A heat island simulation and forecast method of an urban planning scheme based on grey neural network CA model

Abstract

The invention provides a heat island simulation and forecast method of an urban planning scheme based on grey neural network CA model, which comprises the following steps: 1) collecting data of urbangreen space, hardened ground, building layout, volume ratio, building density and water body that affect the simulating and predicting of the heat island; 2) calculating and obtaining that kernel density of each factor in each divide area; 3) establishing a regression equation; according to the typical weather test results and regression parameters, using the grey neural network CA model to trainand verify the model, wherein the output results are the heat island effect of the simulation and prediction program. The invention provides a method for controlling and evaluating an outdoor thermalenvironment of a town. According to the spatial influence factors of urban heat island, the spatial distribution of urban heat island intensity at any time is simulated under the given meteorologicalconditions, so people can adjust the distribution of hardened ground such as building layout, green space layout and road square in urban planning according to the simulation results, which provides apowerful analysis tool for ecological town planning.

Description

technical field [0001] The invention relates to a method for simulating and predicting a heat island of an urban planning scheme based on a gray neural network CA model. Background technique [0002] Urban heat island (UHI) is a phenomenon in which cities are warmer than surrounding rural areas. In 1833, the British Howard (Howard) first proposed the urban heat island effect in the scientific magazine "London Climate". In the following 180 years, western countries carried out extensive research on urban heat islands, so there are endless literature on this aspect. The urban heat island effect varies with time and geographical location. The research on the heat island effect can be divided into two types according to the object: surface temperature and atmospheric temperature near the surface. In terms of research time, there are long-term change studies, annual change studies, and daily change studies. The scope of research ranges from hundreds of square kilometers at th...

AI Technical Summary

Problems solved by technology

[0005] Whether it is numerical models, urban canopies, or cellular automata, existing research is based on larger or smaller scales, and the WRF simulation scale is less than 3km 2 The resolution is difficult to accurately simulate, and the existing cellular automata models are also based on the 120m-scale surface heat island, mostly based on NDBI, DNVI and other indices, which cannot be connected with real cities to achieve real atmospheric temperature, and it is also difficult to match the actual planning The index connection of the design project, not to mention the practical application

Moreover, at present, neural networks and gray systems are used to study heat islands, most of which are based on time-based GM(1,1) models, and this model is also very unsuitable for simulation predictions based on current ground objects, because GM(1,1) The model is based on the prediction of the future based on time series data, while the prediction of the heat island is based on the prediction of the combined impact of multiple factors at the same time

Images