CN116736703A

CN116736703A - Intelligent monitoring system and method for sewage treatment

Info

Publication number: CN116736703A
Application number: CN202310592011.2A
Authority: CN
Inventors: 文祥; 文乐为; 陈建国; 吴建鑫; 刘湘
Original assignee: HUNAN SANLIAN ENVIRONMENTAL PROTECTION SCI-TECH CO LTD
Current assignee: HUNAN SANLIAN ENVIRONMENTAL PROTECTION SCI-TECH CO LTD
Priority date: 2023-05-24
Filing date: 2023-05-24
Publication date: 2023-09-12

Abstract

An intelligent monitoring system for sewage treatment and a method thereof, which acquire ammonia concentration values at a plurality of preset time points in a preset time period acquired by a gas sensor; the deep neural network model based on deep learning is utilized to fit the mapping relation between the time sequence characteristic of the ammonia concentration value in the sewage treatment process and the control of the aerator, and the aerator is adaptively controlled based on the concentration value of the ammonia in the sewage, so that the optimal control of the sewage treatment process is realized, and the treatment efficiency and quality are improved.

Description

Intelligent monitoring system and method for sewage treatment

Technical Field

The application relates to the technical field of intelligent monitoring, in particular to an intelligent monitoring system and an intelligent monitoring method for sewage treatment.

Background

The sewage treatment plant belongs to the high energy consumption industry, and the biological treatment system is taken as a core unit to remove main pollutants in sewage, wherein the energy consumption required by the operation of the anaerobic and aerobic (aeration) treatment systems is more than 50% of the whole sewage plant. Therefore, the aeration system of the sewage treatment plant is optimized and improved, the stability and the energy resource utilization efficiency of the aeration system are improved, and the requirements of standard discharge of effluent of the sewage treatment plant and energy conservation and consumption reduction are met.

In the sewage biological treatment process, the generation and release of gas in the reactor are directly related. The gas generation mechanism can be used for effectively reacting the degradation degree of pollutants by the components and the yield of the gas products in the reactor, and the related data of the detection and analysis of the gas products can be used as important parameters of aeration control.

Most of the traditional sewage treatment aeration control methods currently adopt manual PID control or manual experience modeling control, so that real-time monitoring data and enough data volume are difficult to obtain; the sewage treatment aeration system is a multivariable, strong-coupling and strong-nonlinearity complex system, has the characteristics of uncertainty, time-varying property, time-stagnation property, large inertia and the like, and performs physical, chemical and biological multiphase reactions simultaneously, so that the coupling relation existing among the multivariable complex system is difficult to capture by a method relying on a traditional mathematical model, the complex system which changes in real time cannot be adapted, and the accurate regulation and control of the sewage treatment automatic control system are difficult to realize.

Therefore, an optimized intelligent monitoring scheme for sewage treatment is desired.

Disclosure of Invention

The present application has been made to solve the above-mentioned technical problems. The embodiment of the application provides an intelligent monitoring system for sewage treatment and a method thereof, which are used for acquiring ammonia concentration values at a plurality of preset time points in a preset time period acquired by a gas sensor; the deep neural network model based on deep learning is utilized to fit the mapping relation between the time sequence characteristic of the ammonia concentration value in the sewage treatment process and the control of the aerator, and the aerator is adaptively controlled based on the concentration value of the ammonia in the sewage, so that the optimal control of the sewage treatment process is realized, and the treatment efficiency and quality are improved.

In a first aspect, an intelligent monitoring system for wastewater treatment is provided, comprising:

the ammonia concentration acquisition module is used for acquiring ammonia concentration values at a plurality of preset time points in a preset time period acquired by the gas sensor;

the vectorization module is used for arranging the ammonia concentration values of the plurality of preset time points into ammonia concentration time sequence input vectors according to the time dimension;

the difference module is used for calculating the difference value between the ammonia gas concentration time sequence input vectors of every two adjacent positions in the ammonia gas concentration time sequence input vectors to obtain an ammonia gas concentration time sequence change input vector;

the integration module is used for cascading the ammonia concentration time sequence variation input vector and the ammonia concentration time sequence input vector to obtain an ammonia concentration dynamic and static time sequence input vector;

the characteristic extraction module is used for enabling the ammonia concentration dynamic and static time sequence input vector to pass through a double-pipeline model comprising a first convolutional neural network model and a second convolutional neural network model to obtain a multi-scale ammonia concentration dynamic and static time sequence characteristic vector; and

the monitoring result generation module is used for enabling the multi-scale ammonia concentration dynamic and static time sequence feature vector to pass through a classifier to obtain a classification result, and the classification result is used for indicating whether the aerator is started or not.

In the above-mentioned sewage treatment's intelligent monitoring system, the integration module is used for: cascading the ammonia concentration time sequence variation input vector and the ammonia concentration time sequence input vector by using the following cascading formula to obtain an ammonia concentration dynamic and static time sequence input vector; wherein, the cascade formula is:

V _c ＝Concat[V _a ，V _b ]

wherein V is _a Representing the time sequence change input vector of the ammonia concentration, V _b Representing the ammonia concentration time sequence input vector, concat [. Cndot. ], of]Representing a cascade function, V _c And the dynamic and static time sequence input vector of the ammonia concentration is represented.

In the above-mentioned intelligent monitoring system for sewage treatment, the feature extraction module includes; the first scale feature extraction unit is used for inputting the ammonia concentration dynamic and static time sequence input vector into a first convolution neural network model of the double-pipeline model to obtain a first scale ammonia concentration feature vector, wherein the first convolution neural network model is provided with a one-dimensional convolution kernel of a first scale; the second scale feature extraction unit is used for inputting the ammonia concentration dynamic and static time sequence input vector into a second convolution neural network model of the double-pipeline model to obtain a second scale ammonia concentration feature vector, wherein the second convolution neural network model is provided with a one-dimensional convolution kernel of a second scale, and the first scale is different from the second scale; and the multi-scale fusion unit is used for cascading the first-scale ammonia concentration feature vector and the second-scale ammonia concentration feature vector to obtain the multi-scale ammonia concentration dynamic and static time sequence feature vector.

In the above-mentioned intelligent monitoring system for sewage treatment, the first scale feature extraction unit is configured to: and respectively carrying out convolution processing based on a one-dimensional convolution kernel, mean pooling processing based on a feature matrix and linear activation processing on the ammonia concentration dynamic and static time sequence input vector in forward transfer of layers by using each layer of a first convolution neural network model of the double-pipeline model to take the output of the last layer of the first convolution neural network model as the first-scale ammonia concentration feature vector, wherein the input of the first layer of the first convolution neural network model is the ammonia concentration dynamic and static time sequence input vector.

In the above-mentioned intelligent monitoring system for sewage treatment, the second scale feature extraction unit is configured to: and respectively carrying out convolution processing based on a one-dimensional convolution kernel, mean pooling processing based on a feature matrix and linear activation processing on the ammonia concentration dynamic and static time sequence input vector in forward transfer of layers by using each layer of a second convolution neural network model of the double-pipeline model to ensure that the output of the last layer of the second convolution neural network model is the second-scale ammonia concentration feature vector, wherein the input of the first layer of the second convolution neural network model is the ammonia concentration dynamic and static time sequence input vector.

In the above-mentioned intelligent monitoring system for sewage treatment, the monitoring result generation module includes: the full-connection coding unit is used for carrying out full-connection coding on the multi-scale ammonia concentration dynamic and static time sequence feature vectors by using a plurality of full-connection layers of the classifier so as to obtain coded classification feature vectors; and the classification unit is used for passing the coding classification feature vector through a Softmax classification function of the classifier to obtain the classification result.

The intelligent monitoring system for sewage treatment further comprises a training module for training the double-pipeline model comprising the first convolutional neural network model and the second convolutional neural network model and the classifier; wherein, training module includes: a training ammonia concentration acquisition unit for acquiring training ammonia concentration values at a plurality of predetermined time points within a predetermined period of time acquired by the gas sensor; the training vectorization unit is used for arranging training ammonia concentration values of the plurality of preset time points into training ammonia concentration time sequence input vectors according to the time dimension; the training difference unit is used for calculating the difference value between the training ammonia gas concentration time sequence input vectors of every two adjacent positions in the training ammonia gas concentration time sequence input vectors to obtain training ammonia gas concentration time sequence change input vectors; the training integration unit is used for cascading the training ammonia concentration time sequence variation input vector and the training ammonia concentration time sequence input vector to obtain a training ammonia concentration dynamic and static time sequence input vector; the training feature extraction unit is used for enabling the training ammonia concentration dynamic and static time sequence input vector to pass through a double-pipeline model comprising a first convolutional neural network model and a second convolutional neural network model to obtain a first-scale training ammonia concentration dynamic and static time sequence feature vector and a second-scale training ammonia concentration dynamic and static time sequence feature vector; the multi-scale training fusion module is used for cascading the first-scale training ammonia concentration dynamic and static time sequence feature vector and the second-scale training ammonia concentration dynamic and static time sequence feature vector to obtain a multi-scale training ammonia concentration dynamic and static time sequence feature vector; the training optimization unit is used for calculating the flow refinement loss function value of the first-scale ammonia concentration dynamic and static time sequence feature vector and the second-scale ammonia concentration dynamic and static time sequence feature vector; the classification loss function value calculation unit is used for enabling the multi-scale training ammonia concentration dynamic and static time sequence feature vectors to pass through the classifier to obtain a classification loss function value; and a training unit for calculating a weighted sum of the classification loss function value and the stream refinement loss function value as a loss function value to train the dual-pipeline model including the first convolutional neural network model and the second convolutional neural network model and the classifier.

In the above-mentioned intelligent monitoring system for sewage treatment, the training optimizing unit is used for: calculating the flow refinement loss function value of the first-scale ammonia concentration dynamic and static time sequence feature vector and the second-scale ammonia concentration dynamic and static time sequence feature vector according to the following optimization formula; wherein, the optimization formula is:

wherein V is ₁ Is the dynamic and static time sequence characteristic vector of the first scale ammonia concentration, V ₂ Is the dynamic and static time sequence characteristic vector of the ammonia concentration of the second scale,representing the square of the two norms of the vector, +.>Indicates the subtraction by position, +.]An exponential operation representing a vector representing a calculation of a natural exponential function value raised to a power by a characteristic value of each position in the vector, ">Representing the streaming refinement loss function value.

In a second aspect, an intelligent monitoring method for sewage treatment is provided, which includes:

acquiring ammonia concentration values at a plurality of preset time points in a preset time period acquired by a gas sensor;

arranging the ammonia concentration values at a plurality of preset time points into ammonia concentration time sequence input vectors according to the time dimension;

calculating the difference value between ammonia gas concentration time sequence input vectors of every two adjacent positions in the ammonia gas concentration time sequence input vector to obtain an ammonia gas concentration time sequence change input vector;

Cascading the ammonia concentration time sequence variation input vector and the ammonia concentration time sequence input vector to obtain an ammonia concentration dynamic and static time sequence input vector;

the ammonia concentration dynamic and static time sequence input vector is passed through a double-pipeline model comprising a first convolutional neural network model and a second convolutional neural network model to obtain a multi-scale ammonia concentration dynamic and static time sequence feature vector; and

and the multi-scale ammonia concentration dynamic and static time sequence feature vector passes through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the aerator is started or not.

In the above-mentioned intelligent monitoring method for sewage treatment, cascading the ammonia concentration time sequence variation input vector and the ammonia concentration time sequence input vector to obtain an ammonia concentration dynamic and static time sequence input vector comprises: cascading the ammonia concentration time sequence variation input vector and the ammonia concentration time sequence input vector by using the following cascading formula to obtain an ammonia concentration dynamic and static time sequence input vector; wherein, the cascade formula is:

V _c ＝Concat[V _a ，V _b ]

Compared with the prior art, the intelligent monitoring system and the method for sewage treatment provided by the application have the advantages that the ammonia concentration values of a plurality of preset time points in the preset time period acquired by the gas sensor are obtained; the deep neural network model based on deep learning is utilized to fit the mapping relation between the time sequence characteristic of the ammonia concentration value in the sewage treatment process and the control of the aerator, and the aerator is adaptively controlled based on the concentration value of the ammonia in the sewage, so that the optimal control of the sewage treatment process is realized, and the treatment efficiency and quality are improved.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings 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 only some embodiments of the present application, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.

Fig. 1 is an application scenario diagram of an intelligent monitoring system for sewage treatment according to an embodiment of the present application.

Fig. 2 is a block diagram of an intelligent monitoring system for wastewater treatment according to an embodiment of the present application.

Fig. 3 is a block diagram of the feature extraction module in the intelligent monitoring system for sewage treatment according to an embodiment of the present application.

Fig. 4 is a block diagram of the monitoring result generation module in the intelligent monitoring system for sewage treatment according to the embodiment of the present application.

Fig. 5 is a block diagram of the training module in the intelligent monitoring system for sewage treatment according to an embodiment of the present application.

Fig. 6 is a flow chart of an intelligent monitoring method for sewage treatment according to an embodiment of the present application.

Fig. 7 is a schematic diagram of a system architecture of an intelligent monitoring method for sewage treatment according to an embodiment of the present application.

Detailed Description

The following description of the technical solutions according to the embodiments of the present application will be given with reference to the accompanying drawings in the embodiments of the present application, and it is apparent that the described embodiments are some embodiments of the present application, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.

Unless defined otherwise, all technical and scientific terms used in the embodiments of the application have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in the present application is for the purpose of describing particular embodiments only and is not intended to limit the scope of the present application.

In describing embodiments of the present application, unless otherwise indicated and limited thereto, the term "connected" should be construed broadly, for example, it may be an electrical connection, or may be a communication between two elements, or may be a direct connection, or may be an indirect connection via an intermediate medium, and it will be understood by those skilled in the art that the specific meaning of the term may be interpreted according to circumstances.

It should be noted that, the term "first\second\third" related to the embodiment of the present application is merely to distinguish similar objects, and does not represent a specific order for the objects, it is to be understood that "first\second\third" may interchange a specific order or sequence where allowed. It is to be understood that the "first\second\third" distinguishing objects may be interchanged where appropriate such that embodiments of the application described herein may be practiced in sequences other than those illustrated or described herein.

Aiming at the technical problems, the technical conception of the application is as follows: the nonlinear complex function mapping relation between the ammonia concentration time sequence characteristic and the control of the aerator in the sewage treatment process is fitted by using a deep neural network model based on deep learning, so that the aerator is adaptively controlled based on the concentration value time sequence characteristic of the ammonia in the sewage, thereby realizing the optimal control of the sewage treatment process and improving the treatment efficiency and quality.

Specifically, first, ammonia concentration values at a plurality of predetermined time points within a predetermined period of time acquired by a gas sensor are acquired. Those skilled in the art will appreciate that the concentration of ammonia in the sewage treatment process is one of the important indicators for measuring the sewage treatment effect and quality, so that real-time monitoring of the ammonia concentration is required. The gas sensor can rapidly and accurately collect the ammonia concentration value, the collected data at a plurality of preset time points in a preset time period can reflect the change condition of the gas sensor in the whole time period, and the data can be used for determining the ammonia concentration change trend in the sewage treatment process and providing basis for the control of a follow-up aerator.

And then, arranging the ammonia gas concentration values at the plurality of preset time points into ammonia gas concentration time sequence input vectors according to a time dimension, and calculating the difference value between the ammonia gas concentration values at every two adjacent positions in the ammonia gas concentration time sequence input vectors to obtain ammonia gas concentration time sequence change input vectors. In the sewage treatment process, the concentration value of the ammonia gas changes along with the time, so in the technical scheme of the application, the ammonia gas concentration data at a plurality of preset time points are arranged together according to the time sequence so as to explicitly express the change trend and rule of the ammonia gas concentration from a data source domain. Meanwhile, the change condition of the ammonia concentration value can be represented quantitatively by calculating the difference value between the ammonia concentration values of every two adjacent positions in the ammonia concentration time sequence input vector.

And further, cascading the ammonia concentration time sequence variation input vector and the ammonia concentration time sequence input vector to obtain an ammonia concentration dynamic and static time sequence input vector. The ammonia gas concentration time sequence change input vector and the ammonia gas concentration time sequence input vector are cascaded to obtain an ammonia gas concentration dynamic and static time sequence input vector, so that time sequence change information of ammonia gas and original ammonia gas concentration data are comprehensively considered, and the ammonia gas concentration dynamic and static level information is included. In the sewage treatment process, the dynamic change information of the ammonia concentration reflects the real-time state of the sewage treatment system, and the static level information of the ammonia concentration represents the overall level of sewage treatment quality, which play an important role in the self-adaptive control of the aerator. Therefore, the ammonia concentration time sequence change input vector and the ammonia concentration time sequence input vector are cascaded, so that the ammonia concentration dynamic and static time sequence input vector which more comprehensively and accurately describes the ammonia concentration characteristics in the sewage treatment process can be obtained.

And then, the ammonia concentration dynamic and static time sequence input vector is passed through a double-pipeline model comprising a first convolutional neural network model and a second convolutional neural network model to obtain a multi-scale ammonia concentration dynamic and static time sequence characteristic vector. That is, in the technical scheme of the application, the dual-pipeline model of the first convolutional neural network model and the second convolutional neural network model is used for carrying out multi-scale one-dimensional convolutional coding on the ammonia concentration dynamic and static time sequence input vector so as to capture the correlation pattern characteristics of the ammonia concentration static value and the correlation pattern characteristics of the ammonia concentration variation under different time sequence spaces.

Specifically, by using the first convolutional neural network model and the second convolutional neural network model with one-dimensional convolutional kernels with different scales, dynamic and static characteristics of ammonia concentration can be extracted from different scales, a multi-scale ammonia concentration dynamic and static time sequence characteristic vector is formed, and the change trend and rule of ammonia concentration can be reflected better. Meanwhile, by adopting the structure of the double pipeline model, the mutual influence of the characteristic information of different scales can be avoided, and the accuracy and stability of characteristic extraction are ensured.

And then, the multi-scale ammonia concentration dynamic and static time sequence feature vector is passed through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the aerator is started or not. That is, the classifier is used to determine a class probability tag to which the multi-scale ammonia concentration dynamic and static timing feature vector belongs, the class probability tag including starting the aerator (first tag) and not starting the aerator (second tag). It should be noted that the class probability tag is a control policy tag of the aerator, so that the aerator can be controlled based on the classification result after the classification result is obtained. Therefore, a deep neural network model based on deep learning is utilized to fit a nonlinear complex function mapping relation between ammonia concentration time sequence characteristics and control of an aerator in the sewage treatment process, so that the aerator is adaptively controlled based on concentration value time sequence characteristics of ammonia in sewage, thereby realizing optimal control of the sewage treatment process and improving treatment efficiency and quality.

Particularly, in the technical scheme of the application, when the ammonia concentration dynamic and static time sequence input vector is obtained through a double-pipeline model comprising a first convolutional neural network model and a second convolutional neural network model, a first-scale ammonia concentration dynamic and static time sequence feature vector obtained by the first convolutional neural network model and a second-scale ammonia concentration dynamic and static time sequence feature vector obtained by the second convolutional neural network model are required to be fused to obtain the multi-scale ammonia concentration dynamic and static time sequence feature vector, which is essentially to fuse the sequential time sequence associated feature representations under different scales expressed by the first-scale ammonia concentration dynamic and static time sequence feature vector and the second-scale ammonia concentration dynamic and static time sequence feature vector in a high-dimensional feature space.

Therefore, the applicant of the present application considers that in order to promote the fusion effect of the first-scale ammonia concentration dynamic and static timing feature vector and the second-scale ammonia concentration dynamic and static timing feature vector, it is desirable to promote the correlation between the first-scale ammonia concentration dynamic and static timing feature vector and the second-scale ammonia concentration dynamic and static timing feature vector under the serialized time-series correlation feature representation and the spatial dimension feature representation of the high-dimensional feature space.

Based on the above, the applicant of the present application further introduces a dynamic and static time sequence feature vector V for the first scale ammonia concentration in addition to the classification loss function for the classifier ₁ And the dynamic and static time sequence characteristic vector V of the second scale ammonia concentration ₂ The streaming refinement loss function of (2) is expressed as:

wherein the method comprises the steps ofRepresenting the square of the two norms of the vector.

Here, the stream refinement loss function is based on the first-scale ammonia concentration dynamic and static time sequence feature vector V ₁ And the dynamic and static time sequence characteristic vector V of the second scale ammonia concentration ₂ In the conversion from sequential streaming distribution of time sequence associated features to spatial distribution in high-dimensional feature space, super-resolution improvement of the spatial distribution in the high-dimensional feature space is realized by interpolation under the sequential distribution of vectors synchronously, so that the inter-class probability relation under the balanced sequence is realizedProviding finer alignment for distribution differences in the high-dimensional feature space so as to jointly present cross inter-dimensional (inter-dimensional) context correlation on the sequential time-series correlation feature dimension and the space dimension of the high-dimensional feature space, thereby improving the fusion effect of the multi-scale ammonia concentration dynamic and static time-series feature vector on the first-scale ammonia concentration dynamic and static time-series feature vector and the second-scale ammonia concentration dynamic and static time-series feature vector and improving the accuracy of classification results obtained by the multi-scale ammonia concentration dynamic and static time-series feature vector through a classifier.

Fig. 1 is an application scenario diagram of an intelligent monitoring system for sewage treatment according to an embodiment of the present application. As shown in fig. 1, in this application scenario, first, ammonia concentration values (e.g., C as illustrated in fig. 1) at a plurality of predetermined time points within a predetermined period of time acquired by a gas sensor are acquired; then, the obtained ammonia gas concentration value is input to a server (e.g., S as illustrated in fig. 1) in which an intelligent monitoring algorithm for sewage treatment is deployed, wherein the server is capable of processing the ammonia gas concentration value based on the intelligent monitoring algorithm for sewage treatment to generate a classification result indicating whether to start the aerator.

Having described the basic principles of the present application, various non-limiting embodiments of the present application will now be described in detail with reference to the accompanying drawings.

In one embodiment of the present application, fig. 2 is a block diagram of an intelligent monitoring system for wastewater treatment according to an embodiment of the present application. As shown in fig. 2, a preventive medical inspection system 100 according to an embodiment of the present application includes: an ammonia concentration acquisition module 110 for acquiring ammonia concentration values at a plurality of predetermined time points within a predetermined period of time acquired by the gas sensor; a vectorization module 120, configured to arrange the ammonia concentration values at the plurality of predetermined time points into ammonia concentration time sequence input vectors according to a time dimension; the difference module 130 is configured to calculate a difference value between ammonia concentration time sequence input vectors of every two adjacent positions in the ammonia concentration time sequence input vectors to obtain an ammonia concentration time sequence variation input vector; the integration module 140 is configured to concatenate the ammonia concentration time sequence variation input vector and the ammonia concentration time sequence input vector to obtain an ammonia concentration dynamic and static time sequence input vector; the feature extraction module 150 is configured to pass the ammonia concentration dynamic and static time sequence input vector through a dual-pipeline model including a first convolutional neural network model and a second convolutional neural network model to obtain a multi-scale ammonia concentration dynamic and static time sequence feature vector; and a monitoring result generating module 160, configured to pass the multi-scale ammonia concentration dynamic and static time sequence feature vector through a classifier to obtain a classification result, where the classification result is used to indicate whether to start the aerator.

Specifically, in the embodiment of the present application, the ammonia concentration acquisition module 110 is configured to acquire ammonia concentration values at a plurality of predetermined time points within a predetermined period of time acquired by the gas sensor. Aiming at the technical problems, the technical conception of the application is as follows: the nonlinear complex function mapping relation between the ammonia concentration time sequence characteristic and the control of the aerator in the sewage treatment process is fitted by using a deep neural network model based on deep learning, so that the aerator is adaptively controlled based on the concentration value time sequence characteristic of the ammonia in the sewage, thereby realizing the optimal control of the sewage treatment process and improving the treatment efficiency and quality.

Specifically, in the embodiment of the present application, the vectorizing module 120 and the differentiating module 130 are configured to arrange the ammonia concentration values at the plurality of predetermined time points into an ammonia concentration time sequence input vector according to a time dimension; and the difference value between the ammonia gas concentration time sequence input vectors of every two adjacent positions in the ammonia gas concentration time sequence input vector is calculated to obtain an ammonia gas concentration time sequence change input vector.

Specifically, in the embodiment of the present application, the integrating module 140 is configured to concatenate the ammonia concentration time sequence variation input vector and the ammonia concentration time sequence input vector to obtain an ammonia concentration dynamic and static time sequence input vector. And further, cascading the ammonia concentration time sequence variation input vector and the ammonia concentration time sequence input vector to obtain an ammonia concentration dynamic and static time sequence input vector. The ammonia gas concentration time sequence change input vector and the ammonia gas concentration time sequence input vector are cascaded to obtain an ammonia gas concentration dynamic and static time sequence input vector, so that time sequence change information of ammonia gas and original ammonia gas concentration data are comprehensively considered, and the ammonia gas concentration dynamic and static level information is included.

In the sewage treatment process, the dynamic change information of the ammonia concentration reflects the real-time state of the sewage treatment system, and the static level information of the ammonia concentration represents the overall level of sewage treatment quality, which play an important role in the self-adaptive control of the aerator. Therefore, the ammonia concentration time sequence change input vector and the ammonia concentration time sequence input vector are cascaded, so that the ammonia concentration dynamic and static time sequence input vector which more comprehensively and accurately describes the ammonia concentration characteristics in the sewage treatment process can be obtained.

Wherein, the integration module 140 is configured to: cascading the ammonia concentration time sequence variation input vector and the ammonia concentration time sequence input vector by using the following cascading formula to obtain an ammonia concentration dynamic and static time sequence input vector; wherein, the cascade formula is:

V _c ＝Concat[V _a ，V _b ]

Specifically, in the embodiment of the present application, the feature extraction module 150 is configured to pass the ammonia concentration dynamic and static time sequence input vector through a dual-pipeline model including a first convolutional neural network model and a second convolutional neural network model to obtain a multi-scale ammonia concentration dynamic and static time sequence feature vector. And then, the ammonia concentration dynamic and static time sequence input vector is passed through a double-pipeline model comprising a first convolutional neural network model and a second convolutional neural network model to obtain a multi-scale ammonia concentration dynamic and static time sequence characteristic vector. That is, in the technical scheme of the application, the dual-pipeline model of the first convolutional neural network model and the second convolutional neural network model is used for carrying out multi-scale one-dimensional convolutional coding on the ammonia concentration dynamic and static time sequence input vector so as to capture the correlation pattern characteristics of the ammonia concentration static value and the correlation pattern characteristics of the ammonia concentration variation under different time sequence spaces.

FIG. 3 is a block diagram of the feature extraction module in the intelligent monitoring system for wastewater treatment according to an embodiment of the present application, and as shown in FIG. 3, the feature extraction module 150 includes; a first scale feature extraction unit 151, configured to input the ammonia concentration dynamic and static time sequence input vector into a first convolutional neural network model of the dual-pipeline model to obtain a first scale ammonia concentration feature vector, where the first convolutional neural network model has a one-dimensional convolutional kernel of a first scale; a second scale feature extraction unit 152, configured to input the ammonia concentration dynamic and static time sequence input vector into a second convolutional neural network model of the dual-pipeline model to obtain a second scale ammonia concentration feature vector, where the second convolutional neural network model has a one-dimensional convolution kernel of a second scale, and the first scale is different from the second scale; and a multi-scale fusion unit 153, configured to concatenate the first-scale ammonia concentration feature vector and the second-scale ammonia concentration feature vector to obtain the multi-scale ammonia concentration dynamic and static time sequence feature vector.

It should be noted that the multi-scale neighborhood feature extraction module is essentially a deep neural network model based on deep learning, which is capable of fitting any function by a predetermined training strategy and has a higher feature extraction generalization capability compared to the conventional feature engineering.

The multi-scale neighborhood feature extraction module comprises a plurality of parallel one-dimensional convolution layers, wherein in the process of feature extraction by the multi-scale neighborhood feature extraction module, the plurality of parallel one-dimensional convolution layers respectively carry out one-dimensional convolution coding on input data by one-dimensional convolution check with different scales so as to capture local implicit features of a sequence.

Wherein, the first scale feature extraction unit 151 is configured to: and respectively carrying out convolution processing based on a one-dimensional convolution kernel, mean pooling processing based on a feature matrix and linear activation processing on the ammonia concentration dynamic and static time sequence input vector in forward transfer of layers by using each layer of a first convolution neural network model of the double-pipeline model to take the output of the last layer of the first convolution neural network model as the first-scale ammonia concentration feature vector, wherein the input of the first layer of the first convolution neural network model is the ammonia concentration dynamic and static time sequence input vector.

Further, the second scale feature extraction unit 152 is configured to: and respectively carrying out convolution processing based on a one-dimensional convolution kernel, mean pooling processing based on a feature matrix and linear activation processing on the ammonia concentration dynamic and static time sequence input vector in forward transfer of layers by using each layer of a second convolution neural network model of the double-pipeline model to ensure that the output of the last layer of the second convolution neural network model is the second-scale ammonia concentration feature vector, wherein the input of the first layer of the second convolution neural network model is the ammonia concentration dynamic and static time sequence input vector.

The convolutional neural network (Convolutional Neural Network, CNN) is an artificial neural network and has wide application in the fields of image recognition and the like. The convolutional neural network may include an input layer, a hidden layer, and an output layer, where the hidden layer may include a convolutional layer, a pooling layer, an activation layer, a full connection layer, etc., where the previous layer performs a corresponding operation according to input data, outputs an operation result to the next layer, and obtains a final result after the input initial data is subjected to a multi-layer operation.

The convolutional neural network model has excellent performance in the aspect of image local feature extraction by taking a convolutional kernel as a feature filtering factor, and has stronger feature extraction generalization capability and fitting capability compared with the traditional image feature extraction algorithm based on statistics or feature engineering.

Specifically, in the embodiment of the present application, the monitoring result generating module 160 is configured to pass the multi-scale ammonia concentration dynamic and static time sequence feature vector through a classifier to obtain a classification result, where the classification result is used to indicate whether to start the aerator. And then, the multi-scale ammonia concentration dynamic and static time sequence feature vector is passed through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the aerator is started or not. That is, the classifier is used to determine a class probability tag to which the multi-scale ammonia concentration dynamic and static timing feature vector belongs, the class probability tag including starting the aerator (first tag) and not starting the aerator (second tag).

It should be noted that the class probability tag is a control policy tag of the aerator, so that the aerator can be controlled based on the classification result after the classification result is obtained. Therefore, a deep neural network model based on deep learning is utilized to fit a nonlinear complex function mapping relation between ammonia concentration time sequence characteristics and control of an aerator in the sewage treatment process, so that the aerator is adaptively controlled based on concentration value time sequence characteristics of ammonia in sewage, thereby realizing optimal control of the sewage treatment process and improving treatment efficiency and quality.

It should be noted that the first tag and the second tag do not include the concept of artificial setting, and in fact, during the training process, the computer model does not have the concept of "whether to start the aerator", but only has two classification tags whose output characteristics are the probabilities under the two classification tags, i.e., the sum of p1 and p2 is one. Therefore, the classification result of whether to start the aerator is actually converted into a classification probability distribution conforming to the natural law through classifying the labels, and the physical meaning of the natural probability distribution of the labels is essentially used instead of the language text meaning of whether to start the aerator.

Fig. 4 is a block diagram of the monitoring result generation module in the intelligent monitoring system for sewage treatment according to an embodiment of the present application, as shown in fig. 4, the monitoring result generation module 160 includes: a full-connection encoding unit 161, configured to perform full-connection encoding on the multi-scale ammonia concentration dynamic and static time sequence feature vector by using a plurality of full-connection layers of the classifier to obtain an encoded classification feature vector; and a classification unit 162, configured to pass the encoded classification feature vector through a Softmax classification function of the classifier to obtain the classification result.

Further, the intelligent monitoring system for sewage treatment further comprises a training module for training the double-pipeline model comprising the first convolutional neural network model and the second convolutional neural network model and the classifier; fig. 5 is a block diagram of the training module in the intelligent monitoring system for sewage treatment according to the embodiment of the present application, and as shown in fig. 5, the training module 170 includes: a training ammonia concentration acquisition unit 171 for acquiring training ammonia concentration values at a plurality of predetermined time points within a predetermined period of time acquired by the gas sensor; a training vectorization unit 172, configured to arrange training ammonia concentration values at the plurality of predetermined time points into training ammonia concentration time sequence input vectors according to a time dimension; a training difference unit 173, configured to calculate a difference value between training ammonia concentration time sequence input vectors of every two adjacent positions in the training ammonia concentration time sequence input vectors to obtain a training ammonia concentration time sequence variation input vector; the training integration unit 174 is configured to concatenate the training ammonia concentration time sequence variation input vector and the training ammonia concentration time sequence input vector to obtain a training ammonia concentration dynamic and static time sequence input vector; the training feature extraction unit 175 is configured to pass the training ammonia concentration dynamic and static time sequence input vector through a dual-pipeline model including a first convolutional neural network model and a second convolutional neural network model to obtain a first-scale training ammonia concentration dynamic and static time sequence feature vector and a second-scale training ammonia concentration dynamic and static time sequence feature vector; the multi-scale training fusion module 176 is configured to concatenate the first-scale training ammonia concentration dynamic and static timing feature vector and the second-scale training ammonia concentration dynamic and static timing feature vector to obtain a multi-scale training ammonia concentration dynamic and static timing feature vector; a training optimization unit 177 for calculating a flow refinement loss function value of the first-scale ammonia concentration dynamic and static time sequence feature vector and the second-scale ammonia concentration dynamic and static time sequence feature vector; a classification loss function value calculation unit 178, configured to pass the multi-scale training ammonia concentration dynamic and static time sequence feature vector through the classifier to obtain a classification loss function value; and a training unit 179 for calculating a weighted sum of the classification loss function value and the stream refinement loss function value as a loss function value to train the dual-pipeline model including the first convolutional neural network model and the second convolutional neural network model and the classifier.

Based on the above, the applicant of the present application further introduces a dynamic and static time sequence feature vector V for the first scale ammonia concentration in addition to the classification loss function for the classifier ₁ And the dynamic and static time sequence characteristic vector V of the second scale ammonia concentration ₂ The streaming refinement loss function of (2) is expressed as: calculating the flow refinement loss function value of the first-scale ammonia concentration dynamic and static time sequence feature vector and the second-scale ammonia concentration dynamic and static time sequence feature vector according to the following optimization formula; wherein, the optimization formula is:

wherein V is ₁ Is the dynamic and static time sequence characteristic vector of the first scale ammonia concentration, V ₂ Is the dynamic and static time sequence characteristic vector of the ammonia concentration of the second scale,representing the square of the two norms of the vector, +.>Indicates the subtraction by position, +.]An exponential operation representing a vector representing a calculation of a natural exponential function value raised to a power by a characteristic value of each position in the vector, ">Representing the streaming refinement loss function value. />

Here, the stream refinement loss function is based on the first-scale ammonia concentration dynamic and static time sequence feature vector V ₁ And the dynamic and static time sequence characteristic vector V of the second scale ammonia concentration ₂ In the conversion from the sequential streaming distribution of the time sequence associated features to the space distribution in the high-dimensional feature space, the super-resolution improvement of the space distribution in the high-dimensional feature space is realized by synchronously carrying out interpolation under the sequence distribution of vectors, so that finer alignment is provided for the distribution difference in the high-dimensional feature space through the inter-class probability relation under the balanced sequence, and cross inter-dimensional (inter-dimensional) context association is jointly presented on the sequential time sequence associated feature dimension and the space dimension of the high-dimensional feature space, thereby improving the fusion effect of the multi-scale ammonia concentration dynamic and static time sequence feature vector on the first-scale ammonia concentration dynamic and static time sequence feature vector and the second-scale ammonia concentration dynamic and static time sequence feature vector, and improving the accuracy of the classification result obtained by the classifier of the multi-scale ammonia concentration dynamic and static time sequence feature vector.

In summary, an intelligent monitoring system 100 for wastewater treatment according to an embodiment of the present application is illustrated, which acquires ammonia concentration values at a plurality of predetermined time points within a predetermined period of time acquired by a gas sensor; the deep neural network model based on deep learning is utilized to fit the mapping relation between the time sequence characteristic of the ammonia concentration value in the sewage treatment process and the control of the aerator, and the aerator is adaptively controlled based on the concentration value of the ammonia in the sewage, so that the optimal control of the sewage treatment process is realized, and the treatment efficiency and quality are improved.

As described above, the intelligent monitoring system 100 for sewage treatment according to the embodiment of the present application may be implemented in various terminal devices, such as a server for intelligent monitoring of sewage treatment, and the like. In one example, the intelligent monitoring system 100 for wastewater treatment according to an embodiment of the present application may be integrated into a terminal device as a software module and/or hardware module. For example, the intelligent monitoring system 100 for wastewater treatment may be a software module in the operating system of the terminal device, or may be an application developed for the terminal device; of course, the intelligent monitoring system 100 for wastewater treatment can also be one of a plurality of hardware modules of the terminal device.

Alternatively, in another example, the intelligent monitoring system for wastewater treatment 100 and the terminal device may be separate devices, and the intelligent monitoring system for wastewater treatment 100 may be connected to the terminal device through a wired and/or wireless network and transmit interactive information in a agreed data format.

In one embodiment of the present application, fig. 6 is a flow chart of an intelligent monitoring method for wastewater treatment according to an embodiment of the present application. As shown in fig. 6, an intelligent monitoring method for sewage treatment according to an embodiment of the present application includes: 210, acquiring ammonia concentration values at a plurality of preset time points in a preset time period acquired by a gas sensor; 220, arranging the ammonia concentration values at the plurality of preset time points into ammonia concentration time sequence input vectors according to a time dimension; 230, calculating a difference value between ammonia gas concentration time sequence input vectors of every two adjacent positions in the ammonia gas concentration time sequence input vector to obtain an ammonia gas concentration time sequence change input vector; 240, cascading the ammonia concentration time sequence variation input vector and the ammonia concentration time sequence input vector to obtain an ammonia concentration dynamic and static time sequence input vector; 250, passing the ammonia concentration dynamic and static time sequence input vector through a double-pipeline model comprising a first convolutional neural network model and a second convolutional neural network model to obtain a multi-scale ammonia concentration dynamic and static time sequence feature vector; and 260, passing the multi-scale ammonia concentration dynamic and static time sequence feature vector through a classifier to obtain a classification result, wherein the classification result is used for indicating whether to start the aerator.

Fig. 7 is a schematic diagram of a system architecture of an intelligent monitoring method for sewage treatment according to an embodiment of the present application. As shown in fig. 7, in the system architecture of the intelligent monitoring method of wastewater treatment, first, ammonia concentration values at a plurality of predetermined time points within a predetermined period of time acquired by a gas sensor are acquired; then, arranging the ammonia concentration values at a plurality of preset time points into ammonia concentration time sequence input vectors according to a time dimension; then, calculating the difference value between ammonia gas concentration time sequence input vectors of every two adjacent positions in the ammonia gas concentration time sequence input vector to obtain an ammonia gas concentration time sequence change input vector; then, cascading the ammonia concentration time sequence variation input vector and the ammonia concentration time sequence input vector to obtain an ammonia concentration dynamic and static time sequence input vector; then, the ammonia concentration dynamic and static time sequence input vector is passed through a double-pipeline model comprising a first convolutional neural network model and a second convolutional neural network model to obtain a multi-scale ammonia concentration dynamic and static time sequence feature vector; and finally, the multi-scale ammonia concentration dynamic and static time sequence feature vector is passed through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the aerator is started or not.

In a specific example, in the above intelligent monitoring method for sewage treatment, cascading the ammonia concentration time sequence variation input vector and the ammonia concentration time sequence input vector to obtain an ammonia concentration dynamic and static time sequence input vector includes: cascading the ammonia concentration time sequence variation input vector and the ammonia concentration time sequence input vector by using the following cascading formula to obtain an ammonia concentration dynamic and static time sequence input vector; wherein, the cascade formula is:

V _c ＝Concat[V _a ，V _b ]

In a specific example, in the intelligent monitoring method for sewage treatment, the ammonia concentration dynamic and static time sequence input vector is passed through a double-pipeline model comprising a first convolutional neural network model and a second convolutional neural network model to obtain a multi-scale ammonia concentration dynamic and static time sequence feature vector, which comprises the following steps of; inputting the ammonia concentration dynamic and static time sequence input vector into a first convolution neural network model of the double-pipeline model to obtain a first-scale ammonia concentration feature vector, wherein the first convolution neural network model is provided with a one-dimensional convolution kernel of a first scale; inputting the ammonia concentration dynamic and static time sequence input vector into a second convolution neural network model of the double-pipeline model to obtain a second-scale ammonia concentration feature vector, wherein the second convolution neural network model is provided with a one-dimensional convolution kernel of a second scale, and the first scale is different from the second scale; and cascading the first-scale ammonia concentration feature vector and the second-scale ammonia concentration feature vector to obtain the multi-scale ammonia concentration dynamic and static time sequence feature vector.

In a specific example, in the above intelligent monitoring method for sewage treatment, the ammonia concentration dynamic and static time sequence input vector is input into a first convolutional neural network model of the dual-pipeline model to obtain a first scale ammonia concentration feature vector, where the first convolutional neural network model has a one-dimensional convolutional kernel of a first scale, and the method includes: and respectively carrying out convolution processing based on a one-dimensional convolution kernel, mean pooling processing based on a feature matrix and linear activation processing on the ammonia concentration dynamic and static time sequence input vector in forward transfer of layers by using each layer of a first convolution neural network model of the double-pipeline model to take the output of the last layer of the first convolution neural network model as the first-scale ammonia concentration feature vector, wherein the input of the first layer of the first convolution neural network model is the ammonia concentration dynamic and static time sequence input vector.

In a specific example, in the above intelligent monitoring method for sewage treatment, the inputting the ammonia concentration dynamic and static time sequence input vector into the second convolutional neural network model of the dual-pipeline model to obtain a second scale ammonia concentration feature vector, where the second convolutional neural network model has a one-dimensional convolutional kernel of a second scale, and the first scale is different from the second scale, and includes: and respectively carrying out convolution processing based on a one-dimensional convolution kernel, mean pooling processing based on a feature matrix and linear activation processing on the ammonia concentration dynamic and static time sequence input vector in forward transfer of layers by using each layer of a second convolution neural network model of the double-pipeline model to ensure that the output of the last layer of the second convolution neural network model is the second-scale ammonia concentration feature vector, wherein the input of the first layer of the second convolution neural network model is the ammonia concentration dynamic and static time sequence input vector.

In a specific example, in the above intelligent monitoring method for sewage treatment, the classifying method includes passing the multi-scale ammonia concentration dynamic and static time sequence feature vector through a classifier to obtain a classification result, where the classification result is used to indicate whether to start the aerator, and includes: performing full-connection coding on the multi-scale ammonia concentration dynamic and static time sequence feature vectors by using a plurality of full-connection layers of the classifier to obtain coded classification feature vectors; and passing the coding classification feature vector through a Softmax classification function of the classifier to obtain the classification result.

In a specific example, in the intelligent monitoring method for sewage treatment, training the double-pipeline model including the first convolutional neural network model and the second convolutional neural network model and the classifier is further included; wherein the training the dual-pipeline model including the first convolutional neural network model and the second convolutional neural network model and the classifier comprises: acquiring training ammonia concentration values at a plurality of preset time points in a preset time period acquired by a gas sensor; arranging the training ammonia concentration values at a plurality of preset time points into training ammonia concentration time sequence input vectors according to the time dimension; calculating the difference value between the training ammonia gas concentration time sequence input vectors of every two adjacent positions in the training ammonia gas concentration time sequence input vectors to obtain training ammonia gas concentration time sequence change input vectors; cascading the training ammonia concentration time sequence variation input vector and the training ammonia concentration time sequence input vector to obtain a training ammonia concentration dynamic and static time sequence input vector; the training ammonia concentration dynamic and static time sequence input vector is passed through a double-pipeline model comprising a first convolutional neural network model and a second convolutional neural network model to obtain a first-scale training ammonia concentration dynamic and static time sequence feature vector and a second-scale training ammonia concentration dynamic and static time sequence feature vector; cascading the first scale training ammonia concentration dynamic and static time sequence feature vector and the second scale training ammonia concentration dynamic and static time sequence feature vector to obtain a multi-scale training ammonia concentration dynamic and static time sequence feature vector; calculating flow type refinement loss function values of the first-scale ammonia concentration dynamic and static time sequence feature vector and the second-scale ammonia concentration dynamic and static time sequence feature vector; the multi-scale training ammonia concentration dynamic and static time sequence feature vector is passed through the classifier to obtain a classification loss function value; and a training module that calculates a weighted sum of the classification loss function value and the streaming refinement loss function value as a loss function value to train the dual-pipeline model including the first convolutional neural network model and the second convolutional neural network model and the classifier.

In a specific example, in the above intelligent monitoring method for sewage treatment, calculating a flow refinement loss function value of the first-scale ammonia concentration dynamic and static time sequence feature vector and the second-scale ammonia concentration dynamic and static time sequence feature vector includes: calculating the flow refinement loss function value of the first-scale ammonia concentration dynamic and static time sequence feature vector and the second-scale ammonia concentration dynamic and static time sequence feature vector according to the following optimization formula; wherein, the optimization formula is:

It will be appreciated by those skilled in the art that the specific operation of each step in the above-described intelligent monitoring method for sewage treatment has been described in detail in the above description of the intelligent monitoring system for sewage treatment with reference to fig. 1 to 5, and thus, repetitive description thereof will be omitted.

The present application also provides a computer program product comprising instructions which, when executed, cause an apparatus to perform operations corresponding to the above-described method.

In one embodiment of the present application, there is also provided a computer-readable storage medium storing a computer program for executing the above-described method.

It should be appreciated that embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the forms of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects may be utilized. Furthermore, the computer program product may take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.

Methods, systems, and computer program products of embodiments of the present application are described in the flow diagrams and/or block diagrams. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or 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, embedded processor, 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 specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory 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 memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These 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 steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

The basic principles of the present application have been described above in connection with specific embodiments, however, it should be noted that the advantages, benefits, effects, etc. mentioned in the present application are merely examples and not intended to be limiting, and these advantages, benefits, effects, etc. are not to be considered as essential to the various embodiments of the present application. Furthermore, the specific details disclosed herein are for purposes of illustration and understanding only, and are not intended to be limiting, as the application is not necessarily limited to practice with the above described specific details.

The block diagrams of the devices, apparatuses, devices, systems referred to in the present application are only illustrative examples and are not intended to require or imply that the connections, arrangements, configurations must be made in the manner shown in the block diagrams. As will be appreciated by one of skill in the art, the devices, apparatuses, devices, systems may be connected, arranged, configured in any manner. Words such as "including," "comprising," "having," and the like are words of openness and mean "including but not limited to," and are used interchangeably therewith. The terms "or" and "as used herein refer to and are used interchangeably with the term" and/or "unless the context clearly indicates otherwise. The term "such as" as used herein refers to, and is used interchangeably with, the phrase "such as, but not limited to.

It is also noted that in the apparatus, devices and methods of the present application, the components or steps may be disassembled and/or assembled. Such decomposition and/or recombination should be considered as equivalent aspects of the present application.

The previous description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other aspects without departing from the scope of the application. Thus, the present application is not intended to be limited to the aspects shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Finally, it is further noted that relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or terminal that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or terminal. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article or terminal device comprising the element.

The foregoing description has been presented for purposes of illustration and description. Furthermore, this description is not intended to limit embodiments of the application to the form disclosed herein. Although a number of example aspects and embodiments have been discussed above, a person of ordinary skill in the art will recognize certain variations, modifications, alterations, additions, and subcombinations thereof.

Claims

1. An intelligent monitoring system for sewage treatment, comprising:

2. The intelligent monitoring system for wastewater treatment according to claim 1, wherein the integration module is configured to: cascading the ammonia concentration time sequence variation input vector and the ammonia concentration time sequence input vector by using the following cascading formula to obtain an ammonia concentration dynamic and static time sequence input vector;

wherein, the cascade formula is:

V _c ＝Concat[V _a ，V _b ]

3. The intelligent monitoring system for wastewater treatment according to claim 2, wherein the feature extraction module comprises;

the first scale feature extraction unit is used for inputting the ammonia concentration dynamic and static time sequence input vector into a first convolution neural network model of the double-pipeline model to obtain a first scale ammonia concentration feature vector, wherein the first convolution neural network model is provided with a one-dimensional convolution kernel of a first scale;

the second scale feature extraction unit is used for inputting the ammonia concentration dynamic and static time sequence input vector into a second convolution neural network model of the double-pipeline model to obtain a second scale ammonia concentration feature vector, wherein the second convolution neural network model is provided with a one-dimensional convolution kernel of a second scale, and the first scale is different from the second scale; and

And the multi-scale fusion unit is used for cascading the first-scale ammonia concentration characteristic vector and the second-scale ammonia concentration characteristic vector to obtain the multi-scale ammonia concentration dynamic and static time sequence characteristic vector.

4. The intelligent monitoring system for wastewater treatment according to claim 3, wherein the first scale feature extraction unit is configured to: and respectively carrying out convolution processing based on a one-dimensional convolution kernel, mean pooling processing based on a feature matrix and linear activation processing on the ammonia concentration dynamic and static time sequence input vector in forward transfer of layers by using each layer of a first convolution neural network model of the double-pipeline model to take the output of the last layer of the first convolution neural network model as the first-scale ammonia concentration feature vector, wherein the input of the first layer of the first convolution neural network model is the ammonia concentration dynamic and static time sequence input vector.

5. The intelligent monitoring system for wastewater treatment according to claim 4, wherein the second scale feature extraction unit is configured to: and respectively carrying out convolution processing based on a one-dimensional convolution kernel, mean pooling processing based on a feature matrix and linear activation processing on the ammonia concentration dynamic and static time sequence input vector in forward transfer of layers by using each layer of a second convolution neural network model of the double-pipeline model to ensure that the output of the last layer of the second convolution neural network model is the second-scale ammonia concentration feature vector, wherein the input of the first layer of the second convolution neural network model is the ammonia concentration dynamic and static time sequence input vector.

6. The intelligent monitoring system for wastewater treatment according to claim 5, wherein the monitoring result generation module comprises:

the full-connection coding unit is used for carrying out full-connection coding on the multi-scale ammonia concentration dynamic and static time sequence feature vectors by using a plurality of full-connection layers of the classifier so as to obtain coded classification feature vectors; and

and the classification unit is used for passing the coding classification feature vector through a Softmax classification function of the classifier to obtain the classification result.

7. The intelligent monitoring system for wastewater treatment of claim 6, further comprising a training module for training the dual-pipeline model comprising the first convolutional neural network model and the second convolutional neural network model and the classifier;

wherein, training module includes:

a training ammonia concentration acquisition unit for acquiring training ammonia concentration values at a plurality of predetermined time points within a predetermined period of time acquired by the gas sensor;

the training vectorization unit is used for arranging training ammonia concentration values of the plurality of preset time points into training ammonia concentration time sequence input vectors according to the time dimension;

the training difference unit is used for calculating the difference value between the training ammonia gas concentration time sequence input vectors of every two adjacent positions in the training ammonia gas concentration time sequence input vectors to obtain training ammonia gas concentration time sequence change input vectors;

The training integration unit is used for cascading the training ammonia concentration time sequence variation input vector and the training ammonia concentration time sequence input vector to obtain a training ammonia concentration dynamic and static time sequence input vector;

the training feature extraction unit is used for enabling the training ammonia concentration dynamic and static time sequence input vector to pass through a double-pipeline model comprising a first convolutional neural network model and a second convolutional neural network model to obtain a first-scale training ammonia concentration dynamic and static time sequence feature vector and a second-scale training ammonia concentration dynamic and static time sequence feature vector;

the multi-scale training fusion module is used for cascading the first-scale training ammonia concentration dynamic and static time sequence feature vector and the second-scale training ammonia concentration dynamic and static time sequence feature vector to obtain a multi-scale training ammonia concentration dynamic and static time sequence feature vector;

the training optimization unit is used for calculating the flow refinement loss function value of the first-scale ammonia concentration dynamic and static time sequence feature vector and the second-scale ammonia concentration dynamic and static time sequence feature vector;

the classification loss function value calculation unit is used for enabling the multi-scale training ammonia concentration dynamic and static time sequence feature vectors to pass through the classifier to obtain a classification loss function value; and

And the training unit is used for calculating a weighted sum of the classified loss function value and the streaming refinement loss function value as a loss function value to train the double-pipeline model comprising the first convolutional neural network model and the second convolutional neural network model and the classifier.

8. The intelligent monitoring system for wastewater treatment according to claim 7, wherein the training optimization unit is configured to: calculating the flow refinement loss function value of the first-scale ammonia concentration dynamic and static time sequence feature vector and the second-scale ammonia concentration dynamic and static time sequence feature vector according to the following optimization formula;

wherein, the optimization formula is:

9. An intelligent monitoring method for sewage treatment, which is characterized by comprising the following steps:

10. The intelligent monitoring method for sewage treatment according to claim 9, wherein cascading the ammonia concentration time sequence variation input vector and the ammonia concentration time sequence input vector to obtain an ammonia concentration dynamic and static time sequence input vector comprises: cascading the ammonia concentration time sequence variation input vector and the ammonia concentration time sequence input vector by using the following cascading formula to obtain an ammonia concentration dynamic and static time sequence input vector;

Wherein, the cascade formula is:

V _c ＝Concat[V _a ，V _b ]

wherein V is _a Representing the time sequence change input vector of the ammonia concentration, V _b Represents the ammonia concentration time sequence input vector, cobcat [. Cndot. ], of]Representing a cascade function, V _c And the dynamic and static time sequence input vector of the ammonia concentration is represented.