阅读说明：本技术 一种气体传感器的远程校准方法及系统 (Remote calibration method and system for gas sensor ) 是由 范智勇 陈卓 周清峰 何颂贤 于 2020-07-21 设计创作，主要内容包括：本发明涉及一种气体传感器的远程校准系统及方法,所述系统至少包括：采集模块、标准传感器和云端服务器,所述采集模块至少包括浓度传感器、湿度传感器、温度传感器和用于采集所述浓度传感器的电参数的电参数采集模块,所述云端服务器至少包括建模模块和校准模块,所述建模模块基于所述采集模块发送的原始浓度参数、湿度参数、温度参数、电参数和标准传感器发送的标准数据按照深度学习算法建立校准模型,所述校准模块基于所述校准模型对所述采集模块发送的原始浓度参数进行校准。通过将气体传感器采集状态的电数据作为校准影响因素纳入校准模型中,提高校准模型的校准率,获得更准确的气体浓度参数。(The invention relates to a remote calibration system and a method of a gas sensor, wherein the system at least comprises the following components: collection module, standard sensor and high in the clouds server, collection module includes concentration sensor, humidity transducer, temperature sensor at least and is used for gathering concentration sensor's electric parameter collection module, the high in the clouds server includes modeling module and calibration module at least, modeling module is based on the calibration model is established according to the deep learning algorithm to the original concentration parameter that collection module sent, humidity parameter, temperature parameter, electric parameter and standard sensor sent, calibration module is based on the calibration model is right the original concentration parameter that collection module sent calibrates. The electric data of the gas sensor collecting state is taken as a calibration influence factor to be brought into the calibration model, so that the calibration rate of the calibration model is improved, and more accurate gas concentration parameters are obtained.)

1. A remote calibration system for a gas sensor, comprising at least: an acquisition module (10), a standard sensor (20) and a cloud server (30), characterized in that,

the acquisition module (10) at least comprises a concentration sensor (11), a humidity sensor (12), a temperature sensor (13) and an electrical parameter acquisition module (14) for acquiring electrical parameters of the concentration sensor (11),

the cloud server (30) comprising at least a modeling module (40) and a calibration module (50), wherein,

the modeling module (40) establishes a calibration model according to a deep learning algorithm based on the original concentration parameter, the humidity parameter, the temperature parameter and the electrical parameter sent by the acquisition module (10) and the standard data sent by the standard sensor,

the calibration module (50) calibrates the raw concentration parameters sent by the acquisition module (10) based on the calibration model.

2. The remote calibration system for a gas sensor of claim 1,

the feature vectors of the calibration model of the modelling module (40) are defined as:

wherein

3. The remote calibration system for a gas sensor according to claim 2, wherein the method for establishing the calibration model of the modeling module comprises:

establishing a calibration model based on a deep learning algorithm and initializing the calibration model,

the feature vector is combinedInputting the deep learning algorithm for training and optimization,

up to the predicted concentration of the calibration model output

4. The remote calibration system for gas sensors of claim 3, wherein the calibration module (50) calibrates the raw data sent by the acquisition module (10) based on the calibration model sent by the modeling module (40),

the original data comprises original concentration parameters, humidity parameters, temperature parameters and electric parameters sent by the acquisition module.

5. The remote calibration system for gas sensors according to claim 3, wherein the data processing module in the gas sensor calibrates the raw data sent by the acquisition module (10) based on the calibration model sent by the modeling module (40),

the original data comprises original concentration parameters, humidity parameters, temperature parameters and electric parameters sent by the acquisition module,

and the data processing module sends the calibrated gas concentration parameters to a database (60) of the cloud server.

6. The remote calibration system for a gas sensor according to any one of claims 1 to 5,

the calibration module (50) determines the gas environment characteristics based on the temperature parameters, the humidity parameters and the weather forecast information sent by the third-party platform (15), and the calibration module (50) generates calibration suggestion information containing calibration time and time length thereof based on the gas environment characteristics, or

The calibration module (50) automatically starts a standard sensor (20) to collect standard data of at least one type of gas environment characteristic based on the gas environment characteristic and preset calibration condition information,

the modeling module (40) responds to an optimization instruction sent by the calibration module (50) and optimizes the calibration model based on the raw data sent by the acquisition module (10) and the standard data sent by the standard sensor (20).

7. A method of remote calibration of a gas sensor, the method comprising:

establishing a calibration model according to a deep learning algorithm based on an original concentration parameter, a humidity parameter, a temperature parameter and an electrical parameter which are acquired by a gas sensor and standard data which are sent by a standard sensor,

calibrating the raw concentration parameters sent by the acquisition module (10) based on the calibration model.

8. The remote calibration method for a gas sensor according to claim 7,

the feature vector of the calibration model is defined as:

wherein the content of the first and second substances,

9. The method for remotely calibrating a gas sensor according to claim 8, wherein the method for establishing the calibration model comprises:

establishing a calibration model based on a deep learning algorithm and initializing the calibration model,

the feature vector is combined

up to the predicted concentration of the calibration model outputConcentration parameter y from standard data_tIs less than the error threshold, forming a calibration model.

10. The remote calibration method for a gas sensor according to any one of claims 7 to 9, further comprising:

determining the gas environment characteristics based on the temperature parameters, the humidity parameters and the weather forecast information sent by the third party platform (15),

generating calibration recommendation information containing calibration time and time length thereof based on the gas environment characteristics;

or

Automatically activating a standard sensor (20) to acquire standard data of at least one type of gas environment characteristic based on the gas environment characteristic and preset calibration condition information,

the optimization of the calibration model is performed on the basis of the raw data sent by the acquisition module (1O) and the standard data sent by the standard sensor (20).

Technical Field

The invention belongs to the technical field of gas sensing, and particularly relates to a remote calibration method and system of a gas sensor.

Background

Under the policy of environmental protection and economic development in the new era, atmospheric gridding monitoring is becoming a development trend in the fields of environmental monitoring and atmospheric pollution early warning. The atmospheric gridding monitoring is used as a supplement of the current commercial monitoring station, and has the characteristics of relatively low cost, easy construction and wide coverage. Through the sensor and the integration of wireless transmission module that the calibration was accomplished, be equipped with power management module, can set up miniature air monitoring station and be used for the monitoring of atmosphere meshing. The gas sensors used in the micro air station are often electrochemical sensors, and as the use time increases, the consumption of electrolyte in the electrochemical sensors, the aging of electrodes, and the attenuation of the filtering effect of the filtering membrane can cause the drift of the sensor base line and the change of the sensitivity. To increase the service life of the micro air station, the gas sensors must be periodically maintained or calibrated.

There are two common maintenance schemes for electrochemical sensors. Firstly, an aging experiment is carried out before leaving a factory, the exposure process of the gas sensor in the atmosphere is simulated, then a compensation algorithm is generated to correct the response of the sensor in advance, and a certain degree of aging resistance and self calibration after installation are realized. The second scheme is to regularly maintain the micro air station, replace a new gas sensor, or return the original gas sensor to the factory for secondary calibration in the laboratory. However, the service life of the sensor can be prolonged to a certain extent by the compensation algorithm generated by aging simulation before delivery, and then the compensation algorithm is difficult to maintain accurate compensation after the sensor is used for a long time due to unpredictability of atmospheric environment temperature, humidity and gas concentration changes. The method of returning to the factory for recalibration or directly replacing with a new sensor is obviously costly, inefficient and time consuming. Therefore, it is important to perform online calibration of the gas sensor. There are also many on-line calibration systems in the prior art.

For example, patent document CN105938133B discloses a method and a system for online calibration of a wireless gas sensor, the method including: acquiring a gas concentration sequence acquired by each wireless gas sensor; eliminating deviation data in each gas concentration sequence, taking a median of each gas concentration sequence after eliminating the deviation data, and obtaining a gas concentration vector of each wireless gas sensor; grouping the gas concentration vectors according to gas types, and acquiring the weight of each corresponding wireless gas sensor according to the grouped gas concentration vectors by using a fusion algorithm; acquiring first concentrations of various gases according to the weights of the wireless gas sensors of the groups; fitting the gas concentration of each wireless gas sensor at each moment and the corresponding first concentration to obtain the adjusting parameters of each wireless gas sensor; and calibrating each wireless gas sensor according to the adjusting parameters. However, the drawback of this patent is that,

furthermore, on the one hand, due to the differences in understanding to the person skilled in the art; on the other hand, since the inventor has studied a lot of documents and patents when making the present invention, but the space is not limited to the details and contents listed in the above, however, the present invention is by no means free of the features of the prior art, but the present invention has been provided with all the features of the prior art, and the applicant reserves the right to increase the related prior art in the background.

Disclosure of Invention

The word "module" as used herein describes any type of hardware, software, or combination of hardware and software that is capable of performing the functions associated with the "module".

In view of the deficiencies of the prior art, the present invention provides a remote calibration system for a gas sensor, comprising: collection module, standard sensor and high in the clouds server, its characterized in that, collection module includes concentration sensor, humidity transducer, temperature sensor at least and is used for gathering the electric parameter collection module of concentration sensor's electric parameter, the high in the clouds server includes modeling module and calibration module at least, wherein, modeling module is based on the calibration model is established according to the deep learning algorithm to the original concentration parameter that collection module sent, humidity parameter, temperature parameter, electric parameter and standard sensor sent, calibration module is based on the calibration model is right the original concentration parameter that collection module sent calibrates. The electric data of the gas sensor collecting state is taken as a calibration influence factor to be brought into the calibration model, so that the calibration rate of the calibration model is improved, and more accurate gas concentration parameters are obtained.

Preferably, the feature vector of the calibration model of the modeling module is defined as:

wherein the content of the first and second substances,

representing a feature vector, V, corresponding to a predicted value of the gas concentration_WEIndicating the working electrode output voltage, V, of the concentration sensor_AUXThe output voltage of the auxiliary electrode is shown, T is a temperature parameter, and RH is a humidity parameter. The characteristic vector of the invention comprises electric data and temperature and humidity parameters, which is more beneficial to reducing the influence of environmental factors on gas concentration measurement, so that the gas concentration data after calibration is more accurate.

Preferably, the method for establishing the calibration model of the modeling module includes:

establishing a calibration model based on a deep learning algorithm, initializing the calibration model, and analyzing the feature vectors

Inputting the deep learning algorithm for training and optimizing until the predicted concentration output by the calibration modelConcentration parameter y from standard data_tIs less than the error threshold, forming a calibration model. The method is trained by a deep learning method, so that the error between the obtained predicted concentration value and the labeled value is small, and the optimization of a calibration model is facilitated.

Preferably, the calibration module calibrates the raw data sent by the acquisition module based on the calibration model sent by the modeling module, where the raw data includes the raw concentration parameter, the humidity parameter, the temperature parameter, and the electrical parameter sent by the acquisition module.

Preferably, a data processing module in the gas sensor calibrates the original data sent by the acquisition module based on the calibration model sent by the modeling module, the original data includes an original concentration parameter, a humidity parameter, a temperature parameter and an electrical parameter sent by the acquisition module, and the data processing module sends the calibrated gas concentration parameter to a database of a cloud server.

Preferably, the calibration module determines the gas environment characteristics based on the temperature parameters, the humidity parameters and the weather forecast information sent by the third-party platform, the calibration module generates calibration suggestion information including calibration time and duration thereof based on the gas environment characteristics, or the calibration module automatically starts a standard sensor to collect standard data of at least one type of gas environment characteristics based on the gas environment characteristics and preset calibration condition information, and the modeling module performs optimization of the calibration model based on the raw data sent by the collection module and the standard data sent by the standard sensor in response to an optimization instruction sent by the calibration module. Through the determination of the sum of the gas environment characteristics, the calculation and the selection of the calibration module are facilitated, and the environment characteristics which are more favorable for calibration are selected, so that the influence of the temperature and humidity characteristics of the environment is brought into the calibration model, and the more accurate calibration model is obtained.

The invention also provides a remote calibration method of the gas sensor, which is characterized by comprising the following steps: establishing a calibration model according to a deep learning algorithm based on an original concentration parameter, a humidity parameter, a temperature parameter and an electrical parameter acquired by a gas sensor and standard data sent by a standard sensor, and calibrating the original concentration parameter sent by the acquisition module based on the calibration model. According to the remote calibration method, the electric data of the gas sensor collecting state is taken as the calibration influence factor to be incorporated into the calibration model, so that the calibration rate of the calibration model is improved, and more accurate gas concentration parameters are obtained.

In the remote calibration method of the gas sensor of the present invention, the eigenvector of the calibration model is defined as:

whereinRepresenting a feature vector, V, corresponding to a predicted value of the gas concentration_WEIndicating the working electrode output voltage, V, of the concentration sensor_AUXThe output voltage of the auxiliary electrode is shown, T is a temperature parameter, and RH is a humidity parameter. The characteristic vector of the invention realizes the compensation of temperature and humidity and ensures the accurate measurement of the target gas in the atmospheric environment with variable temperature and humidity.

The invention discloses a remote calibration method of a gas sensor, which comprises the following steps:

establishing a calibration model based on a deep learning algorithm and initializing the calibration model,

the feature vector is combinedInputting the deep learning algorithm for training and optimizing until the predicted concentration output by the calibration model

Concentration parameter y from standard data_tIs less than the error threshold, forming a calibration model. The method for establishing the calibration model is more beneficial to improving the accuracy of calibration.

Preferably, the remote calibration method further comprises:

the method comprises the steps of determining gas environment characteristics based on temperature parameters, humidity parameters and weather forecast information sent by a third-party platform, generating calibration suggestion information comprising calibration time and time length based on the gas environment characteristics, or automatically starting a standard sensor to acquire standard data of at least one type of gas environment characteristics based on the gas environment characteristics and preset calibration condition information, and optimizing a calibration model based on original data sent by an acquisition module and the standard data sent by the standard sensor. The calibration method remotely starts the standard sensor for calibration through the preset condition information without manual intervention of workers, and eliminates the adverse effect of manually starting a calibration program.

The invention has the beneficial technical effects that:

the gas sensor remote calibration system provided by the invention can realize in-situ calibration of the gas sensor and carry out temperature and humidity compensation without returning the sensor to a factory for calibration. Meanwhile, as the remote calibration can be carried out at any time, the service life of the sensor can be better prolonged, and the accurate measurement for a longer time is realized.

Drawings

FIG. 1 is a schematic diagram of the logic blocks of the system for remotely calibrating a gas sensor of the present invention;

FIG. 2 is a schematic diagram of data training in the process of establishing a calibration model according to the present invention;

FIG. 3 is a data plot of the invention after data calibration.

List of reference numerals

10: an acquisition module; 11: a concentration sensor; 12: a humidity sensor; 13: a temperature sensor; 14: an electrical parameter acquisition module; 15: a third party platform; 20: a standard sensor; 30: a cloud server; 40: a modeling module; 50: a calibration module; 60: a database.

Detailed Description

The following detailed description is made with reference to the accompanying drawings.

The invention provides a remote calibration system and a remote calibration method for a gas sensor. The invention also discloses a modeling method and a modeling system based on the gas sensor remote calibration.