CN117849302A - Multi-parameter water quality on-line monitoring method - Google Patents

Abstract

The invention relates to a multi-parameter water quality on-line monitoring method, which comprises the following steps: acquiring multiple parameters of a water sample acquired by an acquisition system and time stamps corresponding to the multiple parameters of the water sample, and acquiring a multivariate time sequence data matrix based on the multiple parameters of the water sample and the time stamps corresponding to the multiple parameters of the water sample; analyzing the multivariate time sequence data matrix through a preset CRNN model to obtain each analysis result of the water sample; inputting each analysis result of the water sample into a preset detection model for detection, correspondingly obtaining each detection result of the water sample, and judging whether each detection result of the water sample is in a preset range of a standard water sample result; if any detection result is not in the preset range of the standard water sample result, triggering an alarm system; the automatic processing flow from parameter analysis to alarm triggering is realized, manual intervention is reduced, and efficiency is improved.

Description

A multi-parameter water quality online monitoring method

Technical Field

The present invention relates to the technical field of water quality detection, and in particular to a multi-parameter water quality online monitoring method, device, equipment and storage medium.

Background technique

Today's society has increasingly higher requirements for the quality of water resources, and water quality monitoring has become an important environmental protection measure. Traditional water quality testing methods usually rely on laboratory chemical analysis, which is not only time-consuming and labor-intensive, but also cannot provide real-time water quality information. However, with the rapid development of information technology, especially the application and promotion of Internet of Things technology, it has become possible to achieve online monitoring of water quality. Online monitoring can accurately monitor water quality changes in real time, which is of great significance to protecting water resources and preventing water pollution accidents. In the prior art, although there are some online monitoring systems, there are still some shortcomings, such as low accuracy, insufficient real-time performance, and inability to perform comprehensive interpretation of complex parameters.

Summary of the invention

The main purpose of the present invention is to provide a multi-parameter water quality online monitoring method, device, equipment and storage medium, which makes the complex parameter analysis more accurate through the application of deep learning CRNN model.

To achieve the above object, the present invention provides a multi-parameter water quality online monitoring method, comprising the following steps:

Acquire multiple parameters of the water sample collected by the collection system and timestamps corresponding to the multiple parameters of the water sample, and obtain a multivariate time series data matrix based on the multiple parameters of the water sample and the timestamps corresponding to the multiple parameters of the water sample;

The multivariate time series data matrix is analyzed by a preset CRNN model to obtain various analysis results of the water sample; wherein the CRNN model includes a CNN layer and a GRU layer; the CNN layer includes a double-layer one-dimensional convolutional recurrent neural network and an activation function; the activation function is respectively connected to multiple GRU modules in a single-layer GRU layer;

Input each analysis result of the water sample into a preset detection model for detection, obtain each detection result of the water sample accordingly, and judge whether each detection result of the water sample is within the preset range of the standard water sample result;

If any of the test results is not within the preset range of the standard water sample results, an alarm system is triggered.

As a further solution of the present invention, multiple parameters of the water sample collected by the collection system are obtained, and a multivariate time series data matrix is obtained based on the multiple parameters of the water sample and the timestamps corresponding to the multiple parameters of the water sample, including:

A preset collection system is used to collect multiple parameters of the water sample and timestamps corresponding to the multiple parameters of the water sample; wherein the multiple parameters of the water sample include temperature, pH, dissolved oxygen, and turbidity of the water sample; the timestamps corresponding to the multiple parameters of the water sample include temperature timestamp, pH timestamp, dissolved oxygen timestamp, and turbidity timestamp;

Mapping and aligning the temperature, pH, dissolved oxygen, and turbidity of the water sample with the temperature timestamp, pH timestamp, dissolved oxygen timestamp, and turbidity timestamp of the water sample to obtain an initial multivariate time series data matrix;

Inputting the initial multivariate time series data matrix into a preset time window model for partitioning to obtain a first target multivariate time series data matrix;

Perform matrix judgment on the first target multivariate time series data matrix to determine whether there are missing elements in the first target multivariate time series data matrix; if there are missing elements in the first target multivariate time series data matrix, perform median calculation on the multiple parameters of the water sample and the timestamps corresponding to the multiple parameters of the water sample to obtain the elements to be supplemented, and fill the elements to be supplemented into the positions of the missing elements to obtain the second target multivariate time series data matrix; wherein, the second target multivariate time series data matrix is used as the multivariate time series data matrix.

As a further solution of the present invention, the multivariate time series data matrix is analyzed by a preset CRNN model to obtain various analysis results of the water sample, including:

Performing data preprocessing on the multivariate time series data matrix through an input layer in the CRNN model to obtain a preprocessing matrix;

Performing data compression on the preprocessing matrix through a double-layer one-dimensional convolutional recurrent neural network to obtain a compressed matrix;

Performing data feature extraction on the compression matrix to obtain a compression extraction matrix;

Performing nonlinear changes on the compression extraction matrix through an activation function to obtain a target matrix of water quality;

The target matrix of the water quality is subjected to matrix analysis by multiple GRU modules to obtain various analysis results of the water samples.

As a further solution of the present invention, each analysis result of the water sample is input into a preset detection model for detection, and correspondingly each detection result of the water sample is obtained, including:

Inputting the analysis results of the water samples into a preset detection model for detection to obtain the corresponding detection values of the water samples;

Calculating the detection value of each water sample respectively to obtain the classification weight corresponding to each water sample;

The classification weights corresponding to each of the water samples are weighted and summed through the attention mechanism to obtain a preliminary target detection result;

Determining whether there are repeated target detection results in the preliminary target detection results;

If the preliminary target detection result contains repeated target detection results, the NMS algorithm is used to eliminate the repeated target detection results to obtain a standard target detection result.

As a further solution of the present invention, each analysis result of the water sample is input into a preset detection model for detection, and corresponding detection values of each water sample are obtained, including:

Performing feature extraction on each analysis result of the water sample to obtain each feature data of the water sample; wherein each feature data of the water sample includes a temperature value, a pH value, a dissolved oxygen value, and a turbidity value of the water sample;

Input each characteristic data of the water sample into a preset detection model for detection, and determine whether there is any abnormal data in each characteristic data of the water sample; wherein the abnormal data includes abnormalities in each characteristic data of the water sample and missing of each characteristic data;

If there are abnormal data in each characteristic data of the water sample, the abnormal data is replaced to obtain standard test data;

The standard test data are sorted by a preset test value sequence model to obtain corresponding test value sequences of each water sample; wherein the test value sequences of each water sample are the test results of each water sample.

The present invention also provides a multi-parameter water quality online monitoring device, comprising:

An acquisition module, used for acquiring multiple parameters of the water sample collected from the acquisition system and timestamps corresponding to the multiple parameters of the water sample, and obtaining a multivariate time series data matrix based on the multiple parameters of the water sample and the timestamps corresponding to the multiple parameters of the water sample;

An analysis module, used to analyze the multivariate time series data matrix through a preset CRNN model to obtain various analysis results of the water sample; wherein the CRNN model includes a CNN layer and a GRU layer; the CNN layer includes a double-layer one-dimensional convolutional recurrent neural network and an activation function; the activation function is respectively connected to multiple GRU modules in a single-layer GRU layer;

The detection module is used to input each analysis result of the water sample into a preset detection model for detection, obtain each detection result of the water sample accordingly, and determine whether each detection result of the water sample is within a preset range of the standard water sample result;

The judgment module is used to trigger an alarm system if any of the test results is not within a preset range of the standard water sample results.

The present invention also provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of any one of the above methods when executing the computer program.

The present invention also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of any of the above-mentioned methods are implemented.

The multi-parameter water quality online monitoring method, device, computer equipment and storage medium provided by the present invention include the following steps: obtaining multiple parameters of a water sample collected from a collection system and timestamps corresponding to the multiple parameters of the water sample, and obtaining a multivariate time series data matrix based on the multiple parameters of the water sample and the timestamps corresponding to the multiple parameters of the water sample; analyzing the multivariate time series data matrix through a preset CRNN model to obtain various analysis results of the water sample; wherein the CRNN model includes a CNN layer and a GRU layer; the CNN layer includes a double-layer one-dimensional convolutional recurrent neural network and an activation function; the activation function is combined with a single-layer GRU layer to obtain a multivariate time series data matrix; the multivariate time series data matrix is analyzed by ...; the multivariate time series data matrix is analyzed by a preset CRNN model; the multivariate time series data matrix is analyzed by a preset CRNN model; the multivariate time series data matrix is analyzed by a preset CRNN model; the multivariate time series data matrix is analyzed by a preset CRNN model; the multivariate time series data matrix is analyzed by a preset CRNN model; the multivariate time series data matrix is analyzed by a preset CRNN model; the multivariate time series data matrix is analyzed by a preset CRNN model; the multi Multiple GRU modules in the RU layer are connected respectively; each analysis result of the water sample is input into the preset detection model for detection, and the corresponding detection results of the water sample are obtained, and it is determined whether each detection result of the water sample is within the preset range of the standard water sample result; if any of the detection results is not within the preset range of the standard water sample result, the alarm system is triggered; by using the preset deep learning CRNN model to analyze the collected parameters, accurate multi-parameter analysis results are obtained, which solves the technical problems of low accuracy and insufficient real-time performance, and realizes the real-time analysis of continuously collected water samples and rapid feedback of water quality conditions. The application of the deep learning CRNN model makes complex parameter analysis more accurate and realizes the automated processing flow from parameter analysis to alarm triggering, reducing manual intervention and improving efficiency.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1 is a schematic diagram of the steps of a multi-parameter water quality online monitoring method according to an embodiment of the present invention;

2 is a block diagram of a multi-parameter water quality online monitoring device according to an embodiment of the present invention;

FIG. 3 is a schematic block diagram of the structure of a computer device according to an embodiment of the present invention.

The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings.

Detailed ways

In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

As shown in FIG. 1 , FIG. 1 is a schematic diagram of the steps of a multi-parameter water quality online monitoring method according to an embodiment of the present invention;

In one embodiment of the present invention, a multi-parameter water quality online monitoring method is provided, comprising the following steps:

In step S1, multiple parameters of water samples collected from a collection system and timestamps corresponding to the multiple parameters of the water samples are obtained, and a multivariate time series data matrix is obtained based on the multiple parameters of the water samples and timestamps corresponding to the multiple parameters of the water samples.

Specifically, use sensors or other collection devices to collect water samples from specific water bodies. Measure and record a variety of water quality parameters, such as temperature, pH, dissolved oxygen, turbidity, chemical oxygen demand (COD), biological oxygen demand (BOD), heavy metal content, etc. Each time data is collected, the specific time of data collection is also recorded, and a timestamp is associated with each parameter value. Check the completeness and accuracy of the data to exclude errors or outliers. Normalize parameters of different dimensions or magnitudes so that they are on the same scale for subsequent analysis. Ensure that the timestamps of all parameters are aligned, handle missing values or interpolate to fill in the gaps in the time series. Align the time series data of each parameter according to the timestamp and construct a matrix in which each row represents a time point and each column represents a parameter. If there are N parameters and T time points, the resulting matrix dimension will be T×N. Use statistical methods or machine learning algorithms to analyze multivariate time series data to identify patterns, trends, and periodicities in the data. Train models based on time series data to predict future changes in water quality parameters. Monitor the changes of water quality parameters in real time and detect abnormal situations, such as pollution incidents, in a timely manner; wherein the collection system is a real-time online collection system.

The above steps can achieve the following technical effects: Real-time and accurate monitoring of water quality status, timely detection of water quality problems. Provide data support for water resource management and decision-making, and predict water quality trends based on historical and real-time data. Take measures to reduce environmental damage through early detection of pollution and abnormal events. Optimize the use and treatment of water resources, improve water treatment efficiency and sustainable use of water resources. The multivariate time series data matrix obtained by the acquisition system can not only be used to monitor and evaluate water quality conditions, but also support a wider range of water resource management and environmental protection activities.

In step S2, the multivariate time series data matrix is analyzed by a preset CRNN model to obtain various analysis results of the water samples; wherein the CRNN model includes a CNN layer and a GRU layer; the CNN layer includes a double-layer one-dimensional convolutional recurrent neural network and an activation function; the activation function is respectively connected to multiple GRU modules in a single-layer GRU layer.

Specifically, the pre-built CRNN model is pre-trained and is specifically designed for analyzing time series data. The CRNN model combines the advantages of convolutional neural networks (CNNs) and recurrent neural networks (RNNs). The multivariate time series data matrix is first passed through a two-layer one-dimensional convolutional neural network. The CNN layer can effectively extract spatial features in the time series data (such as the association between different parameters). The convolution layer is usually followed by an activation function, such as ReLU or Sigmoid, which is used to increase the nonlinearity of the network and help the model learn complex data patterns. The data processed by the CNN is then passed to a single-layer GRU (Gated Recurrent Unit) layer. GRU is a type of RNN that is well suited for processing time series data and can capture the dynamic changes of data over time. The spatial features extracted by the CNN layer are combined with the temporal features captured by the GRU layer. The CRNN model comprehensively analyzes this information and outputs the various analysis results for the water sample.

The following technical effects can be obtained through the above steps: Through the combination of CNN and GRU, the preset CRNN model can efficiently extract the spatial and temporal features in the water quality data, improving the accuracy of the analysis. The GRU layer is particularly suitable for processing time series data, and can better understand and predict data patterns that change over time, such as periodic changes in water quality. The analysis of combined spatial and temporal features enables the model to more accurately predict future changes in water quality, which is very helpful for early warning and risk assessment. The CRNN model can handle multivariate and large amounts of data, and is suitable for complex and high-dimensional environmental monitoring data. In general, the use of the CRNN model in step S2 to analyze the multivariate time series data matrix can improve the accuracy and efficiency of water quality analysis, which is of great significance for environmental monitoring and management.

In step S3, each analysis result of the water sample is input into a preset detection model for detection, and corresponding detection results of the water sample are obtained, and it is determined whether each detection result of the water sample is within a preset range of the standard water sample result.

Specifically, each analysis result of the water sample obtained in step S2 (such as the water quality parameters obtained by the CRNN model analysis) is input into one or more preset detection models. The preset detection model is trained based on different machine learning algorithms for specific water quality assessment tasks. The detection model processes and analyzes the input data and outputs the corresponding test results. Including the binary classification problem of whether the water quality parameters meet the standards, or multi-classification or regression analysis of the specific degree of pollution. According to the test results, it is determined whether each analysis result falls within the preset standard water sample result range. The above range is set based on regulations, industry standards or experience. All the test results are combined to evaluate the overall water quality. If the test results show that some parameters are not within the standard range, the reasons can be further analyzed and corresponding improvement measures can be taken.

Through the above steps, the following technical effects can be achieved: the quality of water samples can be accurately evaluated and monitored to ensure that water quality is safe and up to standard. By automating the analysis and testing process, the errors of manual operation are reduced, the processing efficiency is improved, and the standardization of the analysis process is achieved. If the test results show that the water quality parameters do not meet the standards, timely measures can be taken, such as warning relevant personnel, adjusting the treatment process, etc., so as to effectively manage water quality risks. Through this series of steps, it can play an important role in maintaining public health and environmental quality, especially in the fields of water resources management, drinking water safety monitoring and environmental protection.

In step S4, if any of the test results is not within the preset range of the standard water sample results, an alarm system is triggered.

Specifically, the results of all water sample tests are comprehensively evaluated and compared with the preset water quality standard range to determine whether there are any situations that exceed the standard range. If any test result is found to exceed the predetermined standard water sample result range, the system will automatically trigger an alarm. The above process is automated to ensure an immediate response without human intervention. Once the alarm is triggered, the alarm system will immediately issue an alarm through a preset communication channel (such as SMS, email, application notification, etc.), providing detailed information on which item or items exceed the standard. The alarm notification contains preliminary response instructions or suggestions, instructing the recipient to take appropriate preliminary measures according to the specific situation, such as suspending the use of water sources, increasing the frequency of testing, etc. All alarm events and corresponding test results are recorded for subsequent analysis and problem tracking.

The following technical effects can be achieved through the above steps: A systematic automatic alarm mechanism can ensure that potential water quality problems are identified and responded to in the first place, reducing the delays caused by manual detection and judgment. Through immediate alarms and preliminary crisis management measures, health risks and environmental damage caused by water quality problems can be effectively reduced. The data recorded by the system provides a basis for subsequent analysis and decision-making, which is conducive to identifying the root causes of water quality problems and the direction of improvement. A reliable water quality monitoring and alarm system can increase public confidence in water quality safety management and enhance public trust in water supply agencies. In summary, by setting up an automated alarm system to respond to abnormal test results, water quality problems can be discovered and handled in a timely manner, thereby protecting public health, reducing environmental risks, and supporting scientific decision-making processes.

In a specific embodiment, a plurality of parameters of a water sample collected by a collection system are obtained, and a multivariate time series data matrix is obtained based on the plurality of parameters of the water sample and the timestamps corresponding to the plurality of parameters of the water sample, including:

A preset collection system is used to collect multiple parameters of the water sample and timestamps corresponding to the multiple parameters of the water sample; wherein the multiple parameters of the water sample include temperature, pH, dissolved oxygen, and turbidity of the water sample; the timestamps corresponding to the multiple parameters of the water sample include temperature timestamp, pH timestamp, dissolved oxygen timestamp, and turbidity timestamp;

Mapping and aligning the temperature, pH, dissolved oxygen, and turbidity of the water sample with the temperature timestamp, pH timestamp, dissolved oxygen timestamp, and turbidity timestamp of the water sample to obtain an initial multivariate time series data matrix;

Inputting the initial multivariate time series data matrix into a preset time window model for partitioning to obtain a first target multivariate time series data matrix;

Perform matrix judgment on the first target multivariate time series data matrix to determine whether there are missing elements in the first target multivariate time series data matrix; if there are missing elements in the first target multivariate time series data matrix, perform median calculation on the multiple parameters of the water sample and the timestamps corresponding to the multiple parameters of the water sample to obtain the elements to be supplemented, and fill the elements to be supplemented into the positions of the missing elements to obtain the second target multivariate time series data matrix; wherein, the second target multivariate time series data matrix is used as the multivariate time series data matrix.

Specifically, a preset collection system is used to collect multiple parameters (temperature, pH, dissolved oxygen, turbidity) of water samples and corresponding timestamps. The collected water sample parameters are mapped and aligned with the corresponding timestamps to form an initial multivariate time series data matrix. The above steps ensure the time correlation of each parameter. The initial data matrix is input into the preset time window model for partitioning to form a first target multivariate time series data matrix. The time window model helps to capture the dynamic changes of time series data. Check whether the first target data matrix has missing elements. If so, calculate the median of each parameter, supplement the missing elements, and form a second target multivariate time series data matrix. The above steps can improve data integrity.

The following technical effects can be obtained through the above steps: By supplementing the missing data, the integrity of the data matrix is ensured, which is very important for subsequent analysis and model training. The use of a time window model can better capture the characteristics of time series data, such as periodicity and trends, which is crucial for understanding water quality change patterns. Mapping alignment and the processing of missing data improve data quality and provide a more reliable basis for subsequent analysis. A high-quality multivariate time series data matrix is the basis for implementing complex analysis and building accurate prediction models, which is very important for water quality monitoring and management. In summary, the data processing steps in this embodiment are intended to ensure the integrity, accuracy and availability of the collected water quality data, laying a solid foundation for further analysis and model application.

In a specific embodiment, the multivariate time series data matrix is analyzed by a preset CRNN model to obtain various analysis results of the water sample, including:

Performing data preprocessing on the multivariate time series data matrix through an input layer in the CRNN model to obtain a preprocessing matrix;

Performing data compression on the preprocessing matrix through a double-layer one-dimensional convolutional recurrent neural network to obtain a compressed matrix;

Performing data feature extraction on the compression matrix to obtain a compression extraction matrix;

Performing nonlinear changes on the compression extraction matrix through an activation function to obtain a target matrix of water quality;

The target matrix of the water quality is subjected to matrix analysis by multiple GRU modules to obtain various analysis results of the water samples.

Specifically, the multivariate time series data matrix is preprocessed through the input layer of the CRNN model to obtain a preprocessed matrix suitable for further analysis. The above steps usually include standardization, normalization, etc. to ensure that the data is within the appropriate range. The preprocessed matrix is processed using a double-layer one-dimensional convolutional layer to reduce the dimension of the data and extract key information to obtain a compressed matrix. The above steps help reduce the amount of calculation and improve processing efficiency. Feature extraction is performed on the compressed matrix to obtain a compressed extraction matrix containing the characteristics of key water quality indicators. The key to the above steps is to identify and extract the most important information for water quality analysis. The compressed extraction matrix is nonlinearly transformed through activation functions (such as ReLU or Sigmoid) to obtain a water quality target matrix that is more suitable for complex analysis. The above steps increase the model's ability to handle complex data. Multiple GRU modules are used to analyze the target matrix of water quality, and finally the various analysis results of the water sample are obtained. The GRU module can capture long-term dependencies in time series data and is very effective in understanding and predicting water quality changes.

The following technical effects can be obtained through the above steps: Through data compression and feature extraction steps, the CRNN model can quickly identify the most important information from a large amount of time series data, significantly improving the efficiency of data processing. After adopting nonlinear transformation, the CRNN model can capture and express more complex data patterns, improving the accuracy and reliability of analysis. The use of the GRU module makes the model particularly good at processing time series data, and can effectively capture the dynamic changes of water quality over time, thereby providing more accurate water quality analysis results. The analysis results obtained by the CRNN model can help relevant personnel understand the current status and possible trends of water quality, and provide a scientific basis for water quality management and decision-making. In summary, the CRNN model has shown high efficiency and accuracy in processing and analyzing multivariate time series data, and is particularly suitable for complex environmental monitoring and analysis tasks that need to consider time dependence, such as water quality analysis.

In a specific embodiment, each analysis result of the water sample is input into a preset detection model for detection, and correspondingly each detection result of the water sample is obtained, including:

Inputting the analysis results of the water samples into a preset detection model for detection to obtain the corresponding detection values of the water samples;

Calculating the detection value of each water sample respectively to obtain the classification weight corresponding to each water sample;

The classification weights corresponding to each of the water samples are weighted and summed through the attention mechanism to obtain a preliminary target detection result;

Determining whether there are repeated target detection results in the preliminary target detection results;

If the preliminary target detection result contains repeated target detection results, the NMS algorithm is used to eliminate the repeated target detection results to obtain a standard target detection result.

Specifically, multiple analysis results of each water sample (such as temperature, pH value, dissolved oxygen content, etc.) are input into a pre-trained machine learning/deep learning model to obtain the detection value of each water sample. The detection value obtained for each water sample is further used to calculate its corresponding classification weight, which reflects the possibility or importance of the water sample belonging to a certain category. The attention mechanism is used to perform weighted summation according to the classification weights of each water sample to obtain the preliminary target detection result. The attention mechanism helps the model focus on more informative features and improve the accuracy of detection. Check whether there are repeated detection results in the preliminary target detection results. In the context of water quality detection, problems such as repeated identification of similar pollutants are involved. If there are repeated target detection results, the non-maximum suppression (NMS) algorithm is used to suppress or eliminate those repeated results. The NMS algorithm ensures the accuracy and uniqueness of the final detection results by retaining the detection results with higher weights and removing other repeated results with low weights (or low confidence).

The following technical effects can be obtained through the above steps: By combining the attention mechanism and the NMS algorithm, false detections and duplications can be effectively reduced, and key indicators and pollutants in water samples can be more accurately identified and analyzed. The automated process reduces the need for manual testing, and at the same time improves the speed and efficiency of data processing by quickly eliminating duplicate results. The use of the attention mechanism enables the model to be better adapted to various water quality testing scenarios and complexities, and improves the model's ability to identify and classify unknown water samples. The standardized target detection results finally obtained can provide more accurate data support for water quality management and help make more scientific decisions. In summary, this embodiment demonstrates an advanced data processing process that combines modern machine learning technology and is applied in the field of water quality testing, demonstrating its obvious technical effects in improving accuracy, optimizing efficiency, and enhancing model generalization capabilities.

In a specific embodiment, each analysis result of the water sample is input into a preset detection model for detection, and corresponding detection values of each water sample are obtained, including:

Performing feature extraction on each analysis result of the water sample to obtain each feature data of the water sample; wherein each feature data of the water sample includes a temperature value, a pH value, a dissolved oxygen value, and a turbidity value of the water sample;

Input each characteristic data of the water sample into a preset detection model for detection, and determine whether there is any abnormal data in each characteristic data of the water sample; wherein the abnormal data includes abnormalities in each characteristic data of the water sample and missing of each characteristic data;

If there are abnormal data in each characteristic data of the water sample, the abnormal data is replaced to obtain standard test data;

The standard test data are sorted by a preset test value sequence model to obtain corresponding test value sequences of each water sample; wherein the test value sequences of each water sample are the test results of each water sample.

Specifically, key characteristic data are extracted from the analysis results of each water sample, including temperature, pH value, dissolved oxygen content, turbidity, etc. Temperature, pH value, dissolved oxygen content, turbidity, etc. are common indicators for measuring water quality and can reflect the basic physical and chemical state of water. The extracted characteristic data are input into the preset detection model. The detection model can determine whether the water sample meets the normal water quality range based on these characteristic data and identify abnormal characteristic data. Abnormal characteristic data is abnormal indicator value or missing data. For the detected abnormal data, a certain method (such as replacing it with the mean, median or estimated value based on the prediction model) is used to correct it to ensure that each indicator has a reasonable data value so that subsequent processing can proceed smoothly. The processed standard test data is sorted by the preset test value sequence model to generate a test value sequence that can reflect the water quality status of each water sample.

The following technical effects can be obtained through the above steps: through advanced data processing methods, abnormal data can be effectively identified and corrected, ensuring the accuracy of water quality detection data, thereby improving the reliability of water quality detection. Automated feature extraction and abnormal data processing not only improve the speed of processing data, but also reduce the probability of human error, making large-scale water quality monitoring possible. By taking multiple water quality indicators into consideration, the water quality status can be more comprehensively evaluated, which helps to accurately identify water quality problems and provide more scientific guidance for water treatment. The generated water sample test value sequence can intuitively reflect the overall status of water quality, provide a scientific basis for water quality management and control, and support data-based decision-making. In summary, this embodiment can effectively improve the efficiency and accuracy of water quality detection through scientific data processing procedures and advanced technical methods, and has important technical value in ensuring water resource security and guiding water treatment.

The above describes the multi-parameter water quality online monitoring method in the embodiment of the present invention. The following describes the multi-parameter water quality online monitoring device in the embodiment of the present invention. Please refer to Figure 2. An embodiment of the multi-parameter water quality online monitoring device in the embodiment of the present invention includes:

An acquisition module 21 is used to acquire multiple parameters of the water sample collected from the acquisition system and timestamps corresponding to the multiple parameters of the water sample, and obtain a multivariate time series data matrix based on the multiple parameters of the water sample and the timestamps corresponding to the multiple parameters of the water sample;

The analysis module 22 is used to analyze the multivariate time series data matrix through a preset CRNN model to obtain various analysis results of the water sample; wherein the CRNN model includes a CNN layer and a GRU layer; the CNN layer includes a double-layer one-dimensional convolutional recurrent neural network and an activation function; the activation function is respectively connected to multiple GRU modules in a single-layer GRU layer;

The detection module 23 is used to input each analysis result of the water sample into a preset detection model for detection, obtain each detection result of the water sample accordingly, and determine whether each detection result of the water sample is within a preset range of the standard water sample result;

The judgment module 24 is used to trigger an alarm system if any of the test results is not within a preset range of the standard water sample results.

In this embodiment, for the specific implementation of each unit in the above device embodiment, please refer to the description in the above method embodiment, which will not be repeated here.

Referring to FIG3 , a computer device is also provided in an embodiment of the present invention, and the internal structure of the computer device may be as shown in FIG3 . The computer device includes a processor, a memory, a display screen, an input device, a network interface, and a database connected via a system bus. Among them, the processor designed by the computer is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store the corresponding data in this embodiment. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, the above method is implemented.

Those skilled in the art can understand that the structure shown in FIG. 3 is merely a block diagram of a portion of the structure related to the solution of the present invention, and does not constitute a limitation on the computer device to which the solution of the present invention is applied.

An embodiment of the present invention further provides a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, the above method is implemented. It can be understood that the computer-readable storage medium in this embodiment can be a volatile readable storage medium or a non-volatile readable storage medium.

Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media provided by the present invention and used in the embodiments may include non-volatile and/or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double-speed data rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM.

It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, device, article or method including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, device, article or method. In the absence of further restrictions, an element defined by the sentence "includes a ..." does not exclude the presence of other identical elements in the process, device, article or method including the element.

The above description is only a preferred embodiment of the present invention, and does not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made by using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims (8)

1. A multi-parameter water quality on-line monitoring method is characterized in that: the method comprises the following steps:

acquiring multiple parameters of a water sample acquired by an acquisition system and time stamps corresponding to the multiple parameters of the water sample, and acquiring a multivariate time sequence data matrix based on the multiple parameters of the water sample and the time stamps corresponding to the multiple parameters of the water sample;

analyzing the multivariate time sequence data matrix through a preset CRNN model to obtain each analysis result of the water sample; wherein the CRNN model comprises a CNN layer and a GRU layer; the CNN layer comprises a double-layer one-dimensional convolution cyclic neural network and an activation function; the activation function is respectively connected with a plurality of GRU modules in the single GRU layer;

inputting each analysis result of the water sample into a preset detection model for detection, correspondingly obtaining each detection result of the water sample, and judging whether each detection result of the water sample is in a preset range of a standard water sample result;

and if any detection result is not in the preset range of the standard water sample result, triggering an alarm system.

2. The multi-parameter water quality on-line monitoring method according to claim 1, wherein: acquiring multiple parameters of a water sample acquired by an acquisition system, and acquiring a multivariate time sequence data matrix based on the multiple parameters of the water sample and timestamps corresponding to the multiple parameters of the water sample, wherein the multivariate time sequence data matrix comprises the following components:

collecting various parameters of the water sample and corresponding time stamps of the various parameters of the water sample by adopting a preset collecting system; wherein, the various parameters of the water sample comprise the temperature, pH, dissolved oxygen and turbidity of the water sample; the time stamps corresponding to various parameters of the water sample comprise a temperature time stamp, a pH time stamp, a dissolved oxygen time stamp and a turbidity time stamp;

mapping and aligning the temperature, pH, dissolved oxygen and turbidity of the water sample with a temperature time stamp, a pH value time stamp, a dissolved oxygen time stamp and a turbidity time stamp of the water sample to obtain an initial multivariable time sequence data matrix;

inputting the initial multi-variable time sequence data matrix into a preset time window model for division to obtain a first target multi-variable time sequence data matrix;

matrix judgment is carried out on the first target multi-variable time sequence data matrix, and whether elements lack in the first target multi-variable time sequence data matrix or not is judged; if the first target multivariable time sequence data matrix lacks elements, performing median calculation on various parameters of the water sample and timestamps corresponding to the various parameters of the water sample to obtain elements to be supplemented, and supplementing the elements to be supplemented to the positions lacking the elements to obtain a second target multivariable time sequence data matrix; wherein the second target multivariate time series data matrix is used as a multivariate time series data matrix.

3. The multi-parameter water quality on-line monitoring method according to claim 1, wherein: analyzing the multivariate time sequence data matrix through a preset CRNN model to obtain each analysis result of the water sample, wherein the analysis result comprises the following steps:

performing data preprocessing on the multivariate time series data matrix through an input layer in a CRNN model to obtain a preprocessing matrix;

performing data compression on the preprocessing matrix through a double-layer one-dimensional convolution cyclic neural network to obtain a compression matrix;

extracting data features of the compression matrix to obtain a compression extraction matrix;

nonlinear change is carried out on the compression extraction matrix through an activation function, and a target matrix of water quality is obtained;

and performing matrix analysis on the target matrix of the water quality through a plurality of GRU modules to obtain each analysis result of the water sample.

4. The multi-parameter water quality on-line monitoring method according to claim 1, wherein: inputting each analysis result of the water sample into a preset detection model for detection, and correspondingly obtaining each detection result of the water sample, wherein the detection method comprises the following steps of:

inputting each analysis result of the water sample into a preset detection model for detection to obtain a detection value of each corresponding water sample;

respectively calculating the detection values of the water samples to obtain classification weights corresponding to the water samples;

carrying out weighted summation on the classification weights corresponding to the water samples through an attention mechanism to obtain a preliminary target detection result;

judging whether the preliminary target detection result has a repeated target detection result or not;

and if the preliminary target detection result contains a repeated target detection result, rejecting the repeated target detection result by adopting an NMS algorithm to obtain a standard target detection result.

5. The multi-parameter on-line monitoring method of water quality according to claim 4, wherein: inputting each analysis result of the water sample into a preset detection model for detection, correspondingly obtaining detection values of each corresponding water sample, wherein the detection values comprise:

respectively extracting the characteristics of each analysis result of the water sample to obtain each characteristic data of the water sample; wherein, each characteristic data of the water sample comprises a temperature value, a pH value, a dissolved oxygen value and a turbidity value of the water sample;

inputting each characteristic data of the water sample into a preset detection model for detection, and judging whether each characteristic data of the water sample has abnormal data or not; wherein the abnormal data comprise abnormal characteristic data of the water sample and missing characteristic data;

if abnormal data exist in each characteristic data of the water sample, replacing the abnormal data to obtain standard detection data;

sequencing the standard detection data through a preset detection value sequence model to obtain a detection value sequence of each corresponding water sample; wherein the detection value sequence of each water sample is each detection result of the water sample.

6. The utility model provides a multiparameter quality of water on-line monitoring device which characterized in that includes:

the acquisition module is used for acquiring various parameters of the water sample acquired by the acquisition system and time stamps corresponding to the various parameters of the water sample, and acquiring a multivariate time sequence data matrix based on the various parameters of the water sample and the time stamps corresponding to the various parameters of the water sample;

the analysis module is used for analyzing the multivariate time sequence data matrix through a preset CRNN model to obtain each analysis result of the water sample; wherein the CRNN model comprises a CNN layer and a GRU layer; the CNN layer comprises a double-layer one-dimensional convolution cyclic neural network and an activation function; the activation function is respectively connected with a plurality of GRU modules in the single GRU layer;

the detection module is used for respectively inputting each analysis result of the water sample into a preset detection model for detection, correspondingly obtaining each detection result of the water sample, and judging whether each detection result of the water sample is in a preset range of a standard water sample result;

and the judging module is used for triggering the alarm system if any detection result is not in the preset range of the standard water sample result.

7. A computer device comprising a memory and a processor, the memory having stored therein a computer program, characterized in that the processor, when executing the computer program, implements the steps of the method of any of claims 1 to 5.

8. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1 to 5.