CN111596006A

CN111596006A - Calibration method of atmosphere online monitor and monitor

Info

Publication number: CN111596006A
Application number: CN202010355441.9A
Authority: CN
Inventors: 敖小强; 李永帅; 刘建坡; 李妍
Original assignee: Beijing SDL Technology Co Ltd
Current assignee: Beijing SDL Technology Co Ltd
Priority date: 2020-04-29
Filing date: 2020-04-29
Publication date: 2020-08-28

Abstract

The application relates to a calibration method of an atmosphere online monitor, which comprises the following steps: in a preset time period, simultaneously monitoring gas of a target environment by using a standard monitor and a monitor to be calibrated to respectively obtain sensor raw data, wherein the sensor raw data comprises a sensor output signal and a standard monitoring output of the standard monitor; and training an RNN-LSTM neural network by using the sensor raw data to obtain a calibration model of the monitor to be calibrated.

Description

Calibration method of atmosphere online monitor and monitor

Technical Field

The application belongs to the field of atmospheric monitoring, and particularly relates to a calibration method of an atmospheric on-line monitor and the monitor.

Background

At present, a plurality of sensors are generally used in an on-line atmosphere monitor, and a plurality of signals are output. The signal response dynamic process transfer function is complex due to the non-linearity of the sensor. As the number of sensors increases, the correction parameters become too large. The inventors of the present application found that: the calibration is carried out by using the traditional method, the manual workload is large, the difficulty is high, and the calibration is difficult to adapt to the calibration work of a complex system.

The inventors of the present application found that the present method of calibrating a measurement instrument using a neural network has only a static mapping without a signal dynamic response process prediction part. The dynamic response process of the output data of the instrument corrected by the method is not accurate enough, and the error in the dynamic process is large.

Disclosure of Invention

The application aims to provide a calibration method of an atmosphere on-line monitor and a detector.

One embodiment of the present application provides a calibration method for an online atmosphere monitor, including: in a preset time period, simultaneously monitoring gas of a target environment by using a standard monitor and a monitor to be calibrated to respectively obtain sensor raw data, wherein the sensor raw data comprises a sensor output signal and a standard monitoring output of the standard monitor; and training an RNN-LSTM neural network by using the sensor raw data to obtain a calibration model of the monitor to be calibrated.

Another embodiment of the present application provides an online atmosphere monitor, comprising: a sensor array sensing gas output sensor raw data of a target environment; and the atmosphere monitoring module is used for carrying out data processing on the original data based on a calibration model to output a calibration value, wherein the calibration model is an RNN-LSTM neural network model.

By using the method and the monitor, the measuring instrument can be calibrated by the neural network by introducing the RNN-LSTM network in the calibration process, and a signal dynamic response process prediction part is not provided. By using the method and the monitor, the derivative quantity in the differential equation in the dynamic mathematical model can be mapped through the time sequence data relationship, the dynamic response process of the output data of the monitor is more accurate, and the error in the dynamic process is reduced (the derivative quantity in the differential equation is deducted).

Drawings

FIG. 1 illustrates a method for calibrating an on-line atmospheric monitor according to one embodiment of the present application.

FIG. 2 shows a schematic diagram of the connections of the apparatus for calibrating the monitor using the method of FIG. 1.

Fig. 3 shows a schematic composition diagram of an atmosphere on-line monitor according to another embodiment of the present application.

Fig. 4 shows another embodiment of the present application, an atmospheric monitoring instrument calibration device.

FIG. 5 shows a block diagram of an electronic device according to an example embodiment.

Detailed Description

The following description will be made by way of specific embodiments of the present disclosure on the "calibration method for an on-line air monitor and monitor", and those skilled in the art will understand the advantages and effects of the present disclosure from the disclosure of the present disclosure. The invention is capable of other and different embodiments and its several details are capable of modification in various other respects, all without departing from the spirit and scope of the present invention. The drawings of the present invention are for illustrative purposes only and are not intended to be drawn to scale. The following embodiments will further explain the related art of the present invention in detail, but the disclosure is not intended to limit the scope of the present invention.

The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are some, but not all, embodiments of the present application. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.

It should be understood that the terms "first," "second," "third," and "fourth," etc. in the claims, description, and drawings of the present application are used for distinguishing between different objects and not for describing a particular order. The terms "comprises" and "comprising," when used in the specification and claims of this application, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.

It is also to be understood that the terminology used in the description of the present application herein is for the purpose of describing particular embodiments only, and is not intended to be limiting of the application. As used in the specification and claims of this application, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It should be further understood that the term "and/or" as used in the specification and claims of this application refers to any and all possible combinations of one or more of the associated listed items and includes such combinations.

As shown in fig. 1, the method 1000 may include: s110 and S120.

In S110, the gas in the target environment may be monitored simultaneously by the standard monitor and the monitor to be calibrated within a preset time period. And obtaining raw sensor data within the preset time period, wherein the raw sensor data comprises a sensor output signal. In S110, a standard monitor output of a standard monitor may also be obtained. Wherein the standard monitoring output may include a concentration of at least one predetermined component contained in the gas in the target environment.

Alternatively, the monitor to be calibrated may comprise a plurality of sensors. Alternatively, the plurality of sensors may output a plurality of signals. Optionally the plurality of signals may comprise a plurality of dynamic signals. Optionally the plurality of dynamic signals may comprise at least one dynamic non-linear signal.

Optionally, the target environment may be a preset simulation environment. Optionally, the monitor to be calibrated and the standard monitor may be simultaneously placed in an atmosphere simulation apparatus, and a preset contaminated gas may be introduced into the atmosphere simulation apparatus according to a sequence established by a preset process file. The output data of the standard monitor and the sensor raw data output by the sensor array of the monitor to be calibrated can be automatically recorded.

As shown in FIG. 1, the RNN-LSTM neural network may be trained using the raw sensor data to obtain a calibration model of the monitor to be calibrated in S120. Wherein the calibration model can be used to calculate a monitored output of the gas of the target environment from raw outputs of the plurality of sensors. Alternatively, the target environment may be a real environment. The calibration model may be a linear model, a nonlinear model, a dynamic model, or a multiple-input/multiple-output model. Further, the model may be a multiple-input/multiple-output nonlinear dynamical model.

The RNN-LSTM neural network is a mathematical model that performs distributed parallel information processing by mimicking behavioral characteristics of biological neural networks. The network achieves the aim of information processing by adjusting the mutual connection relationship among a large number of internal nodes depending on the complexity of the system.

The artificial neural network has self-learning and self-adapting capabilities, the internal relation and the law of the artificial neural network and the input/output data can be analyzed through a batch of pre-provided input/output data corresponding to each other, and finally a complex nonlinear system function is formed through the laws, and the learning and analyzing process is called as 'training'. Each input connection of a neuron has a synaptic connection strength, represented by a connection weight, by which the signal to be generated is amplified, each input quantity corresponding to an associated weight. The processing unit quantizes the weighted inputs, adds the weighted inputs to obtain a sum of weighted values, and calculates an output quantity, wherein the output quantity is a function of the sum of the weights, and the function is generally called a transfer function.

RNN-LSTM recurrent artificial neural networks: RNN differs from feed-forward neural networks in that RNN can use its internal memory to process input sequences at arbitrary timing. LSTM is a variation of RNN, and comprises a forgetting gate, an input gate, and an output gate.

In conventional RNN, the training algorithm is BPTT (Back-propagation Through Time, Back propagation). However, when the time period is long, the BPTT causes the RNN network to return the required residual error exponentially, which results in slow update of the network weight and failure to exhibit the long-term RNN memory effect, and therefore a memory unit is required to store the RNN memory.

Therefore, an improved model of RNN is proposed: long-short Term Memory model (LSTM). The special RNN network model solves the problem of RNN model gradient diffusion. LSTM has a "triple gate": and an input gate i, an output gate o and a forgetting gate f limit the value range to be within (0,1) by using a Sigmoid function. The three gates can control the information flow direction at different moments, and proper information is selected to enter the central cells by controlling the forgetting gate and the input gate, so that irrelevant information is rejected; the information after cell processing is output at the most appropriate timing by controlling the output gate.

In addition to LSTM, GRUs, bi-directional RNNs or SRUs may be selected as a temporal neural network model for residual duration prediction according to some embodiments of the present invention.

In S120, parameters are automatically adapted by using the sensor raw data and the output of the standard monitor within the preset time period, and by using a recurrent artificial neural network having a history information memory function and a feedback function, and by using a machine learning algorithm.

Optionally, after S120, the method may further include: and transmitting the trained calibration model into the monitor to be calibrated. Optionally, a memory (not shown) and a processor (not shown) may be included within the monitor under test. Wherein the memory may be configured to store the calibration model and the processor may be configured to execute the calibration model. When the above model is executed, the processor may calculate gas monitoring data for the target environment using raw signals from the sensors and output the data. Optionally, the monitoring data may include a concentration of at least one predetermined component in the target environment.

As shown in fig. 2, the method 1000 involves a monitor 11 to be calibrated, a standard monitor 13 and a calibration device 12.

Alternatively, the monitor 11 to be calibrated (the instrument to be calibrated) may contain a sensor array 111. Alternatively, the sensor array 111 may include a plurality of sensors of a plurality of types. Alternatively, the sensor array 111 may be configured to sense a plurality of technical indicators of the gas in the target environment and output a plurality of sensor raw signals of a plurality of sensors. The reference monitor 13 (reference instrument) may be used to output a reference monitor output (reference instrument output).

The calibration device 12 can be connected to the monitor 11 to be calibrated and the reference monitor 13, respectively. Alternatively, the calibration device 12 may be electrically connected to at least one of the monitor 11 to be calibrated and the monitoring standard 13, or may be coupled to at least one of the monitor 11 to be calibrated and the monitoring standard 13. Alternatively, the calibration device 12 may be communicatively connected to at least one of the monitor 11 to be calibrated and the reference monitor 13. Alternatively, the calibration device 12 may receive a raw sensor signal from the monitor 11 to be calibrated and a standard monitor output from the standard monitor 13. Optionally, the calibration device 12 may also transmit a calibrated calibration model to the monitor 11 to be calibrated.

Optionally, the calibration device 12 may include a standard monitor output receiving module (not shown) for receiving a standard monitor output from the standard monitor 13. The calibration device 12 may further include a sensor raw signal receiving module (not shown) for receiving a sensor raw signal output from the monitor 11 to be calibrated. Optionally, the calibration apparatus 12 may further include an RNN-LSTM recurrent artificial neural network model training module (not shown), which is respectively connected to the standard monitoring output receiving module and the sensor raw signal receiving module. Alternatively, the RNN-LSTM recurrent artificial neural network model training module may be used to train the RNN-LSTM recurrent artificial neural network model using the sensor raw signals and the standard monitoring output. Alternatively, the RNN-LSTM recurrent artificial neural network model training module may include a processor (not shown) for model training and a memory (not shown) for storing training programs and training data. Wherein the processor may be a processor array of multiple processors operable in parallel. Optionally, the calibration device 12 may further comprise a model transmission module (not shown) for transmitting the trained calibration model to the monitor 11 to be calibrated.

As shown in fig. 3, the monitor 2000 may include a sensor array 21 and an atmospheric monitoring module 22.

The sensor array 21 may include a plurality of sensors for atmospheric quality monitoring, among other things. Alternatively, the sensor array 21 may be used to sense the gas in the target environment and collect a plurality of technical data required to monitor the gas in the target environment. And may output the collected sensor raw data to the atmosphere monitoring module 22. Alternatively, the target environment may be a real environment.

Alternatively, the sensor raw data may be a plurality of sets of outputs of a plurality of sensors within a preset time period. Optionally, there are multiple sets of dynamic data in the data output by the plurality of sensors. Optionally, there are multiple sets of dynamic nonlinear data in the multiple sets of dynamic data.

The atmosphere monitoring module 22 may process the raw data of the sensor by using a trained calibration model, which is an RNN-LSTM neural network model, and calculate the monitoring data of the gas in the target environment, and may output the monitoring data. Optionally, the monitoring data may comprise a concentration of at least one predetermined component contained within the target environment.

Optionally, the atmospheric monitoring module 22 may include a memory (not shown) and an atmospheric monitoring processor (not shown). The memory may be used to store a calibration model that the atmospheric monitoring processor may execute. When the calibration model is executed, the atmospheric monitoring processor may process the sensor raw data using the calibration model and obtain gas monitoring data for the target environment.

Optionally, the monitor 2000 may further include a calibration module (not shown) that interfaces with the calibration device of fig. 2. In the calibration state, the calibration module may be used to output raw sensor data from the sensor array 21 and may be used to receive a trained calibration model from the aforementioned calibration apparatus. Optionally, the calibration module may also transmit the received calibration model to the atmospheric monitoring module 22.

The apparatus 3000 as shown in fig. 4 may include a sensor raw data receiving unit 31, a standard monitoring data receiving unit 32, and a model training unit 33.

The sensor raw data receiving unit 31 may be configured to receive sensor raw data from the monitor 35 to be calibrated. Alternatively, the raw sensor data may be quantized data obtained by converting at least one predetermined physical quantity into at least one sensor (not shown) in a sensor array (not shown) in the monitor 35 to be calibrated while sensing gas in the target environment.

Alternatively, the sensor raw data may be a plurality of dynamic data output by a plurality of sensors. Alternatively, the sensor raw data may be a plurality of dynamic nonlinear data output by a plurality of sensors. Alternatively, the raw data may be at least one dynamic non-linear data output by the at least one sensor over a preset time period.

Optionally, the target environment may be a preset simulation environment. Optionally, the monitor 35 to be calibrated and the standard monitor may be simultaneously placed in an atmosphere simulation apparatus, and a preset contaminated gas may be introduced into the atmosphere simulation apparatus according to a sequence established by a preset process file. The output data of standard monitor 36 and the raw sensor data output by the sensor array of monitor 35 to be calibrated may be automatically recorded.

Alternatively, the sensor raw data receiving unit 31 may be connected to the monitor 35 to be calibrated in a communication manner, and may acquire the sensor raw data of the monitor 35 to be calibrated by a communication means. Alternatively, the sensor raw data receiving unit 31 may be connected to the monitor 35 to be calibrated by wire or wirelessly. Alternatively, the sensor raw data receiving unit 31 may be connected to the monitor 35 to be calibrated via a public network connection or a private network connection. Alternatively, the sensor raw data receiving unit 31 may receive the sensor raw data through a storage medium. The storage medium may include externally accessible memory within the monitor under test, memory connected to intermediate electronics between the raw data receiving unit 31 and the monitor 35 under calibration, and may include a removable storage medium. Alternatively, the raw data receiving unit 31 and the monitor 35 to be calibrated may not be connected, and the sensor raw data may be transmitted through a removable storage medium. Alternatively, the sensor raw data receiving unit 31 may acquire the sensor raw data of the monitor 35 to be calibrated in real time. Or acquiring the sensor raw data of the instrument to be calibrated, which is stored in a preset storage medium and is generated by sensing the gas in the target environment within the past preset time period. This past preset time period matches in time with the standard monitoring data mentioned below. The manner in which the sensor raw data receiving unit 31 acquires the sensor raw data may not be limited to this.

The standard monitoring data receiving unit 32 may be configured to receive the atmospheric monitoring data output from the standard monitor 36. The atmospheric monitoring data may be a concentration of at least one predetermined component in the atmosphere. Alternatively, the at least one predetermined component may be at least one contaminant. Alternatively, the predetermined composition may be a gas composition, a nebulized liquid composition, or a suspended particulate matter.

Optionally, the connection mode of the standard monitoring data receiving unit 32 and the standard monitor 36 may be the same as the connection mode of the sensor raw data receiving unit 31 and the monitor 35 to be calibrated, which is not described herein again. The manner of acquiring the standard monitoring data by the standard monitoring data receiving unit 32 may also be similar to the manner of acquiring the raw sensor data by the raw sensor data receiving unit 31, and is not described herein again.

Alternatively, the model training unit 33 may train the RNN-LSTM neural network model using the aforementioned sensor raw data and the aforementioned standard monitoring data, and use the model as a calibration model. Optionally, the model training unit may comprise a model training processor 331 and a memory 332.

Wherein memory 332 may be used to store the aforementioned sensor raw data, the aforementioned standard monitoring data and model training programs, and data generated by the model training process. Alternatively, the memory 332 may include at least one of a single-port memory, a dual-port memory, and a multi-port memory. Optionally, the memory may include at least one of dynamic memory, static memory, and non-volatile memory.

The model training processor 331 may execute a model training program stored in the memory 332 and train the RNN-LSTM neural network model using the sensor raw data and the standard monitoring data. Wherein the RNN-LSTM neural network model may be stored as a calibration model in the monitor shown in fig. 3. The monitor can process raw sensor data generated by a sensor array in the monitor by using the calibration model, and calculate to obtain atmospheric monitoring data of a target environment. The target environment may be a real environment and the atmospheric monitoring data may preset the concentration of the constituent.

Alternatively, the model training processor 331 may be a dedicated processor for model training. Alternatively, the model training processor 331 may be a processor array of multiple processors. Alternatively, the plurality of processors may exchange data with each other through the memory 332. Alternatively, the plurality of sensors may cooperate in a pipelined manner. Alternatively, the plurality of sensors may be divided into a plurality of groups, with the sensors within a group working in pipeline cooperation and the sensors between groups working in parallel.

Optionally, the apparatus 3000 may further comprise a model transmission unit 34. The model transmission unit 34 may be used to transmit the model trained by the model training unit 33 to the monitor 35 to be calibrated. Optionally, the model transmission unit 34 may be connected to the monitor 35 to be calibrated. The connection between the optional model transmission unit 34 and the monitor 35 to be calibrated may be the same as the connection between the sensor raw data receiving unit 31 and the monitor 35 to be calibrated, and is not described herein again.

An electronic device 200 according to this embodiment of the present application is described below with reference to fig. 5. The electronic device 200 shown in fig. 5 is only an example, and should not bring any limitation to the functions and the scope of use of the embodiments of the present application.

As shown in fig. 5, the electronic device 200 is embodied in the form of a general purpose computing device. The components of the electronic device 200 may include, but are not limited to: at least one processing unit 210, at least one memory unit 220, a bus 230 connecting different system components (including the memory unit 220 and the processing unit 210), a display unit 240, and the like.

Wherein the storage unit stores program code executable by the processing unit 210 to cause the processing unit 210 to perform the methods according to various exemplary embodiments of the present application described herein. For example, the processing unit 210 may perform the method as shown in fig. 1.

The memory unit 220 may include readable media in the form of volatile memory units, such as a random access memory unit (RAM)2201 and/or a cache memory unit 2202, and may further include a read only memory unit (ROM) 2203.

The storage unit 220 may also include a program/utility 2204 having a set (at least one) of program modules 2205, such program modules 2205 including, but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which, or some combination thereof, may comprise an implementation of a network environment.

Bus 230 may be one or more of several types of bus structures, including a memory unit bus or memory unit controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.

The electronic device 200 may also communicate with one or more external devices 300 (e.g., keyboard, pointing device, bluetooth device, etc.), with one or more devices that enable a user to interact with the electronic device 200, and/or with any devices (e.g., router, modem, etc.) that enable the electronic device 200 to communicate with one or more other computing devices. Such communication may occur via an input/output (I/O) interface 250. Also, the electronic device 200 may communicate with one or more networks (e.g., a Local Area Network (LAN), a Wide Area Network (WAN), and/or a public network such as the Internet) via the network adapter 260. The network adapter 260 may communicate with other modules of the electronic device 200 via the bus 230. It should be appreciated that although not shown in the figures, other hardware and/or software modules may be used in conjunction with the electronic device 200, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, among others.

As will be appreciated by one skilled in the art, aspects of the present application may be embodied as a system, method or computer program product. Accordingly, this application may take the form of an entirely hardware embodiment, an entirely software embodiment (including firmware, resident software, micro-code, etc.) or an embodiment combining software and hardware aspects that may all generally be referred to as a "circuit," module "or" system. Furthermore, the present application may take the form of a computer program product embodied in any tangible expression medium having computer-usable program code embodied in the medium.

The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each block of the flowchart illustrations and/or block diagrams, and combinations of blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions/acts specified in the flowchart and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable medium that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable medium produce an article of manufacture including instruction means which implement the function/act specified in the flowchart and/or block diagram block or blocks.

The computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide processes for implementing the functions/acts specified in the flowchart and/or block diagram block or blocks.

The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present application. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.

The foregoing detailed description of the embodiments of the present application has been presented to illustrate the principles and implementations of the present application, and the description of the embodiments is only intended to facilitate the understanding of the methods and their core concepts of the present application. Meanwhile, a person skilled in the art should, according to the idea of the present application, change or modify the embodiments and applications of the present application based on the scope of the present application. In view of the above, the description should not be taken as limiting the application.

Claims

1. A calibration method of an atmosphere online monitor comprises the following steps:

in a preset time period, simultaneously monitoring gas of a target environment by using a standard monitor and a monitor to be calibrated to respectively obtain sensor raw data, wherein the sensor raw data comprises a sensor output signal and a standard monitoring output of the standard monitor;

and training an RNN-LSTM neural network by using the sensor raw data to obtain a calibration model of the monitor to be calibrated.

2. The method of claim 1, wherein the target environment is a preset simulated environment.

3. The method of claim 1, wherein the monitor to be calibrated comprises a plurality of sensors.

4. The method of claim 3, wherein the plurality of sensors output a plurality of signals.

5. The method of claim 4, wherein the plurality of signals comprises a plurality of dynamic signals.

6. The method of claim 5, wherein the plurality of signals comprises a plurality of dynamic nonlinear signals.

7. The method of claim 1, wherein the target environment is a real environment.

8. An online atmosphere monitor comprising:

a sensor array sensing gas output sensor raw data of a target environment;

an atmosphere monitoring module for performing data processing on the raw data based on a calibration model to output a calibration value, wherein the calibration model is an RNN-LSTM neural network model,

the atmospheric monitoring module includes an atmospheric monitoring processor that executes the calibration model.

9. The monitor of claim 1, wherein the sensor raw data is dynamic data.

10. The monitor of claim 9, wherein the dynamic data is dynamic nonlinear data.