A Lake Water Level Forecasting Method Based on Multi-factor Similarity Analysis

Abstract

The invention discloses a lake water level prediction method based on multi-factor similarity analysis. This method aims at the application shortcomings of existing mechanism models and big data models, and combines the mechanism logic of lake water level prediction and the big data foundation accumulated in the history of lake basins. Starting from the mechanistic data logic, deduce the similarity between the real-time monitoring lake basin data and historical monitoring data, build a lake water level prediction model based on multi-factor similarity analysis, and introduce a multi-factor similarity weight strategy according to the change of dry and flood seasons, which can meet the needs of multiple factors. It is necessary to forecast the water level of lakes under the conditions of hydrological and meteorological monitoring basic conditions and long-term measured data, and it can realize rolling forecast based on real-time monitoring data. The advantages are: in lake basins with a variety of hydrological and meteorological monitoring basic conditions and long-term measured data, the results of this method are more applicable than the currently widely used methods, and continuous precision optimization can be carried out through real-time monitoring data accumulation .

Description

technical field [0001] The invention relates to the technical field of hydrological forecasting, in particular to a lake water level forecasting method based on multi-factor similarity analysis. Background technique [0002] Forecasting the water level of lakes is to make qualitative or quantitative predictions of the hydrological situation of lakes in a certain period of time in the future by using the previous and current hydrometeorological information of lakes. However, due to the relatively backward collection and transmission methods and technologies of hydrometeorological and other information in the early and current stages of the lake, the timeliness of information is poor; or it is difficult to obtain data collected by hydrological and meteorological stations in the upper and lower reaches of the lake, and the easily obtained data is not as good as the traditional hydrological model. Applicable; or the watershed data do not meet the calculation of the traditional m...

AI Technical Summary

Problems solved by technology

The mechanism prediction method can well simulate the hydrological conditions of the watershed and make predictions that conform to the mechanism logic, but it often requires a large amount of hydrological data, which limits its use conditions and reduces the prediction accuracy in areas with insufficient data; the big data prediction method mainly Prediction is made from data mining methods such as artificial neural networks, but because it is only mined from the data relationship of several hydrological data sequences, it lacks the support of mechanism logic, so it is not suitable for water level prediction of large and medium-sized lakes or complex watersheds.

Images