effectively solved; parameter sensitivity is quantified, and parameter posteriori distribution is obtained, and the value range of sensitive parameters is narrowed, and a basis is provided for parameter constant value selection and model sensitivity improvement.

Description

technical field [0001] The invention relates to a water quality parameter analysis method, in particular to a water environment mathematical model water quality parameter uncertainty and sensitivity analysis method, and belongs to the technical field of water environment mathematical model parameter uncertainty analysis. Background technique [0002] Uncertainty and sensitivity analysis are research hotspots in numerical models. The uncertainty of mathematical model simulation results mainly comes from three aspects: parameter uncertainty, input data uncertainty, and model structure uncertainty. Among them, the uncertainty of parameters is mainly due to the fact that the actual meanings of each parameter are different, some parameters are difficult to obtain accurate values, and the parameters estimated by experience or measured values ​​may deviate from the actual values. The uncertainty and sensitivity analysis of parameters plays an important role in optimizing the value ...

AI Technical Summary

Problems solved by technology

The existing uncertainty or sensitivity analysis methods cannot effectively solve this problem, and the existing research usually analyzes the uncertainty results or parameter sensitivity separately, and often obtains the uncertainty range and parameter sensitivity of the model results. In fact, there is a close relationship between uncertainty and sensitivity, and the combination of uncertainty and sensitivity analysis is necessary for quantitative analysis of parameter sensitivity

[0004] In terms of parameter estimation, the acquisition channels of parameter value range and prior distribution are very limited. Reducing the value range of parameters is an important result of uncertainty analysis. However, the improvement of prior distribution is rare, and parameters cannot be optimized at the same time. It is one of the defects of the existing analysis methods to trace back to its distribution law through the value range of parameters

Images