Multi-factor correction method for effective wave height value prediction result of single-station waves

Abstract

The invention provides a correction method for a station wave height prediction result. According to the method, based on historical data of a to-be-corrected station before an arbitrary moment T, a weight mapping matrix W and an offset matrix B are self-defined to train an extreme learning machine model, and the weight mapping matrix W, the offset matrix B and an output weight matrix P after training are obtained; and for a to-be-corrected moment after the arbitrary moment T, the weight mapping matrix W, the offset matrix B and the output weight matrix P are adopted to predict a correction error, and the correction error is adopted to correct the value prediction result. Through the multi-factor correction method, the error between a predicted value and a truly observed value can be reduced, the accuracy of the value prediction result is improved, and working efficiency is improved. The method has important significance for timely predicting an ocean storm surge disaster, effectivelycarrying out disaster prevention and reduction work and ensuring the safety of life and properties of people.

Description

technical field [0001] The invention relates to the technical field of ocean forecasting, in particular to a multi-element correction method for numerical forecasting results of ocean wave heights. Background technique [0002] Storm surges caused by typhoons or hurricanes often cause serious economic losses and casualties in coastal areas. If the effective wave height of waves can be accurately predicted, relevant departments can take corresponding measures to reduce economic losses and casualties. Therefore, the reasonable calculation of wave height plays an important guiding role in the construction of coastal engineering. [0003] Most of the existing wave forecasting is the combination of numerical forecasting method and artificial forecasting. Firstly, the historical observation data is used to predict the significant wave height of future waves through numerical models, and then experienced forecasters make corresponding corrections to the numerical forecast results....

AI Technical Summary

Problems solved by technology

This manpower-based correction method of numerical forecast results relies heavily on the experience of the forecaster, which is not only time-consuming and labor-intensive but also highly subjective.

Images