WO2023056682A1 - Ozone layer prediction algorithm based on artificial intelligence - Google Patents

Ozone layer prediction algorithm based on artificial intelligence Download PDF

Info

Publication number
WO2023056682A1
WO2023056682A1 PCT/CN2021/129850 CN2021129850W WO2023056682A1 WO 2023056682 A1 WO2023056682 A1 WO 2023056682A1 CN 2021129850 W CN2021129850 W CN 2021129850W WO 2023056682 A1 WO2023056682 A1 WO 2023056682A1
Authority
WO
WIPO (PCT)
Prior art keywords
ozone
neural network
value
day
concentration
Prior art date
Application number
PCT/CN2021/129850
Other languages
French (fr)
Chinese (zh)
Inventor
史剑
杜辉
张文
郭海龙
曾智
张雪艳
汪浩笛
Original Assignee
中国人民解放军国防科技大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 中国人民解放军国防科技大学 filed Critical 中国人民解放军国防科技大学
Publication of WO2023056682A1 publication Critical patent/WO2023056682A1/en

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/28—Design optimisation, verification or simulation using fluid dynamics, e.g. using Navier-Stokes equations or computational fluid dynamics [CFD]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/0418—Architecture, e.g. interconnection topology using chaos or fractal principles
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/084—Backpropagation, e.g. using gradient descent
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2113/00—Details relating to the application field
    • G06F2113/08—Fluids
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
    • G06F2119/14—Force analysis or force optimisation, e.g. static or dynamic forces
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
    • Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
    • Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation

Definitions

  • the invention belongs to the technical field of ozone layer forecasting, and in particular relates to an artificial intelligence-based ozone layer forecasting algorithm.
  • the ozone layer is the stratosphere of the atmosphere where the concentration of ozone is high. The part with the greatest concentration is located at an altitude of 20-25 kilometers. If the ozone of the ozone layer is corrected to the standard situation, its thickness is only about 3 millimeters on average. Ozone content varies with latitude, season and weather. Ultraviolet radiation is absorbed by ozone at high altitudes, which has a warming effect on the atmosphere. At the same time, it protects the living things on the earth from the damage of far ultraviolet radiation. The small amount of ultraviolet radiation that passes through has a bactericidal effect and is of great benefit to living things.
  • the purpose of the present invention is to provide an artificial intelligence-based ozone layer forecasting algorithm to solve the problems raised in the above-mentioned background technology.
  • the ozone layer prediction algorithm based on artificial intelligence comprises the following steps:
  • Establish an ozone concentration monitoring station select a location that has no impact on the surrounding environment and the discharge of active pollutants as the ozone concentration monitoring station;
  • S2. collect and obtain historical meteorological data: obtain the daily concentration data of each ozone concentration monitoring site history every day, the daily concentration data predicted every day of each ozone concentration monitoring site history and the reference corresponding to the history of each ozone concentration monitoring site every day As a result, the reference result corresponding to the history of each ozone concentration monitoring site every day is the difference between the daily concentration data of each ozone concentration monitoring site history and the daily concentration data of each ozone concentration monitoring site history every day;
  • Select influencing factors select yesterday's O 3 -8h value, PM2.5 daily average concentration, NO2 daily average concentration, forecasted daily average air pressure, daily average wind speed and daily average temperature as influencing factors;
  • Preliminary prediction of the O 3 -8h value in one day by fitting, the predicted value and the measured value of the meteorological parameters are substituted into the prediction equation, and the to-be-predicted value of each ozone concentration monitoring station is obtained by obtaining the equation Predict the O 3 -8h value of a day;
  • the network is divided into three parts: an input layer, a hidden layer and an output layer, and in the network, the neurons of the input layer, the hidden layer and the output layer are fully connected;
  • the chaotic artificial neural network is to mix the new hyperchaotic system with the artificial neural network, and the chaotic artificial neural network is to carry out the ozone concentration through the artificial neural network, BP and multiple linear regression model predict;
  • the pollutant concentration data in S1 includes sulfur dioxide SO2, nitrogen dioxide NO2, O 3 , O 3 -8h, fine particulate matter with an aerodynamic diameter of 2.5 microns or smaller PM2.5 and aerodynamic diameter PM10 is fine particulate matter 10 microns or smaller.
  • the collection and acquisition of historical meteorological data in S2 also includes obtaining the point source emission characteristics corresponding to any historical moment of each ozone concentration monitoring station, and the point source emission characteristics corresponding to any historical moment of each ozone concentration monitoring station Source emission characteristics are obtained through the following steps:
  • the sensitive area includes all cities bordering on the city where each ozone concentration monitoring site is located;
  • the prediction equation of fitting by way of fitting in said S4 is:
  • c is the predicted value of O 3 -8h on the day to be predicted
  • x1, x2, ..., x6 respectively represent the average concentration of NO2, average temperature, average air pressure, average wind speed, and average concentration of PM2.5 on the day to be predicted
  • the predicted value of , and the predicted or measured value of yesterday's O 3 -8h of the day to be predicted, a0, a1, ..., a6 are regression coefficients.
  • the new hyper-chaotic system in the S5 is described by the following formula:
  • the corresponding hyperchaotic complex system is given by:
  • ⁇ 1 u1+ju2
  • ⁇ 2 u3+ju4
  • ⁇ 3 u5
  • j -1-0.5
  • the bar above the variable represents the complex conjugate of the variable
  • the formula of S8 using chaotic artificial neural network for long-term and short-term forecasting is as follows:
  • n the number of samples
  • xobs,i the observed value
  • xpre,i the predicted value
  • Var the sample variance
  • the chaotic artificial neural network in the S7 adopts CANN operation, and the CANN operation is usually divided into two steps: first, setting parameters, selecting an activation function and an operation mode, and using a sigmoid function for the activation function, and setting The operation mode is set to regression; second, 70%-80% of the sample data are selected for training, and the rest are used as test data, and the weights and thresholds are automatically set in CANN.
  • an early warning of ozone pollution is also included, and the early warning of ozone pollution includes an early warning module, which is used for early warning of ozone pollution and determining the interval and pollution level of ozone pollution.
  • the method for ozone pollution early warning comprises the following steps:
  • the physical parameters include air temperature, total solar radiation irradiance, cloud cover, particle concentration, and nitrogen oxide concentration;
  • S202 Determine the inflection point value of each physical parameter described in S201 according to atmospheric ozone generation conditions, and use the inflection point value to perform interval division and assign a value to each interval;
  • the present invention adopts the CANN operation similar to other neural networks in that it does not depend on the complex relationship between parameters and outputs, but it relies on constant changes in weights, so that parameters and outputs are closely related, avoiding cumbersome mathematical modeling.
  • the present invention has high generalization, can maintain high prediction accuracy while reducing some input parameters, can achieve good results in long-term and short-term ozone prediction, and can accurately and efficiently predict ozone Concentration is conducive to the protection of the environment.
  • Fig. 1 is a flowchart of the present invention.
  • the present invention provides a kind of technical scheme: the ozone layer prediction algorithm based on artificial intelligence comprises the following steps:
  • Establish an ozone concentration monitoring station select a location that has no impact on the surrounding environment and the discharge of active pollutants as the ozone concentration monitoring station;
  • S2. collect and obtain historical meteorological data: obtain the daily concentration data of each ozone concentration monitoring site history every day, the daily concentration data predicted every day of each ozone concentration monitoring site history and the reference corresponding to the history of each ozone concentration monitoring site every day As a result, the reference result corresponding to the history of each ozone concentration monitoring site every day is the difference between the daily concentration data of each ozone concentration monitoring site history and the daily concentration data of each ozone concentration monitoring site history every day;
  • Select influencing factors select yesterday's O3-8h value, PM2.5 daily average concentration, NO2 daily average concentration, forecasted daily average air pressure, daily average wind speed and daily average temperature as the influencing factors;
  • Preliminary prediction of the O 3 -8h value in one day by fitting, the predicted value and the measured value of the meteorological parameters are substituted into the prediction equation, and the to-be-predicted value of each ozone concentration monitoring station is obtained by obtaining the equation Predict the O 3 -8h value of a day;
  • the network is divided into three parts: an input layer, a hidden layer and an output layer, and in the network, the neurons of the input layer, the hidden layer and the output layer are fully connected;
  • the chaotic artificial neural network is to mix the new hyperchaotic system with the artificial neural network, and the chaotic artificial neural network is to carry out the ozone concentration through the artificial neural network, BP and multiple linear regression model predict;
  • the pollutant concentration data in S1 includes sulfur dioxide SO2, nitrogen dioxide NO2, O3 , O3-8h , fine particulate matter PM2.5 with an aerodynamic diameter of 2.5 microns or less and fine particulate matter PM10 with an aerodynamic diameter of 10 microns or less.
  • the collection and acquisition of historical meteorological data in S2 also includes obtaining the point source emission characteristics corresponding to any historical moment of each ozone concentration monitoring site, and any of the ozone concentration monitoring sites
  • the point source emission characteristics corresponding to historical moments are obtained through the following steps:
  • the sensitive areas include all cities bordering on the city where each ozone concentration monitoring site is located;
  • the prediction equation of fitting by way of fitting in said S4 is:
  • c is the predicted value of O 3 -8h on the day to be predicted
  • x1, x2, ..., x6 respectively represent the average concentration of NO2, average temperature, average air pressure, average wind speed, and average concentration of PM2.5 on the day to be predicted
  • the predicted value of , and the predicted or measured value of yesterday's O 3 -8h of the day to be predicted, a0, a1, ..., a6 are regression coefficients.
  • the new hyper-chaotic system in the S5 is described by the following formula:
  • the corresponding hyperchaotic complex system is given by:
  • ⁇ 1 u1+ju2
  • ⁇ 2 u3+ju4
  • ⁇ 3 u5
  • j -1-0.5
  • the bar above the variable represents the complex conjugate of the variable
  • the formula of S8 using chaotic artificial neural network for long-term and short-term forecasting is as follows:
  • n the number of samples
  • xobs,i the observed value
  • xpre,i the predicted value
  • Var the sample variance
  • the chaotic artificial neural network in S7 adopts CANN operation, and the CANN operation is usually divided into two steps: the first, setting parameters, selecting activation function and operation mode, for activation function, use sigmoid function, and set the operation mode to regression; second, select 70%-80% of the sample data for training, and the rest as test data, and the weights and thresholds are automatically set in CANN.
  • an early warning of ozone pollution is also included, and the early warning of ozone pollution includes an early warning module, and the early warning module is used for early warning of ozone pollution, and determining the interval and pollution level of ozone pollution.
  • the method for ozone pollution early warning comprises the following steps:
  • the physical parameters include air temperature, total solar radiation irradiance, cloud cover, particle concentration, and nitrogen oxide concentration;
  • S202 Determine the inflection point value of each physical parameter described in S201 according to atmospheric ozone generation conditions, and use the inflection point value to perform interval division and assign a value to each interval;
  • the present invention adopts CANN operation similar to other neural networks in that it does not depend on the complex relationship between parameters and outputs, but it relies on the constant change of weights, so that parameters and outputs are closely related , avoiding cumbersome mathematical modeling;
  • the present invention has high generalization, can reduce some input parameters while maintaining high prediction accuracy, and can achieve good results in long-term and short-term ozone prediction, and The ozone concentration can be accurately and efficiently predicted, which is beneficial to the protection of the environment.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Computing Systems (AREA)
  • Software Systems (AREA)
  • Artificial Intelligence (AREA)
  • Mathematical Physics (AREA)
  • Computational Linguistics (AREA)
  • Computer Hardware Design (AREA)
  • General Health & Medical Sciences (AREA)
  • Data Mining & Analysis (AREA)
  • Biophysics (AREA)
  • Biomedical Technology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Geometry (AREA)
  • Molecular Biology (AREA)
  • Medical Informatics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Algebra (AREA)
  • Fluid Mechanics (AREA)
  • Mathematical Analysis (AREA)
  • Mathematical Optimization (AREA)
  • Pure & Applied Mathematics (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

Disclosed in the present invention is an ozone layer prediction algorithm based on artificial intelligence, the algorithm comprising the following steps: S1, establishing an ozone concentration monitoring site; S2, collecting and acquiring historical meteorological data; S3, selecting an impact factor; S4, performing preliminary prediction on one-day O3-8h values; S5, establishing a new hyperchaotic system; S6, establishing an artificial neural network; S7, establishing a chaotic artificial neural network; and S8, using the chaotic artificial neural network to perform long- and short-term prediction. The present invention is conducive to simplifying a research process of traditional numerical weather prediction methods; and the traditional numerical weather prediction methods tend to be relatively complex, and have high requirements for calculation. The similarity between a CANN operation used in the present invention and other neural networks is that the CANN operation does not rely on a complex relationship between parameters and outputs, but relies on the constant change of weights, such that the parameters are tightly associated with the outputs, thereby avoiding tedious mathematical modeling.

Description

基于人工智能的臭氧层预报算法Ozone Layer Forecast Algorithm Based on Artificial Intelligence 技术领域technical field
本发明属于臭氧层预报技术领域，具体涉及基于人工智能的臭氧层预报算法。The invention belongs to the technical field of ozone layer forecasting, and in particular relates to an artificial intelligence-based ozone layer forecasting algorithm.
背景技术Background technique
臭氧层是大气层的平流层中臭氧浓度高的层次。浓度最大的部分位于20—25公里的高度处。若把臭氧层的臭氧校订到标准情况，则其厚度平均仅为3毫米左右。臭氧含量随纬度、季节和天气等变化而不同。紫外辐射在高空被臭氧吸收，对大气有增温作用，同时保护了地球上的生物免受远紫外辐射的伤害，透过的少量紫外辐射，有杀菌作用，对生物大有裨益。The ozone layer is the stratosphere of the atmosphere where the concentration of ozone is high. The part with the greatest concentration is located at an altitude of 20-25 kilometers. If the ozone of the ozone layer is corrected to the standard situation, its thickness is only about 3 millimeters on average. Ozone content varies with latitude, season and weather. Ultraviolet radiation is absorbed by ozone at high altitudes, which has a warming effect on the atmosphere. At the same time, it protects the living things on the earth from the damage of far ultraviolet radiation. The small amount of ultraviolet radiation that passes through has a bactericidal effect and is of great benefit to living things.
随着结构性减排措施的不断深入，环保技术的大面积推广，以及清洁能源的逐渐普及，颗粒物浓度实现了逐年下降。然而，臭氧污染却“不降反升”，成为下阶段大气污染防治工作亟待解决的难题。大气臭氧污染成因复杂，给现实治理工作带来很大困难。With the continuous deepening of structural emission reduction measures, the large-scale promotion of environmental protection technologies, and the gradual popularization of clean energy, the concentration of particulate matter has been reduced year by year. However, ozone pollution has "increased instead of falling", which has become an urgent problem to be solved in the next stage of air pollution prevention and control. The causes of atmospheric ozone pollution are complex, which brings great difficulties to the actual treatment work.
目前现有的基于人工智能的臭氧层预报算法还存在一些问题：数学建模过程比较繁琐，不方便对臭氧浓度进行准确的预报，不利于环境保护，为此我们提出基于人工智能的臭氧层预报算法。At present, there are still some problems in the existing artificial intelligence-based ozone layer forecasting algorithm: the mathematical modeling process is cumbersome, it is inconvenient to accurately forecast the ozone concentration, and it is not conducive to environmental protection. Therefore, we propose an artificial intelligence-based ozone layer forecasting algorithm.
发明内容Contents of the invention
本发明的目的在于提供基于人工智能的臭氧层预报算法，以解决上述背景技术中提出的问题。The purpose of the present invention is to provide an artificial intelligence-based ozone layer forecasting algorithm to solve the problems raised in the above-mentioned background technology.
为实现上述目的，本发明提供如下技术方案：基于人工智能的臭氧层预报算法，包括以下步骤：To achieve the above object, the present invention provides following technical scheme: the ozone layer prediction algorithm based on artificial intelligence comprises the following steps:
S1.建立臭氧浓度监测站点：选取对周围环境和活性污染物的排放没有影响的位置作为臭氧浓度监测站点；S1. Establish an ozone concentration monitoring station: select a location that has no impact on the surrounding environment and the discharge of active pollutants as the ozone concentration monitoring station;
S2.收集、获取历史气象数据：获取每一臭氧浓度监测站点历史每天的日 浓度数据、每一臭氧浓度监测站点历史每天预测的日浓度数据和所述每一臭氧浓度监测站点历史每天对应的参考结果，所述每一臭氧浓度监测站点历史每天对应的参考结果为每一臭氧浓度监测站点历史每天预测的日浓度数据与每一臭氧浓度监测站点历史每天的日浓度数据之差；S2. collect and obtain historical meteorological data: obtain the daily concentration data of each ozone concentration monitoring site history every day, the daily concentration data predicted every day of each ozone concentration monitoring site history and the reference corresponding to the history of each ozone concentration monitoring site every day As a result, the reference result corresponding to the history of each ozone concentration monitoring site every day is the difference between the daily concentration data of each ozone concentration monitoring site history and the daily concentration data of each ozone concentration monitoring site history every day;
S3.选取影响因子：选取昨日O 3-8h值、PM2.5日均浓度、NO2日均浓度、预测的日平均气压、日平均风速和日平均气温作为影响因子； S3. Select influencing factors: select yesterday's O 3 -8h value, PM2.5 daily average concentration, NO2 daily average concentration, forecasted daily average air pressure, daily average wind speed and daily average temperature as influencing factors;
S4.对一天O 3-8h值进行初步预测：通过拟合的方式，将气象参数的预测值和实测值代入预测方程中，通过求取方程来对每一臭氧浓度监测站点的要预测的那一天的O 3-8h值进行预测； S4. Preliminary prediction of the O 3 -8h value in one day: by fitting, the predicted value and the measured value of the meteorological parameters are substituted into the prediction equation, and the to-be-predicted value of each ozone concentration monitoring station is obtained by obtaining the equation Predict the O 3 -8h value of a day;
S5.建立新的超混沌系统：提出新的超混沌系统及其相应的复杂系统，以生成神经网络输入层和隐含层之间的连接权重值、隐含层中的神经元阈值，实现更好的预测结果；S5. Establish a new hyperchaotic system: Propose a new hyperchaotic system and its corresponding complex system to generate the connection weight value between the input layer and the hidden layer of the neural network, and the neuron threshold in the hidden layer to achieve more good prediction results;
S6.建立人工神经网络：网络分为三个部分：输入层，隐藏层和输出层，在网络中，输入层，隐藏层和输出层的神经元完全连接；S6. Establish an artificial neural network: the network is divided into three parts: an input layer, a hidden layer and an output layer, and in the network, the neurons of the input layer, the hidden layer and the output layer are fully connected;
S7.建立混沌人工神经网络：所述混沌人工神经网络是将新的超混沌系统与人工神经网络混合，且所述混沌人工神经网络是通过人工神经网络、BP和多元线性回归模型对臭氧浓度进行预测；S7. Set up a chaotic artificial neural network: the chaotic artificial neural network is to mix the new hyperchaotic system with the artificial neural network, and the chaotic artificial neural network is to carry out the ozone concentration through the artificial neural network, BP and multiple linear regression model predict;
S8.利用混沌人工神经网络进行长短期预报。S8. Using the chaotic artificial neural network for long-term and short-term forecasting.
优选的，所述S1中的污染物浓度数据包括二氧化硫SO2、二氧化氮NO2、O 3、O 3-8h、空气动力学直径为2.5微米或更小的细颗粒物PM2.5和空气动力学直径为10微米或更小的细颗粒物PM10。 Preferably, the pollutant concentration data in S1 includes sulfur dioxide SO2, nitrogen dioxide NO2, O 3 , O 3 -8h, fine particulate matter with an aerodynamic diameter of 2.5 microns or smaller PM2.5 and aerodynamic diameter PM10 is fine particulate matter 10 microns or smaller.
优选的，所述S2中收集、获取历史气象数据还包括获取每一臭氧浓度监测站点的任一历史时刻对应的点源排放特征，所述每一臭氧浓度监测站点的任一历史时刻对应的点源排放特征通过以下步骤获取：Preferably, the collection and acquisition of historical meteorological data in S2 also includes obtaining the point source emission characteristics corresponding to any historical moment of each ozone concentration monitoring station, and the point source emission characteristics corresponding to any historical moment of each ozone concentration monitoring station Source emission characteristics are obtained through the following steps:
S101.获取敏感区域，所述敏感区域包括与所述每一臭氧浓度监测站点所 在的城市接壤的所有城市；S101. obtain the sensitive area, the sensitive area includes all cities bordering on the city where each ozone concentration monitoring site is located;
S102.将所述敏感区域划分为若干个相同的矩形，获取所述每一矩形对应的点源排放子特征，根据所述每一矩形的点源排放子特征，获得所述点源排放特征。S102. Divide the sensitive area into several identical rectangles, obtain the point source emission sub-features corresponding to each rectangle, and obtain the point source emission characteristics according to the point source emission sub-features of each rectangle.
优选的，所述S4中通过拟合的方式的拟合的预测方程为：Preferably, the prediction equation of fitting by way of fitting in said S4 is:
Figure PCTCN2021129850-appb-000001
Figure PCTCN2021129850-appb-000001
其中，c为要预测的那一天的O 3-8h预测值，x1，x2，……，x6分别表示要预测那一天的NO2平均浓度、平均气温、平均气压、平均风速、PM2.5平均浓度的预测值，以及要预测那一天的昨日O 3-8h的预测值或实测值，a0，a1，……，a6是回归系数。 Among them, c is the predicted value of O 3 -8h on the day to be predicted, and x1, x2, ..., x6 respectively represent the average concentration of NO2, average temperature, average air pressure, average wind speed, and average concentration of PM2.5 on the day to be predicted The predicted value of , and the predicted or measured value of yesterday's O 3 -8h of the day to be predicted, a0, a1, ..., a6 are regression coefficients.
优选的，所述S5中的新的超混沌系统由以下公式描述：Preferably, the new hyper-chaotic system in the S5 is described by the following formula:
Figure PCTCN2021129850-appb-000002
Figure PCTCN2021129850-appb-000002
其中，a、b、c和d表示实常数，而ξ1、ξ2、ξ3和ξ4表示实变量，符号上的点表示相对于时间t的导数；where a, b, c, and d denote real constants, while ξ1, ξ2, ξ3, and ξ4 denote real variables, and dots on the symbols represent derivatives with respect to time t;
相应的超混沌复杂系统由下式给出：The corresponding hyperchaotic complex system is given by:
Figure PCTCN2021129850-appb-000003
Figure PCTCN2021129850-appb-000003
其中，η1＝u1+ju2，η2＝u3+ju4，η3＝u5，η4＝u6+ju7表示复变量，j＝-1 -0.5，变量上方的条形代表变量的复共轭； Wherein, η1=u1+ju2, η2=u3+ju4, η3=u5, η4=u6+ju7 represent complex variables, j= -1-0.5 , and the bar above the variable represents the complex conjugate of the variable;
系统的真实版本如下：The real version of the system is as follows:
Figure PCTCN2021129850-appb-000004
Figure PCTCN2021129850-appb-000004
优选的，所述S8利用混沌人工神经网络进行长短期预报的公式如下：Preferably, the formula of S8 using chaotic artificial neural network for long-term and short-term forecasting is as follows:
Figure PCTCN2021129850-appb-000005
Figure PCTCN2021129850-appb-000005
Figure PCTCN2021129850-appb-000006
Figure PCTCN2021129850-appb-000006
其中，n表示样本数量，xobs,i表示观测值，xpre,i表示预测值，Var表示样本方差。Among them, n represents the number of samples, xobs,i represents the observed value, xpre,i represents the predicted value, and Var represents the sample variance.
优选的，所述S7中的混沌人工神经网络采用CANN操作，所述CANN操作通常分为两个步骤：第一，设置参数，选择激活函数和操作模式，对于激活函数，使用sigmoid函数，并且将操作模式设置为回归；第二，选择70％-80％的样本数据进行训练，其余作为测试数据，权重和阈值在CANN中自动设置。Preferably, the chaotic artificial neural network in the S7 adopts CANN operation, and the CANN operation is usually divided into two steps: first, setting parameters, selecting an activation function and an operation mode, and using a sigmoid function for the activation function, and setting The operation mode is set to regression; second, 70%-80% of the sample data are selected for training, and the rest are used as test data, and the weights and thresholds are automatically set in CANN.
优选的，还包括臭氧污染预警，所述臭氧污染预警包括预警模块，所述预警模块用于对臭氧污染进行预警，确定臭氧污染的区间和污染等级。Preferably, an early warning of ozone pollution is also included, and the early warning of ozone pollution includes an early warning module, which is used for early warning of ozone pollution and determining the interval and pollution level of ozone pollution.
优选的，所述臭氧污染预警的方法包括以下步骤：Preferably, the method for ozone pollution early warning comprises the following steps:
S201.通过分析大气臭氧生成的外部条件，得到影响大气臭氧污染的敏感物理参量；所述物理参量包括气温、太阳总辐射辐照度、云量、颗粒物浓度以及氮氧化物浓度；S201. By analyzing the external conditions of atmospheric ozone generation, sensitive physical parameters affecting atmospheric ozone pollution are obtained; the physical parameters include air temperature, total solar radiation irradiance, cloud cover, particle concentration, and nitrogen oxide concentration;
S202.根据大气臭氧生成条件确定S201每一项所述物理参量的拐点值，并利用拐点值进行区间划分后对每一个区间进行赋值；S202. Determine the inflection point value of each physical parameter described in S201 according to atmospheric ozone generation conditions, and use the inflection point value to perform interval division and assign a value to each interval;
S203.获取臭氧浓度监测站点地区在一段时间内的所述物理参量的历史数据，利用Logistic回归模型，得到每一项物理参量的权值；S203. Obtain the historical data of the physical parameters in the ozone concentration monitoring site area for a period of time, and use the Logistic regression model to obtain the weight of each physical parameter;
S204.根据S202的区间赋值以及S203所述物理参量的权值，计算臭氧浓度影响因素结果值，并确定预警区间和相应的预警等级。S204. According to the interval assignment in S202 and the weight value of the physical parameter in S203, calculate the result value of the ozone concentration influencing factors, and determine the early warning interval and corresponding early warning level.
与现有技术相比，本发明的有益效果是：Compared with prior art, the beneficial effect of the present invention is:
(1)本发明采用CANN操作与其他神经网络的相似之处在于它不依赖于参数和输出之间的复杂关系，但它依赖于权值的不断变化，使参数与输出紧密关联，避免了繁琐的数学建模。(1) The present invention adopts the CANN operation similar to other neural networks in that it does not depend on the complex relationship between parameters and outputs, but it relies on constant changes in weights, so that parameters and outputs are closely related, avoiding cumbersome mathematical modeling.
(2)本发明具有较高的泛化性，能够在减少一些输入参数的同时保持较高的预测精度，在长期和短期的臭氧预测中都能取得良好的效果，且能准确高效地预测臭氧浓度，有利于对环境进行保护。(2) The present invention has high generalization, can maintain high prediction accuracy while reducing some input parameters, can achieve good results in long-term and short-term ozone prediction, and can accurately and efficiently predict ozone Concentration is conducive to the protection of the environment.
附图说明Description of drawings
图1为本发明的流程图。Fig. 1 is a flowchart of the present invention.
具体实施方式Detailed ways
下面将结合本发明实施例中的附图，对本发明实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本发明一部分实施例，而不是全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都属于本发明保护的范围。The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some, not all, embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the protection scope of the present invention.
实施例1Example 1
请参阅图1，本发明提供一种技术方案：基于人工智能的臭氧层预报算法，包括以下步骤：Please refer to Fig. 1, the present invention provides a kind of technical scheme: the ozone layer prediction algorithm based on artificial intelligence comprises the following steps:
S1.建立臭氧浓度监测站点：选取对周围环境和活性污染物的排放没有影响的位置作为臭氧浓度监测站点；S1. Establish an ozone concentration monitoring station: select a location that has no impact on the surrounding environment and the discharge of active pollutants as the ozone concentration monitoring station;
S2.收集、获取历史气象数据：获取每一臭氧浓度监测站点历史每天的日浓度数据、每一臭氧浓度监测站点历史每天预测的日浓度数据和所述每一臭氧浓度监测站点历史每天对应的参考结果，所述每一臭氧浓度监测站点历史每天对应的参考结果为每一臭氧浓度监测站点历史每天预测的日浓度数据与每一臭氧浓度监测站点历史每天的日浓度数据之差；S2. collect and obtain historical meteorological data: obtain the daily concentration data of each ozone concentration monitoring site history every day, the daily concentration data predicted every day of each ozone concentration monitoring site history and the reference corresponding to the history of each ozone concentration monitoring site every day As a result, the reference result corresponding to the history of each ozone concentration monitoring site every day is the difference between the daily concentration data of each ozone concentration monitoring site history and the daily concentration data of each ozone concentration monitoring site history every day;
S3.选取影响因子：选取昨日O 3-8h值、PM2.5日均浓度、NO2日均浓度、 预测的日平均气压、日平均风速和日平均气温作为影响因子； S3. Select influencing factors: select yesterday's O3-8h value, PM2.5 daily average concentration, NO2 daily average concentration, forecasted daily average air pressure, daily average wind speed and daily average temperature as the influencing factors;
S4.对一天O 3-8h值进行初步预测：通过拟合的方式，将气象参数的预测值和实测值代入预测方程中，通过求取方程来对每一臭氧浓度监测站点的要预测的那一天的O 3-8h值进行预测； S4. Preliminary prediction of the O 3 -8h value in one day: by fitting, the predicted value and the measured value of the meteorological parameters are substituted into the prediction equation, and the to-be-predicted value of each ozone concentration monitoring station is obtained by obtaining the equation Predict the O 3 -8h value of a day;
S5.建立新的超混沌系统：提出新的超混沌系统及其相应的复杂系统，以生成神经网络输入层和隐含层之间的连接权重值、隐含层中的神经元阈值，实现更好的预测结果；S5. Establish a new hyperchaotic system: Propose a new hyperchaotic system and its corresponding complex system to generate the connection weight value between the input layer and the hidden layer of the neural network, and the neuron threshold in the hidden layer to achieve more good prediction results;
S6.建立人工神经网络：网络分为三个部分：输入层，隐藏层和输出层，在网络中，输入层，隐藏层和输出层的神经元完全连接；S6. Establish an artificial neural network: the network is divided into three parts: an input layer, a hidden layer and an output layer, and in the network, the neurons of the input layer, the hidden layer and the output layer are fully connected;
S7.建立混沌人工神经网络：所述混沌人工神经网络是将新的超混沌系统与人工神经网络混合，且所述混沌人工神经网络是通过人工神经网络、BP和多元线性回归模型对臭氧浓度进行预测；S7. Set up a chaotic artificial neural network: the chaotic artificial neural network is to mix the new hyperchaotic system with the artificial neural network, and the chaotic artificial neural network is to carry out the ozone concentration through the artificial neural network, BP and multiple linear regression model predict;
S8.利用混沌人工神经网络进行长短期预报。S8. Using the chaotic artificial neural network for long-term and short-term forecasting.
本实施例中，优选的，所述S1中的污染物浓度数据包括二氧化硫SO2、二氧化氮NO2、O 3、O 3-8h、空气动力学直径为2.5微米或更小的细颗粒物PM2.5和空气动力学直径为10微米或更小的细颗粒物PM10。 In this embodiment, preferably, the pollutant concentration data in S1 includes sulfur dioxide SO2, nitrogen dioxide NO2, O3 , O3-8h , fine particulate matter PM2.5 with an aerodynamic diameter of 2.5 microns or less and fine particulate matter PM10 with an aerodynamic diameter of 10 microns or less.
本实施例中，优选的，所述S2中收集、获取历史气象数据还包括获取每一臭氧浓度监测站点的任一历史时刻对应的点源排放特征，所述每一臭氧浓度监测站点的任一历史时刻对应的点源排放特征通过以下步骤获取：In this embodiment, preferably, the collection and acquisition of historical meteorological data in S2 also includes obtaining the point source emission characteristics corresponding to any historical moment of each ozone concentration monitoring site, and any of the ozone concentration monitoring sites The point source emission characteristics corresponding to historical moments are obtained through the following steps:
S101.获取敏感区域，所述敏感区域包括与所述每一臭氧浓度监测站点所在的城市接壤的所有城市；S101. Obtain sensitive areas, the sensitive areas include all cities bordering on the city where each ozone concentration monitoring site is located;
S102.将所述敏感区域划分为若干个相同的矩形，获取所述每一矩形对应的点源排放子特征，根据所述每一矩形的点源排放子特征，获得所述点源排放特征。S102. Divide the sensitive area into several identical rectangles, obtain the point source emission sub-features corresponding to each rectangle, and obtain the point source emission characteristics according to the point source emission sub-features of each rectangle.
本实施例中，优选的，所述S4中通过拟合的方式的拟合的预测方程为：In this embodiment, preferably, the prediction equation of fitting by way of fitting in said S4 is:
Figure PCTCN2021129850-appb-000007
Figure PCTCN2021129850-appb-000007
其中，c为要预测的那一天的O 3-8h预测值，x1，x2，……，x6分别表示要预测那一天的NO2平均浓度、平均气温、平均气压、平均风速、PM2.5平均浓度的预测值，以及要预测那一天的昨日O 3-8h的预测值或实测值，a0，a1，……，a6是回归系数。 Among them, c is the predicted value of O 3 -8h on the day to be predicted, and x1, x2, ..., x6 respectively represent the average concentration of NO2, average temperature, average air pressure, average wind speed, and average concentration of PM2.5 on the day to be predicted The predicted value of , and the predicted or measured value of yesterday's O 3 -8h of the day to be predicted, a0, a1, ..., a6 are regression coefficients.
本实施例中，优选的，所述S5中的新的超混沌系统由以下公式描述：In this embodiment, preferably, the new hyper-chaotic system in the S5 is described by the following formula:
Figure PCTCN2021129850-appb-000008
Figure PCTCN2021129850-appb-000008
其中，a、b、c和d表示实常数，而ξ1、ξ2、ξ3和ξ4表示实变量，符号上的点表示相对于时间t的导数；where a, b, c, and d denote real constants, while ξ1, ξ2, ξ3, and ξ4 denote real variables, and dots on the symbols represent derivatives with respect to time t;
相应的超混沌复杂系统由下式给出：The corresponding hyperchaotic complex system is given by:
Figure PCTCN2021129850-appb-000009
Figure PCTCN2021129850-appb-000009
其中，η1＝u1+ju2，η2＝u3+ju4，η3＝u5，η4＝u6+ju7表示复变量，j＝-1 -0.5，变量上方的条形代表变量的复共轭； Wherein, η1=u1+ju2, η2=u3+ju4, η3=u5, η4=u6+ju7 represent complex variables, j= -1-0.5 , and the bar above the variable represents the complex conjugate of the variable;
系统的真实版本如下：The real version of the system is as follows:
Figure PCTCN2021129850-appb-000010
Figure PCTCN2021129850-appb-000010
本实施例中，优选的，所述S8利用混沌人工神经网络进行长短期预报的公式如下：In this embodiment, preferably, the formula of S8 using chaotic artificial neural network for long-term and short-term forecasting is as follows:
Figure PCTCN2021129850-appb-000011
Figure PCTCN2021129850-appb-000011
Figure PCTCN2021129850-appb-000012
Figure PCTCN2021129850-appb-000012
其中，n表示样本数量，xobs,i表示观测值，xpre,i表示预测值，Var表示样本方差。Among them, n represents the number of samples, xobs,i represents the observed value, xpre,i represents the predicted value, and Var represents the sample variance.
本实施例中，优选的，所述S7中的混沌人工神经网络采用CANN操作，所述CANN操作通常分为两个步骤：第一，设置参数，选择激活函数和操作模式，对于激活函数，使用sigmoid函数，并且将操作模式设置为回归；第二，选择70％-80％的样本数据进行训练，其余作为测试数据，权重和阈值在CANN中自动设置。In this embodiment, preferably, the chaotic artificial neural network in S7 adopts CANN operation, and the CANN operation is usually divided into two steps: the first, setting parameters, selecting activation function and operation mode, for activation function, use sigmoid function, and set the operation mode to regression; second, select 70%-80% of the sample data for training, and the rest as test data, and the weights and thresholds are automatically set in CANN.
本实施例中，优选的，还包括臭氧污染预警，所述臭氧污染预警包括预警模块，所述预警模块用于对臭氧污染进行预警，确定臭氧污染的区间和污染等级。In this embodiment, preferably, an early warning of ozone pollution is also included, and the early warning of ozone pollution includes an early warning module, and the early warning module is used for early warning of ozone pollution, and determining the interval and pollution level of ozone pollution.
本实施例中，优选的，所述臭氧污染预警的方法包括以下步骤：In the present embodiment, preferably, the method for ozone pollution early warning comprises the following steps:
S201.通过分析大气臭氧生成的外部条件，得到影响大气臭氧污染的敏感物理参量；所述物理参量包括气温、太阳总辐射辐照度、云量、颗粒物浓度以及氮氧化物浓度；S201. By analyzing the external conditions of atmospheric ozone generation, sensitive physical parameters affecting atmospheric ozone pollution are obtained; the physical parameters include air temperature, total solar radiation irradiance, cloud cover, particle concentration, and nitrogen oxide concentration;
S202.根据大气臭氧生成条件确定S201每一项所述物理参量的拐点值，并利用拐点值进行区间划分后对每一个区间进行赋值；S202. Determine the inflection point value of each physical parameter described in S201 according to atmospheric ozone generation conditions, and use the inflection point value to perform interval division and assign a value to each interval;
S203.获取臭氧浓度监测站点地区在一段时间内的所述物理参量的历史数据，利用Logistic回归模型，得到每一项物理参量的权值；S203. Obtain the historical data of the physical parameters in the ozone concentration monitoring site area for a period of time, and use the Logistic regression model to obtain the weight of each physical parameter;
S204.根据S202的区间赋值以及S203所述物理参量的权值，计算臭氧浓度影响因素结果值，并确定预警区间和相应的预警等级。S204. According to the interval assignment in S202 and the weight value of the physical parameter in S203, calculate the result value of the ozone concentration influencing factors, and determine the early warning interval and corresponding early warning level.
本发明的原理及优点：本发明采用CANN操作与其他神经网络的相似之处在于它不依赖于参数和输出之间的复杂关系，但它依赖于权值的不断变化，使参数与输出紧密关联，避免了繁琐的数学建模；本发明具有较高的泛化性，能够在减少一些输入参数的同时保持较高的预测精度，在长期和短期的臭氧预测中都能取得良好的效果，且能准确高效地预测臭氧浓度，有利于对环境进行保护。Principles and advantages of the present invention: the present invention adopts CANN operation similar to other neural networks in that it does not depend on the complex relationship between parameters and outputs, but it relies on the constant change of weights, so that parameters and outputs are closely related , avoiding cumbersome mathematical modeling; the present invention has high generalization, can reduce some input parameters while maintaining high prediction accuracy, and can achieve good results in long-term and short-term ozone prediction, and The ozone concentration can be accurately and efficiently predicted, which is beneficial to the protection of the environment.
尽管已经示出和描述了本发明的实施例，对于本领域的普通技术人员而 言，可以理解在不脱离本发明的原理和精神的情况下可以对这些实施例进行多种变化、修改、替换和变型，本发明的范围由所附权利要求及其等同物限定。Although the embodiments of the present invention have been shown and described, those skilled in the art can understand that various changes, modifications and substitutions can be made to these embodiments without departing from the principle and spirit of the present invention. and modifications, the scope of the invention is defined by the appended claims and their equivalents.

Claims (9)

  1. 基于人工智能的臭氧层预报算法，其特征在于，包括以下步骤：The ozone layer prediction algorithm based on artificial intelligence is characterized in that, comprising the following steps:
    S1.建立臭氧浓度监测站点：选取对周围环境和活性污染物的排放没有影响的位置作为臭氧浓度监测站点；S1. Establish an ozone concentration monitoring station: select a location that has no impact on the surrounding environment and the discharge of active pollutants as the ozone concentration monitoring station;
    S2.收集、获取历史气象数据：获取每一臭氧浓度监测站点历史每天的日浓度数据、每一臭氧浓度监测站点历史每天预测的日浓度数据和所述每一臭氧浓度监测站点历史每天对应的参考结果，所述每一臭氧浓度监测站点历史每天对应的参考结果为每一臭氧浓度监测站点历史每天预测的日浓度数据与每一臭氧浓度监测站点历史每天的日浓度数据之差；S2. collect and obtain historical meteorological data: obtain the daily concentration data of each ozone concentration monitoring site history every day, the daily concentration data predicted every day of each ozone concentration monitoring site history and the reference corresponding to the history of each ozone concentration monitoring site every day As a result, the reference result corresponding to the history of each ozone concentration monitoring site every day is the difference between the daily concentration data of each ozone concentration monitoring site history and the daily concentration data of each ozone concentration monitoring site history every day;
    S3.选取影响因子：选取昨日O 3-8h值、PM2.5日均浓度、NO2日均浓度、预测的日平均气压、日平均风速和日平均气温作为影响因子； S3. Select influencing factors: select yesterday's O 3 -8h value, PM2.5 daily average concentration, NO2 daily average concentration, forecasted daily average air pressure, daily average wind speed and daily average temperature as influencing factors;
    S4.对一天O 3-8h值进行初步预测：通过拟合的方式，将气象参数的预测值和实测值代入预测方程中，通过求取方程来对每一臭氧浓度监测站点的要预测的那一天的O 3-8h值进行预测； S4. Preliminary prediction of the O 3 -8h value in one day: by fitting, the predicted value and the measured value of the meteorological parameters are substituted into the prediction equation, and the to-be-predicted value of each ozone concentration monitoring station is obtained by obtaining the equation Predict the O 3 -8h value of a day;
    S5.建立新的超混沌系统：提出新的超混沌系统及其相应的复杂系统，以生成神经网络输入层和隐含层之间的连接权重值、隐含层中的神经元阈值，实现更好的预测结果；S5. Establish a new hyperchaotic system: Propose a new hyperchaotic system and its corresponding complex system to generate the connection weight value between the input layer and the hidden layer of the neural network, and the neuron threshold in the hidden layer to achieve more good prediction results;
    S6.建立人工神经网络：网络分为三个部分：输入层，隐藏层和输出层，在网络中，输入层，隐藏层和输出层的神经元完全连接；S6. Establish an artificial neural network: the network is divided into three parts: an input layer, a hidden layer and an output layer, and in the network, the neurons of the input layer, the hidden layer and the output layer are fully connected;
    S7.建立混沌人工神经网络：所述混沌人工神经网络是将新的超混沌系统与人工神经网络混合，且所述混沌人工神经网络是通过人工神经网络、BP和多元线性回归模型对臭氧浓度进行预测；S7. Set up a chaotic artificial neural network: the chaotic artificial neural network is to mix the new hyperchaotic system with the artificial neural network, and the chaotic artificial neural network is to carry out the ozone concentration through the artificial neural network, BP and multiple linear regression model predict;
    S8.利用混沌人工神经网络进行长短期预报。S8. Using the chaotic artificial neural network for long-term and short-term forecasting.
  2. 根据权利要求1所述的基于人工智能的臭氧层预报算法，其特征在于：所述S1中的污染物浓度数据包括二氧化硫SO2、二氧化氮NO2、O 3、O 3-8h、空气动力学直径为2.5微米或更小的细颗粒物PM2.5和空气动力学直径为10 微米或更小的细颗粒物PM10。 The ozone layer prediction algorithm based on artificial intelligence according to claim 1, is characterized in that: the pollutant concentration data in the described S1 comprises sulfur dioxide SO , nitrogen dioxide NO , O 3 , O 3 -8h, an aerodynamic diameter of PM2.5, which is fine particulate matter of 2.5 microns or less, and PM10, which has an aerodynamic diameter of 10 microns or less.
  3. 根据权利要求1所述的基于人工智能的臭氧层预报算法，其特征在于：所述S2中收集、获取历史气象数据还包括获取每一臭氧浓度监测站点的任一历史时刻对应的点源排放特征，所述每一臭氧浓度监测站点的任一历史时刻对应的点源排放特征通过以下步骤获取：The ozone layer forecasting algorithm based on artificial intelligence according to claim 1, is characterized in that: collecting in the described S2, obtaining historical meteorological data also includes obtaining the point source emission characteristics corresponding to any historical moment of each ozone concentration monitoring site, The point source emission characteristics corresponding to any historical moment of each ozone concentration monitoring site are obtained through the following steps:
    S101.获取敏感区域，所述敏感区域包括与所述每一臭氧浓度监测站点所在的城市接壤的所有城市；S101. Obtain sensitive areas, the sensitive areas include all cities bordering on the city where each ozone concentration monitoring site is located;
    S102.将所述敏感区域划分为若干个相同的矩形，获取所述每一矩形对应的点源排放子特征，根据所述每一矩形的点源排放子特征，获得所述点源排放特征。S102. Divide the sensitive area into several identical rectangles, obtain the point source emission sub-features corresponding to each rectangle, and obtain the point source emission characteristics according to the point source emission sub-features of each rectangle.
  4. 根据权利要求1所述的基于人工智能的臭氧层预报算法，其特征在于：所述S4中通过拟合的方式的拟合的预测方程为：The artificial intelligence-based ozone layer prediction algorithm according to claim 1, is characterized in that: the predictive equation by the fitting of the mode of fitting among the described S4 is:
    Figure PCTCN2021129850-appb-100001
    Figure PCTCN2021129850-appb-100001
    其中，c为要预测的那一天的O 3-8h预测值，x1，x2，……，x6分别表示要预测那一天的NO2平均浓度、平均气温、平均气压、平均风速、PM2.5平均浓度的预测值，以及要预测那一天的昨日O 3-8h的预测值或实测值，a0，a1，……，a6是回归系数。 Among them, c is the predicted value of O 3 -8h on the day to be predicted, and x1, x2, ..., x6 respectively represent the average concentration of NO2, average temperature, average air pressure, average wind speed, and average concentration of PM2.5 on the day to be predicted The predicted value of , and the predicted or measured value of yesterday's O 3 -8h of the day to be predicted, a0, a1, ..., a6 are regression coefficients.
  5. 根据权利要求1所述的基于人工智能的臭氧层预报算法，其特征在于：所述S5中的新的超混沌系统由以下公式描述：The artificial intelligence-based ozone layer forecasting algorithm according to claim 1, is characterized in that: the new ultra-chaotic system among the described S5 is described by following formula:
    Figure PCTCN2021129850-appb-100002
    Figure PCTCN2021129850-appb-100002
    其中，a、b、c和d表示实常数，而ξ1、ξ2、ξ3和ξ4表示实变量，符号上的点表示相对于时间t的导数；where a, b, c, and d denote real constants, while ξ1, ξ2, ξ3, and ξ4 denote real variables, and dots on the symbols represent derivatives with respect to time t;
    相应的超混沌复杂系统由下式给出：The corresponding hyperchaotic complex system is given by:
    Figure PCTCN2021129850-appb-100003
    Figure PCTCN2021129850-appb-100003
    其中，η1＝u1+ju2，η2＝u3+ju4，η3＝u5，η4＝u6+ju7表示复变量，j＝-1 -0.5，变量上方的条形代表变量的复共轭； Wherein, η1=u1+ju2, η2=u3+ju4, η3=u5, η4=u6+ju7 represent complex variables, j= -1-0.5 , and the bar above the variable represents the complex conjugate of the variable;
    系统的真实版本如下：The real version of the system is as follows:
    Figure PCTCN2021129850-appb-100004
    Figure PCTCN2021129850-appb-100004
  6. 根据权利要求1所述的基于人工智能的臭氧层预报算法，其特征在于：所述S8利用混沌人工神经网络进行长短期预报的公式如下：The ozone layer forecasting algorithm based on artificial intelligence according to claim 1, is characterized in that: the formula that described S8 utilizes chaotic artificial neural network to carry out long-term and short-term forecasting is as follows:
    Figure PCTCN2021129850-appb-100005
    Figure PCTCN2021129850-appb-100005
    Figure PCTCN2021129850-appb-100006
    Figure PCTCN2021129850-appb-100006
    其中，n表示样本数量，xobs,i表示观测值，xpre,i表示预测值，Var表示样本方差。Among them, n represents the number of samples, xobs,i represents the observed value, xpre,i represents the predicted value, and Var represents the sample variance.
  7. 根据权利要求1所述的基于人工智能的臭氧层预报算法，其特征在于：所述S7中的混沌人工神经网络采用CANN操作，所述CANN操作通常分为两个步骤：第一，设置参数，选择激活函数和操作模式，对于激活函数，使用 sigmoid函数，并且将操作模式设置为回归；第二，选择70％-80％的样本数据进行训练，其余作为测试数据，权重和阈值在CANN中自动设置。The ozone layer forecasting algorithm based on artificial intelligence according to claim 1, is characterized in that: the chaotic artificial neural network among the described S7 adopts CANN operation, and described CANN operation is usually divided into two steps: the first, set parameter, select Activation function and operation mode, for the activation function, use the sigmoid function, and set the operation mode to regression; second, select 70%-80% of the sample data for training, and the rest as test data, weights and thresholds are automatically set in CANN .
  8. 根据权利要求1所述的基于人工智能的臭氧层预报算法，其特征在于：还包括臭氧污染预警，所述臭氧污染预警包括预警模块，所述预警模块用于对臭氧污染进行预警，确定臭氧污染的区间和污染等级。The ozone layer forecasting algorithm based on artificial intelligence according to claim 1, is characterized in that: also comprises ozone pollution early warning, and described ozone pollution early warning comprises early warning module, and described early warning module is used for early warning to ozone pollution, determines ozone pollution range and pollution level.
  9. 根据权利要求8所述的基于人工智能的臭氧层预报算法，其特征在于：所述臭氧污染预警的方法包括以下步骤：The artificial intelligence-based ozone layer forecasting algorithm according to claim 8, is characterized in that: the method for ozone pollution early warning comprises the following steps:
    S201.通过分析大气臭氧生成的外部条件，得到影响大气臭氧污染的敏感物理参量；所述物理参量包括气温、太阳总辐射辐照度、云量、颗粒物浓度以及氮氧化物浓度；S201. By analyzing the external conditions of atmospheric ozone generation, sensitive physical parameters affecting atmospheric ozone pollution are obtained; the physical parameters include air temperature, total solar radiation irradiance, cloud cover, particle concentration, and nitrogen oxide concentration;
    S202.根据大气臭氧生成条件确定S201每一项所述物理参量的拐点值，并利用拐点值进行区间划分后对每一个区间进行赋值；S202. Determine the inflection point value of each physical parameter described in S201 according to atmospheric ozone generation conditions, and use the inflection point value to perform interval division and assign a value to each interval;
    S203.获取臭氧浓度监测站点地区在一段时间内的所述物理参量的历史数据，利用Logistic回归模型，得到每一项物理参量的权值；S203. Obtain the historical data of the physical parameters in the ozone concentration monitoring site area for a period of time, and use the Logistic regression model to obtain the weight of each physical parameter;
    S204.根据S202的区间赋值以及S203所述物理参量的权值，计算臭氧浓度影响因素结果值，并确定预警区间和相应的预警等级。S204. According to the interval assignment in S202 and the weight value of the physical parameter in S203, calculate the result value of the ozone concentration influencing factors, and determine the early warning interval and corresponding early warning level.
PCT/CN2021/129850 2021-10-09 2021-11-10 Ozone layer prediction algorithm based on artificial intelligence WO2023056682A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202111175211.5 2021-10-09
CN202111175211.5A CN113901714B (en) 2021-10-09 2021-10-09 Artificial intelligence-based ozone layer forecasting algorithm

Publications (1)

Publication Number Publication Date
WO2023056682A1 true WO2023056682A1 (en) 2023-04-13

Family

ID=79190607

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/129850 WO2023056682A1 (en) 2021-10-09 2021-11-10 Ozone layer prediction algorithm based on artificial intelligence

Country Status (2)

Country Link
CN (1) CN113901714B (en)
WO (1) WO2023056682A1 (en)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116776745A (en) * 2023-08-18 2023-09-19 南昌云宜然科技有限公司 Method and system for predicting pollutant concentration and track movement based on edge calculation
CN117031582A (en) * 2023-06-27 2023-11-10 华南理工大学 Ozone hour concentration forecasting method based on recursive space-time learning and simulation monitoring fusion
CN117113089A (en) * 2023-10-16 2023-11-24 北京英视睿达科技股份有限公司 Methane data complement method, device, equipment and medium based on carbon monoxide
CN117265251A (en) * 2023-09-20 2023-12-22 索罗曼(广州)新材料有限公司 Titanium flat bar oxygen content online monitoring system and method thereof
CN117879789A (en) * 2024-03-13 2024-04-12 数盾信息科技股份有限公司 Satellite communication message data transmission method based on high-speed encryption
CN117879789B (en) * 2024-03-13 2024-05-14 数盾信息科技股份有限公司 Satellite communication message data transmission method based on high-speed encryption

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117291319B (en) * 2023-11-27 2024-02-20 新禾数字科技(无锡)有限公司 O based on machine learning 3 Residual prediction method

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106019409A (en) * 2016-05-11 2016-10-12 北京市环境保护监测中心 Partition prediction method and system for ozone concentration
CN109257159A (en) * 2018-11-07 2019-01-22 中南大学 The building method of novel higher-dimension hyperchaotic system

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112684118B (en) * 2020-12-31 2022-12-20 南京信息工程大学 Convenient early warning method for atmospheric ozone pollution

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106019409A (en) * 2016-05-11 2016-10-12 北京市环境保护监测中心 Partition prediction method and system for ozone concentration
CN109257159A (en) * 2018-11-07 2019-01-22 中南大学 The building method of novel higher-dimension hyperchaotic system

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
YONG WANG, YANG JIN, WANG YING: "Image Encryption Algorithm Based on Improved Henon Hyperchaotic System Combined with AES Algorithm", COMPUTER ENGINEERING AND APPLICATIONS, HUABEI JISUAN JISHU YANJIUSUO, CN, vol. 55, no. 22, 26 November 2018 (2018-11-26), CN , pages 180 - 186, XP093055852, ISSN: 1002-8331, DOI: 10.3778/j.issn.1002-8331.1807-0230 *

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117031582A (en) * 2023-06-27 2023-11-10 华南理工大学 Ozone hour concentration forecasting method based on recursive space-time learning and simulation monitoring fusion
CN116776745A (en) * 2023-08-18 2023-09-19 南昌云宜然科技有限公司 Method and system for predicting pollutant concentration and track movement based on edge calculation
CN116776745B (en) * 2023-08-18 2023-10-24 南昌云宜然科技有限公司 Method and system for predicting pollutant concentration and track movement based on edge calculation
CN117265251A (en) * 2023-09-20 2023-12-22 索罗曼(广州)新材料有限公司 Titanium flat bar oxygen content online monitoring system and method thereof
CN117265251B (en) * 2023-09-20 2024-04-09 索罗曼(广州)新材料有限公司 Titanium flat bar oxygen content online monitoring system and method thereof
CN117113089A (en) * 2023-10-16 2023-11-24 北京英视睿达科技股份有限公司 Methane data complement method, device, equipment and medium based on carbon monoxide
CN117113089B (en) * 2023-10-16 2024-01-23 北京英视睿达科技股份有限公司 Methane data complement method, device, equipment and medium based on carbon monoxide
CN117879789A (en) * 2024-03-13 2024-04-12 数盾信息科技股份有限公司 Satellite communication message data transmission method based on high-speed encryption
CN117879789B (en) * 2024-03-13 2024-05-14 数盾信息科技股份有限公司 Satellite communication message data transmission method based on high-speed encryption

Also Published As

Publication number Publication date
CN113901714A (en) 2022-01-07
CN113901714B (en) 2022-08-23

Similar Documents

Publication Publication Date Title
WO2023056682A1 (en) Ozone layer prediction algorithm based on artificial intelligence
Wang et al. Regional prediction of ground-level ozone using a hybrid sequence-to-sequence deep learning approach
Xu et al. Research and application of a hybrid model based on dynamic fuzzy synthetic evaluation for establishing air quality forecasting and early warning system: A case study in China
CN106651036A (en) Air quality forecasting system
Yao et al. A support vector machine approach to estimate global solar radiation with the influence of fog and haze
CN109002915B (en) Photovoltaic power station short-term power prediction method based on Kmeans-GRA-Elman model
Mahanta et al. Urban air quality prediction using regression analysis
Cai et al. A combined filtering strategy for short term and long term wind speed prediction with improved accuracy
Yousif et al. Analysis and forecasting of weather conditions in Oman for renewable energy applications
Alassery et al. An artificial intelligence-based solar radiation prophesy model for green energy utilization in energy management system
Wang et al. Comparative study of machine learning approaches for predicting short-term photovoltaic power output based on weather type classification
CN111428942B (en) Line icing thickness prediction method for extracting micro-terrain factors based on variable grid technology
Kothandaraman et al. Intelligent forecasting of air quality and pollution prediction using machine learning
Reja et al. A review of the evaluation of urban wind resources: Challenges and perspectives
CN108802857A (en) A kind of Meteorology Forecast System based on meteorological data
CN109085296B (en) Judge that ozone generates sensibility and precursor control method, device, storage medium, terminal
CN108446783A (en) A kind of prediction of new fan operation power and monitoring method
CN113537515A (en) PM2.5 prediction method, system, device and storage medium
CN116050666B (en) Photovoltaic power generation power prediction method for irradiation characteristic clustering
Li et al. Multi-model ensemble forecast method of PM2. 5 concentration based on wavelet neural networks
CN116484998A (en) Distributed photovoltaic power station power prediction method and system based on meteorological similar day
Faruq et al. River water level forecasting for flood warning system using deep learning long short-term memory network
CN107578122A (en) A kind of Load Forecasting and system based on sendible temperature and date type
Valsaraj et al. Machine learning-based simplified methods using shorter wind measuring masts for the time ahead wind forecasting at higher altitude for wind energy applications
CN113112085A (en) New energy station power generation load prediction method based on BP neural network

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21959770

Country of ref document: EP

Kind code of ref document: A1