CN213813502U - Near-surface ozone concentration prediction system based on machine learning - Google Patents
Near-surface ozone concentration prediction system based on machine learning Download PDFInfo
- Publication number
- CN213813502U CN213813502U CN202022597854.6U CN202022597854U CN213813502U CN 213813502 U CN213813502 U CN 213813502U CN 202022597854 U CN202022597854 U CN 202022597854U CN 213813502 U CN213813502 U CN 213813502U
- Authority
- CN
- China
- Prior art keywords
- acquisition terminal
- data acquisition
- data
- display
- keyboard
- Prior art date
- Legal status (The legal status 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 status listed.)
- Active
Links
Images
Abstract
The utility model discloses a nearly ground ozone concentration prediction system based on machine learning, which comprises an outer shell, shell leading flank top corner is provided with the pilot lamp, pilot lamp one side is provided with the display, display one side is provided with the keyboard, shell leading flank bottom one side is provided with power connection, power connection external battery, battery output end is provided with the chip, pilot lamp, display, keyboard and chip all are connected with the battery. The utility model discloses a remain and influenced great characteristic vector to ozone prediction to avoided subjective screening to cause the rejection of effective characteristic vector, utilized data acquisition terminal, characteristic screening treater, central model training ware and data calculation unit to mutually support, thereby reachd the prediction result of ozone concentration through effective characteristic vector.
Description
Technical Field
The utility model belongs to the technical field of gas concentration prediction, in particular to near-ground ozone concentration prediction system based on machine learning.
Background
With the aggravation of ozone pollution and the deepening of understanding of ozone harm, it is important to provide accurate ozone forecast. The forecasting method of the atmospheric composition can be mainly summarized into a statistical method, numerical forecasting and other advanced technical methods, and the forecasting method based on experience and statistics in the past is simple and convenient to calculate, but cannot meet the requirements in the aspects of theory, timeliness and the like. At present, due to various reasons, such as that a list of emission sources is different from an actual situation, especially the mode cannot consider sudden pollutant emission, the physical and chemical mechanisms of the mode are very complex, parameterization of a physical process still needs to be continuously completed, and the like, a numerical mode has a large error in forecasting atmospheric components such as particles, ozone, and the like. The machine learning method can effectively capture the hidden nonlinear characteristics in atmospheric composition changes, the atmospheric composition is predicted by constructing a model through characteristic variables, compared with numerical prediction, the machine learning method has the advantages of less input data, higher calculation efficiency and ideal prediction effect, and therefore, in recent years, the machine learning method is researched and applied in the fields of weather and atmospheric composition prediction.
The machine learning method needs a large amount of observation and simulation data for training, the selection of input characteristic quantity has an important influence on whether the ozone concentration can be accurately and efficiently predicted, the machine learning model for atmospheric composition prediction usually considers ground atmospheric composition and meteorological element data, does not consider boundary layer and high-altitude meteorological elements, and the characteristic quantity with important weight on ozone prediction can be ignored based on the subjective screening method, so that the ozone prediction result can be influenced.
Therefore, it is necessary to invent a near-surface ozone concentration prediction system based on machine learning to solve the above problems.
SUMMERY OF THE UTILITY MODEL
To the above problem, the utility model provides a near-ground ozone concentration prediction system based on machine learning to solve the problem of proposing among the above-mentioned background art.
In order to achieve the above object, the utility model provides a following technical scheme: a near-ground ozone concentration prediction system based on machine learning comprises a shell, wherein an indicator lamp is arranged at the top corner of the front side face of the shell, a display is arranged on one side of the indicator lamp, a keyboard is arranged on one side of the display, a power connector is arranged on one side of the bottom of the front side face of the shell, the power connector is externally connected with a battery, and the battery is connected with a chip;
the chip comprises a data acquisition terminal, a feature screening processor, a central model trainer and a data calculation unit, wherein a data storage card is arranged at the output end of the data acquisition terminal, the keyboard is in one-way connection with the data storage card, the data storage card is in one-way connection with the feature screening processor, the feature screening processor is in one-way connection with the central model trainer, the central model trainer is in one-way connection with the data calculation unit, and the data calculation unit is in one-way connection with the display;
the base is arranged on the rear side face of the shell and connected with the shell in a clamping mode.
Furthermore, the number of the indicator lamps is three, and the indicator lamps are respectively connected with the display, the keyboard and the chip from top to bottom.
Further, the data acquisition terminal comprises an atmospheric composition observation data acquisition terminal, a ground meteorological observation data acquisition terminal and a weather numerical mode data acquisition terminal;
the system comprises an atmospheric composition observation data acquisition terminal, a ground meteorological observation data acquisition terminal and a weather numerical mode data acquisition terminal, wherein the atmospheric composition observation data acquisition terminal, the ground meteorological observation data acquisition terminal and the weather numerical mode data acquisition terminal are respectively connected with a data storage card, the atmospheric composition observation data acquisition terminal is used for acquiring atmospheric composition observation data of past ten-day scales, the ground meteorological observation data acquisition terminal is used for acquiring observation data of past ten-day-scale meteorological automatic stations, and the weather numerical mode data acquisition terminal is used for acquiring high-low air image element data of past ten days.
Further, the data storage card is used for storing the original data acquired by the data acquisition terminal and storing the original data in a classified manner according to the input date and the predicted duration of the keyboard;
the characteristic screening processor is used for reading the original data of the data storage card and optimizing the characteristic quantity with larger influence on ozone prediction in an objective selection mode;
the central model trainer is used for carrying out model training on the feature quantity optimized by the feature screening processor through a training model;
and the data calculation unit is used for calculating the ozone concentration according to the training model and the characteristic quantity and sending the calculation result to the display.
Furthermore, the chip and the data storage card are both arranged inside the shell, the indicator light, the display, the keyboard and the chip are all connected with the battery through power connectors, and the keyboard is composed of ten digital keys.
The utility model discloses a technological effect and advantage:
1. the utility model discloses a remain and influenced great characteristic vector to ozone prediction to avoided subjective screening to cause the rejection of effective characteristic vector, utilized data acquisition terminal, characteristic screening treater, central model training ware and data calculation unit to mutually support, thereby reachd the prediction result of ozone concentration through effective characteristic vector.
2. The utility model discloses an pilot lamp shows the working condition of display, keyboard and chip, lights when the pilot lamp, explains that display, keyboard and chip work are normal, if the pilot lamp extinguishes, maintains display, keyboard and chip and pilot lamp, guarantees that whole ozone concentration prediction system is in normal process state all the time, avoids equipment damage to lead to ozone concentration prediction error.
Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by the practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without creative efforts.
Fig. 1 shows a schematic structural view of the front side of a housing according to an embodiment of the present invention;
fig. 2 is a schematic diagram illustrating a connection structure of a battery and a chip according to an embodiment of the present invention;
fig. 3 shows a schematic view of a base structure of an embodiment of the present invention;
fig. 4 shows a schematic flow chart of the system inside chip according to the embodiment of the present invention;
in the figure: 1. a housing; 11. an indicator light; 12. a display; 13. a keyboard; 14. a power supply connector; 2. a battery; 3. a chip; 31. a data acquisition terminal; 311. an atmospheric composition observation data acquisition terminal; 312. a ground meteorological observation data acquisition terminal; 313. a weather numerical mode data acquisition terminal; 32. a feature screening processor; 33. a central model trainer; 34. a data calculation unit; 4. a data storage card; 5. a base.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the embodiments of the present invention are clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. Based on the embodiments in the present invention, all other embodiments obtained by a person skilled in the art without creative work belong to the protection scope of the present invention.
The utility model provides a near-ground ozone concentration prediction system based on machine learning, as shown in fig. 1-2, including shell 1, 1 top corner of shell is provided with pilot lamp 11, pilot lamp 11 one side is provided with display 12, display 12 one side is provided with keyboard 13, keyboard 13 comprises ten number keys, makes things convenient for keyboard 13 to input different figures according to the number key of difference, and then can make up into arbitrary figure and be listed as. A power connector 14 is arranged on one side of the bottom of the shell 1, the power connector 14 is externally connected with a battery 2, the battery 2 is connected with the chip 3, the chip 3 and the data storage card 4 are both positioned in the shell 1, the chip 3 and the data storage card 4 are protected through the shell 1, the chip 3 and the data storage card 4 are prevented from being influenced by the external environment, and the indicator light 11, the display 12, the keyboard 13 and the chip 3 are all connected with the battery 2 through the power connector 14; the number of the indicator lamps 11 is three, the indicator lamps 11 are respectively connected with the display 12, the keyboard 13 and the chip 3 from top to bottom, after the power connector 14 of the whole ozone concentration prediction system is connected with the battery 2, the indicator lamps 11 are lightened, and the three indicator lamps 11 are respectively connected in the working circuits of the display 12, the keyboard 13 and the chip 3 to show the working conditions of the display 12, the keyboard 13 and the chip 3. Illustratively, when the middle indicator light 11 is turned off and the other two indicator lights 11 are still bright, the keyboard 13 is damaged, the keyboard 13 or the middle indicator light 11 can be maintained, the whole ozone concentration prediction system is always in a normal process state, and the ozone concentration prediction error caused by equipment damage is avoided. After the power connector 14 at the bottom of the front side surface of the shell 1 is connected with the battery 2, the battery 2 supplies power for the whole ozone concentration prediction system, the indicator lamp 11 and the display 12 are lighted, and the display 12 displays the prediction data of the ozone layer concentration collected by the chip 3. The rear side face of the shell 1 is provided with a base 5, and the base 5 is connected with the shell 1 in a clamping mode. Place shell 1 at 5 tops of base, the kelly of 5 both sides of base can support shell 1, avoids shell 1 direct contact to ground.
Exemplarily, the power connector 14 is a three-hole connector, the power connector 14 is correspondingly clamped with a three-hole socket on the power supply 2, the power supply 2 supplies power to the indicator light 11, the display 12, the keyboard 13 and the chip 3 through the power connector 14, the principle is the same as that of a power supply device in the prior art, wherein the indicator light 11 is set as a diode, the type of the diode is HER308, the chip 3 is set as a single chip microcomputer, the type of the single chip microcomputer is SMC62, and the indicator light 11, the display 12, the keyboard 13 and the chip 3 are all common products in the prior art and are known to those skilled in the art.
The chip 3 comprises a data acquisition terminal 31, a feature screening processor 32, a central model trainer 33 and a data calculation unit 34, wherein the data storage card 4 is arranged at the output end of the data acquisition terminal 31, the keyboard 13 is in one-way connection with the data storage card 4, the data storage card 4 is in one-way connection with the feature screening processor 32, the feature screening processor 32 is in one-way connection with the central model trainer 33, the central model trainer 33 is in one-way connection with the data calculation unit 34, and the data calculation unit 34 is in one-way connection with the display 12; illustratively, the data collection terminal 31 is an aggregate of sensors, wherein the atmospheric composition observation data collection terminal 311 is provided as a gas sensor of the type IAQ-CORE C, the ground weather observation data collection terminal 312 is provided as an image collector of the type CASIO-DT-X10, the weather value pattern data collection terminal 313 is provided as a data collector of the type YL6800, and the data memory card 4 is provided as a memory of the type XBOX 360. Wherein, data acquisition terminal 31 and 4 electric connection of data storage card, data storage card 4 are used for the original data signal of storage data acquisition terminal 31 collection to save the signal, its storage mode is the same with the storage principle among the prior art, for the known technique of the technical personnel of the field of the subject, the utility model discloses mainly through line connection's mode for original data transmits between each part, connects each part through electric connection's mode, guarantees data transmission's agility.
The data acquisition terminal 31 comprises an atmospheric composition observation data acquisition terminal 311, a ground meteorological observation data acquisition terminal 312 and a weather numerical mode data acquisition terminal 313; the atmospheric composition observation data acquisition terminal 311, the ground meteorological observation data acquisition terminal 312 and the weather numerical mode data acquisition terminal 313 are respectively connected with the data storage card 4, the atmospheric composition observation data acquisition terminal 311 is used for acquiring atmospheric composition observation data of past ten-day regulations, the ground meteorological observation data acquisition terminal 312 is used for acquiring observation data of past ten-day-regulations meteorological automatic stations, and the weather numerical mode data acquisition terminal 313 is used for acquiring high-low air image element data of past ten days. The data storage card 4 is used for storing the original data acquired by the data acquisition terminal 31 and classifying and storing the original data according to the input date and the predicted time length of the keyboard 13. When the original data collected by the data collection terminal 31 is sent to the data storage card 4 for storage, the keyboard 13 marks dates on the established data set according to the time line sequence, establishes a data subset in the data set according to the ozone prediction time, and then sequentially stores the original data collected by the data collection terminal 31 in the data subset according to the date and prediction time sequence, thereby completing the storage process of the original data by the data storage card 4.
The characteristic screening processor 32 is configured to read the original data of the data storage card 4, and select a characteristic quantity having a large influence on ozone prediction in an objective selection manner, thereby avoiding elimination of an effective characteristic quantity caused by subjective screening.
The central model trainer 33 is configured to train the feature quantities optimized by the feature screening processor 32 through a training model. Illustratively, the central model trainer 33 uses the feature quantities obtained by screening to pass through an input layer, an LSTM network layer and an output layer through an LSTM-RNN (long-short term memory-Deep Neural network), the hidden layer memory cells are set to 10, the activation function is tanh (hyperbolic tangent), the optimizer selects an Adam algorithm learning step length to be 24, the whole sequence is divided into 20 time periods, and iterative training is performed for 200 times, so as to obtain a calculated training model, and further increase the calculation efficiency and the prediction accuracy.
And the data calculation unit 34 is used for calculating the ozone concentration according to the training model and the characteristic quantity and sending the calculation result to the display 12. The data calculation unit 34 sends the result of the ozone concentration prediction to the display 12, and the display 12 finally displays the prediction data of the ozone concentration.
Specifically, the feature filtering processor 32 may be a processor currently used by a computer; the central model trainer 33 may be a model training device capable of implementing the above functions in the prior art, such as a model training device disclosed in chinese patent CN201611073994.5, a model training device disclosed in CN201510362322.5, or the like; the data computation unit 34 may employ computation modules in the prior art, such as an intel STK1AW32SC 91065 computation module, a raspberry pi computation module, and the like.
The utility model discloses the theory of operation:
referring to the attached fig. 1-4 of the specification, after the power connector 14 is connected with the power supply 2, the bright condition of the indicator light 11 is observed, the chip 3, the display 12 and the keyboard 13 are ensured to be in a standby state, the date to be forecasted is input through the keyboard 13 according to the format yyyyymmdd, the ozone duration to be forecasted is input, and the selection can be carried out in 3 data subsets of 24 hours, 48 hours and 72 hours; the data acquisition terminal 31 acquires the original data of each hour in the past ten days according to the instruction input by the keyboard 13 and stores the original data in the data storage card 4; the feature screening processor 32, the central model trainer 33 and the data calculation unit 34 cooperate with each other to predict the ozone concentration for a specific time period by using the training model and the preferred feature quantity according to the predicted time period input by the keyboard 13, and then send the prediction result to the display 12, and the result of ozone prediction is displayed by the display 12.
Although the present invention has been described in detail with reference to the foregoing embodiments, it should be understood by those skilled in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; such modifications and substitutions do not depart from the spirit and scope of the present invention in its corresponding aspects.
Claims (5)
1. A near-surface ozone concentration prediction system based on machine learning, comprising a housing (1), characterized in that: an indicator lamp (11) is arranged at the top corner of the front side face of the shell (1), a display (12) is arranged on one side of the indicator lamp (11), a keyboard (13) is arranged on one side of the display (12), a power supply connector (14) is arranged on one side of the bottom of the front side face of the shell (1), the power supply connector (14) is externally connected with a battery (2), and the battery (2) is connected with the chip (3);
the chip (3) comprises a data acquisition terminal (31), a feature screening processor (32), a central model trainer (33) and a data calculation unit (34), wherein a data storage card (4) is arranged at the output end of the data acquisition terminal (31), the keyboard (13) is in one-way connection with the data storage card (4), the data storage card (4) is in one-way connection with the feature screening processor (32), the feature screening processor (32) is in one-way connection with the central model trainer (33), the central model trainer (33) is in one-way connection with the data calculation unit (34), and the data calculation unit (34) is in one-way connection with the display (12);
the utility model discloses a portable electronic device, including shell (1), shell (1) trailing flank is provided with base (5), base (5) and shell (1) joint.
2. The near-surface ozone concentration prediction system based on machine learning of claim 1, wherein:
the number of the indicator lamps (11) is three, and the indicator lamps (11) are respectively connected with the display (12), the keyboard (13) and the chip (3) from top to bottom.
3. The near-surface ozone concentration prediction system based on machine learning of claim 1, wherein:
the data acquisition terminal (31) comprises an atmospheric composition observation data acquisition terminal (311), a ground meteorological observation data acquisition terminal (312) and a weather numerical mode data acquisition terminal (313);
the system comprises an atmospheric composition observation data acquisition terminal (311), a ground meteorological observation data acquisition terminal (312) and a weather numerical mode data acquisition terminal (313), wherein the atmospheric composition observation data acquisition terminal (311) is respectively connected with a data storage card (4), the atmospheric composition observation data acquisition terminal (311) is used for acquiring past ten-day-standard atmospheric composition observation data, the ground meteorological observation data acquisition terminal (312) is used for acquiring past ten-day-standard meteorological automatic station observation data, and the weather numerical mode data acquisition terminal (313) is used for acquiring past ten-day-high-low air image element data.
4. The near-surface ozone concentration prediction system based on machine learning of claim 3, wherein:
the data storage card (4) is used for storing the original data acquired by the data acquisition terminal (31) and classifying and storing the original data according to the input date and the predicted time length of the keyboard (13);
the characteristic screening processor (32) is used for reading the original data of the data storage card (4) and optimizing the characteristic quantity with larger influence on ozone prediction in an objective selection mode;
the central model trainer (33) is used for carrying out model training on the feature quantity optimized by the feature screening processor (32) through a training model;
and the data calculation unit (34) is used for calculating the ozone concentration according to the training model and the characteristic quantity and sending the calculation result to the display (12).
5. The near-surface ozone concentration prediction system based on machine learning of any one of claims 1-4, wherein:
the chip (3) and the data storage card (4) are both arranged inside the shell (1), the indicator light (11), the display (12), the keyboard (13) and the chip (3) are all connected with the battery (2) through a power connector (14), and the keyboard (13) is composed of ten digital keys.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202022597854.6U CN213813502U (en) | 2020-11-11 | 2020-11-11 | Near-surface ozone concentration prediction system based on machine learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202022597854.6U CN213813502U (en) | 2020-11-11 | 2020-11-11 | Near-surface ozone concentration prediction system based on machine learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN213813502U true CN213813502U (en) | 2021-07-27 |
Family
ID=76932405
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202022597854.6U Active CN213813502U (en) | 2020-11-11 | 2020-11-11 | Near-surface ozone concentration prediction system based on machine learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN213813502U (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113791650A (en) * | 2021-08-16 | 2021-12-14 | 北京农业信息技术研究中心 | Method and system for adjusting ozone concentration |
-
2020
- 2020-11-11 CN CN202022597854.6U patent/CN213813502U/en active Active
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113791650A (en) * | 2021-08-16 | 2021-12-14 | 北京农业信息技术研究中心 | Method and system for adjusting ozone concentration |
| CN113791650B (en) * | 2021-08-16 | 2023-12-22 | 北京农业信息技术研究中心 | Ozone concentration adjusting method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106779223A (en) | A kind of photovoltaic system electricity generation power real-time predicting method and device | |
| CN112785458A (en) | Intelligent management and maintenance system for bridge health big data | |
| CN213813502U (en) | Near-surface ozone concentration prediction system based on machine learning | |
| CN111122667A (en) | Distributed air quality monitoring device and monitoring method thereof | |
| CN115173550A (en) | Distributed photovoltaic power generation real-time monitoring method and system | |
| CN113284026A (en) | Big data wisdom garden management system | |
| CN109242186A (en) | A kind of predicting power of photovoltaic plant system based on cloud platform | |
| CN115016339A (en) | Monitoring method, device and medium for outdoor power equipment | |
| CN109784614A (en) | Thermal power plant produces mobile management system | |
| CN106302820A (en) | Coastal Wetland Environment based on Internet of Things monitoring system | |
| CN107131908A (en) | Multiple-sensor network platform environment monitoring system and method | |
| CN117390403B (en) | Power grid fault detection method and system for new energy lighthouse power station | |
| CN114169797A (en) | Energy management system and method | |
| CN202102046U (en) | Harmonic monitor capable of displaying harmonic flow direction | |
| CN112884532A (en) | Low-power consumption thing networking collection system | |
| CN117010946A (en) | Thermal power plant production and operation cost accounting system and application method thereof | |
| CN116700057A (en) | Park energy-carbon double-control digital management system | |
| CN2752790Y (en) | Automatic meteorological observation apparatus | |
| CN116095723A (en) | Asset management system of communication base station | |
| CN115002171A (en) | Intelligent operation supervision system of sewage treatment facility | |
| CN113901026A (en) | Sulfur hexafluoride material management and control method and system based on big data | |
| CN112153162A (en) | Safety protection system based on Internet of things | |
| CN110824271A (en) | Electric automobile fills electric pile simulation test platform | |
| CN214410267U (en) | High-efficient low-power consumption thing networking collector | |
| CN202770471U (en) | Automatic test device of tablet computer |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| GR01 | Patent grant | ||
| GR01 | Patent grant |