Real-time monitoring method for boiler flue gas waste heat recovery system

Abstract

The present disclosure relates to a real-time monitoring method for a boiler flue gas waste heat recovery system. The method comprises: real-time monitoring and collecting flue gas inlet temperature and flue gas outlet temperature data through a first temperature sensor and a second temperature sensor deployed at the inlet and outlet of a flue gas cooler, and then extracting features through a temperature time series pattern feature extractor and a heat recovery effect time series feature extractor based on a deep neural network, and then calculating the heat recovery effect anchor point and cosine similarity to obtain the recovery effect time series fluctuation feature, so as to determine whether there is an abnormality in the operation state of the waste heat recovery system. In this way, real-time monitoring and abnormality judgment of the waste heat recovery system can be achieved, so as to timely discover problems and abnormalities in the system, thereby facilitating timely measures to repair and avoid system performance degradation or even safety accidents.

Description

Technical Field

[0001] The present disclosure relates to the technical field of waste heat recovery, and in particular to a real-time monitoring method for a boiler flue gas waste heat recovery system. Background Art

[0002] The boiler flue gas waste heat recovery system is a common energy recovery device in industrial production. It can improve energy utilization efficiency by recovering the heat energy in the flue gas. Specifically, the heat recovery system recovers the waste heat in the flue gas in the flue gas cooler and uses it to heat water or steam, thereby realizing energy reuse and reducing energy consumption. However, various problems may occur in the waste heat recovery system during operation, such as pipe blockage, heat exchanger leakage, etc. These problems will affect the performance and safe operation of the system. Therefore, monitoring the operating status of the boiler flue gas waste heat recovery system is crucial to ensure the efficient and safe operation of the system.

[0003] However, traditional boiler flue gas waste heat recovery system monitoring solutions usually adopt manual inspections or regular inspections. Manual inspections or regular inspections cannot achieve real-time monitoring of the waste heat recovery system status, resulting in system problems that may not be discovered during the inspection cycle, causing energy waste or safety hazards. In addition, the manual inspection method is easily affected by the subjective factors of the operator, and there is a possibility of missed inspections or misjudgments. At the same time, the manual inspection method also requires a lot of manpower and material resources, and cannot comprehensively and continuously monitor the system operation status.

[0004] Therefore, an optimized real-time monitoring solution for the boiler flue gas waste heat recovery system is desired. Summary of the invention

[0005] This summary is provided to introduce concepts in a brief form that will be described in detail in the detailed description below. This summary is not intended to identify key features or essential features of the claimed technical solution, nor is it intended to limit the scope of the claimed technical solution.

[0006] In a first aspect, the present disclosure provides a real-time monitoring method for a boiler flue gas waste heat recovery system, the method comprising:

[0007] Acquire a time series of flue gas inlet temperature and a time series of flue gas outlet temperature collected by a first temperature sensor and a second temperature sensor deployed at an inlet and an outlet of a flue gas cooler;

[0008] Performing sequence segmentation and data regularization on the time series of the smoke inlet temperature and the time series of the smoke outlet temperature at the same time scale to obtain a sequence of local time series input vectors of the smoke inlet temperature and a sequence of local time series input vectors of the smoke outlet temperature;

[0009] The temperature time series pattern feature extractor based on a deep neural network is used to extract features from the sequence of local time series input vectors of the smoke inlet temperature and the sequence of local time series input vectors of the smoke outlet temperature to obtain a sequence of local time series associated feature vectors of the smoke inlet temperature and a sequence of local time series associated feature vectors of the smoke outlet temperature;

[0010] Inputting each group of corresponding local time series correlation feature vectors of flue gas inlet temperature and local time series correlation feature vectors of flue gas outlet temperature in the sequence of local time series correlation feature vectors of flue gas inlet temperature and the sequence of local time series correlation feature vectors of flue gas outlet temperature into a heat recovery effect time series feature extractor to obtain a sequence of heat recovery local time series effect feature vectors;

[0011] Calculating a heat recovery effect anchor point of the sequence of heat recovery local time series effect feature vectors, and calculating the cosine similarity of each heat recovery local time series effect feature vector in the sequence of heat recovery local time series effect feature vectors relative to the heat recovery effect anchor point to obtain a recovery effect time series fluctuation feature;

[0012] Based on the time series fluctuation characteristics of the recovery effect, it is determined whether there is any abnormality in the operation state of the waste heat recovery system.

[0013] Optionally, the temperature time series pattern feature extractor based on a deep neural network is a temperature time series pattern feature extractor based on a one-dimensional extended convolutional layer.

[0014] Optionally, a temperature time series pattern feature extractor based on a deep neural network is used to perform feature extraction on the sequence of the local time series input vectors of the smoke inlet temperature and the sequence of the local time series input vectors of the smoke outlet temperature to obtain a sequence of local time series associated feature vectors of the smoke inlet temperature and a sequence of local time series associated feature vectors of the smoke outlet temperature, including: processing the sequence of the local time series input vectors of the smoke inlet temperature by the temperature time series pattern feature extractor based on the one-dimensional extended convolution layer using the following time series pattern feature extraction formula to obtain a sequence of local time series associated feature vectors of the smoke inlet temperature; wherein the time series pattern feature extraction formula is:

[0015]

[0016] C i =f(ω·X i:i+j-l +b)

[0017] Among them, X1,X2,...,X n represents each local time series input vector of the flue gas inlet temperature in the sequence of the local time series input vector of the flue gas inlet temperature, X 1:nA cascade vector representing a sequence of local time series input vectors of the flue gas inlet temperature, Indicates the connection operation, X i:i+j-l X in the sequence representing the local time series input vector of the flue gas inlet temperature i ,X i+1 ,...,X i+j-1 , ω and b represent the weight matrix and bias vector respectively, f(·) represents the convolution operation, C i The i-th local time series correlation feature vector of the flue gas inlet temperature in the sequence of local time series correlation feature vectors of the flue gas inlet temperature is represented.

[0018] Optionally, each group of corresponding local time series correlation feature vectors of flue gas inlet temperature and local time series correlation feature vectors of flue gas outlet temperature in the sequence of local time series correlation feature vectors of flue gas inlet temperature and the sequence of local time series correlation feature vectors of flue gas outlet temperature is input into a heat recovery effect time series feature extractor to obtain a sequence of local time series effect feature vectors of heat recovery, including: each group of corresponding local time series correlation feature vectors of flue gas inlet temperature and local time series correlation feature vectors of flue gas outlet temperature in the sequence of local time series correlation feature vectors of flue gas inlet temperature and the sequence of local time series correlation feature vectors of flue gas outlet temperature is input into the heat recovery effect time series feature extractor and processed with the following heat recovery effect feature extraction formula to obtain a sequence of local time series effect feature vectors of heat recovery; wherein, the heat recovery effect feature extraction formula is:

[0019]

[0020] in, and are the local time series correlation feature vectors of the smoke inlet temperature and the local time series correlation feature vectors of the smoke outlet temperature corresponding to each group in the sequence of the local time series correlation feature vectors of the smoke inlet temperature and the sequence of the local time series correlation feature vectors of the smoke outlet temperature, respectively. is the square of the L2 norm of the vector, L2 is a hyperparameter, v c It is each heat recovery local timing effect feature vector in the sequence of the heat recovery local timing effect feature vectors.

[0021] Optionally, a heat recovery effect anchor point of the sequence of heat recovery local timing effect feature vectors is calculated, and the cosine similarity of each heat recovery local timing effect feature vector in the sequence of heat recovery local timing effect feature vectors relative to the heat recovery effect anchor point is calculated to obtain the recovery effect timing fluctuation characteristic, including: calculating the global mean vector of the sequence of heat recovery local timing effect feature vectors as the heat recovery effect anchor point; calculating the cosine similarity of each heat recovery local timing effect feature vector in the sequence of heat recovery local timing effect feature vectors relative to the heat recovery effect anchor point to obtain a heat recovery effect timing fluctuation feature vector composed of multiple cosine similarities as the heat recovery effect timing fluctuation characteristic.

[0022] Optionally, based on the time-series fluctuation characteristics of the recovery effect, determining whether there is any abnormality in the operating state of the waste heat recovery system includes: passing the time-series fluctuation characteristic vector of the heat recovery effect through a classifier-based monitor to obtain a monitoring result, and the monitoring result is used to indicate whether there is any abnormality in the operating state of the waste heat recovery system.

[0023] Optionally, a training step is also included: used to train the temperature time series pattern feature extractor based on the one-dimensional extended convolution layer, the heat recovery effect time series feature extractor and the classifier-based monitor.

[0024] Optionally, the training step includes: obtaining a time series of training flue gas inlet temperature and a time series of training flue gas outlet temperature collected by a first temperature sensor and a second temperature sensor deployed at the inlet and outlet of the flue gas cooler, as well as a real value of whether there is an abnormality in the operating state of the waste heat recovery system; performing sequence segmentation and data regularization on the time series of the training flue gas inlet temperature and the time series of the training flue gas outlet temperature at the same time scale to obtain a sequence of local time series input vectors of the training flue gas inlet temperature and a sequence of local time series input vectors of the training flue gas outlet temperature; performing feature extraction on the sequence of local time series input vectors of the training flue gas inlet temperature and the sequence of local time series input vectors of the training flue gas outlet temperature respectively through the temperature time series pattern feature extractor based on the deep neural network to obtain a sequence of local time series associated feature vectors of the training flue gas inlet temperature and a sequence of local time series associated feature vectors of the training flue gas outlet temperature; and transmitting the sequence of local time series associated feature vectors of the training flue gas inlet temperature and the sequence of local time series associated feature vectors of the training flue gas outlet temperature to the temperature sensor. Each group of corresponding training flue gas inlet temperature local time series association feature vectors and training flue gas outlet temperature local time series association feature vectors in the sequence of quantities are input into the heat recovery effect time series feature extractor to obtain a sequence of training heat recovery local time series effect feature vectors; the global mean vector of the sequence of training heat recovery local time series effect feature vectors is calculated as the training heat recovery effect anchor point; the training cosine similarity of each training heat recovery local time series effect feature vector in the sequence of training heat recovery local time series effect feature vectors relative to the training heat recovery effect anchor point is calculated to obtain a training heat recovery effect time series fluctuation feature vector composed of multiple training cosine similarities; the training heat recovery effect time series fluctuation feature vector is passed through the classifier-based monitor to obtain a classification loss function value; based on the classification loss function value, the temperature time series pattern feature extractor based on the one-dimensional expanded convolution layer, the heat recovery effect time series feature extractor and the classifier-based monitor are trained, and in each iteration, the training heat recovery effect time series fluctuation feature vector is corrected.

[0025] Optionally, passing the training heat recovery effect time series fluctuation feature vector through the classifier-based monitor to obtain a classification loss function value, including: using the classifier-based monitor to process the training heat recovery effect time series fluctuation feature vector with the following training classification formula to generate a training classification result, wherein the training classification formula is: softax{(W n ,B n ):…:(W1,B1)|X}, where X represents the time series fluctuation feature vector of the training heat recovery effect, W1 to W n is the weight matrix, B1 to B nrepresents a bias matrix; and calculating a cross entropy value between the training classification result and the true value as the classification loss function value.

[0026] By adopting the above technical solution, the first temperature sensor and the second temperature sensor deployed at the inlet and outlet of the flue gas cooler are used to monitor and collect the flue gas inlet temperature and flue gas outlet temperature data in real time, and then feature extraction is performed through the temperature time series pattern feature extractor and the heat recovery effect time series feature extractor based on the deep neural network, and then the heat recovery effect anchor point and cosine similarity are calculated to obtain the recovery effect time series fluctuation characteristics to determine whether there is an abnormality in the operation status of the waste heat recovery system. In this way, real-time monitoring and abnormality judgment of the waste heat recovery system can be realized, so as to timely discover problems and abnormalities in the system, which is conducive to taking timely measures to repair and avoid system performance degradation or even safety accidents.

[0027] Other features and advantages of the present disclosure will be described in detail in the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] The above and other features, advantages and aspects of the embodiments of the present disclosure will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. Throughout the drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and the originals and elements are not necessarily drawn to scale. In the drawings:

[0029] Figure 1 The present invention is a flow chart showing a real-time monitoring method of a boiler flue gas waste heat recovery system according to an exemplary embodiment.

[0030] Figure 2 It is a block diagram of a real-time monitoring system of a boiler flue gas waste heat recovery system according to an exemplary embodiment.

[0031] Figure 3 It is a block diagram of an electronic device according to an exemplary embodiment.

[0032] Figure 4 It is an application scenario diagram of a real-time monitoring method of a boiler flue gas waste heat recovery system according to an exemplary embodiment. DETAILED DESCRIPTION

[0033] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as being limited to the embodiments described herein, which are instead provided for a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not intended to limit the scope of protection of the present disclosure.

[0034] It should be understood that the various steps described in the method embodiments of the present disclosure may be performed in different orders and / or in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present disclosure is not limited in this respect.

[0035] The term "including" and its variations used herein are open inclusions, i.e., "including but not limited to". The term "based on" means "based at least in part on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". The relevant definitions of other terms will be given in the following description.

[0036] It should be noted that the concepts such as "first" and "second" mentioned in the present disclosure are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0037] It should be noted that the modifications of "one" and "plurality" mentioned in the present disclosure are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise clearly indicated in the context, it should be understood as "one or more".

[0038] The names of the messages or information exchanged between multiple devices in the embodiments of the present disclosure are only used for illustrative purposes and are not used to limit the scope of these messages or information.

[0039] In response to the above technical problems, in the technical solution of the present application, a real-time monitoring method for a boiler flue gas waste heat recovery system is proposed, which can determine whether there is an abnormality in the operating state of the waste heat recovery system by real-time monitoring of the flue gas temperature. Specifically, the first temperature sensor and the second temperature sensor deployed at the inlet and outlet of the flue gas cooler monitor and collect the flue gas inlet temperature and the flue gas outlet temperature data in real time, and introduce an artificial intelligence-based data processing and analysis algorithm at the back end to perform a time series collaborative correlation analysis of the flue gas inlet temperature and the flue gas outlet temperature, so as to detect and evaluate the heat recovery effect, thereby detecting abnormalities in the operating state of the waste heat recovery system. In this way, real-time monitoring and abnormal judgment of the waste heat recovery system can be achieved, so as to timely discover problems and abnormalities in the system, such as pipe blockage, heat exchanger leakage, etc., which is conducive to taking timely measures to repair and avoid system performance degradation or even safety accidents.

[0040] In order to solve the above problems, the present disclosure provides a real-time monitoring method for a boiler flue gas waste heat recovery system, which uses a first temperature sensor and a second temperature sensor deployed at the inlet and outlet of a flue gas cooler to monitor and collect flue gas inlet temperature and flue gas outlet temperature data in real time, and then uses a temperature time series pattern feature extractor and a heat recovery effect time series feature extractor based on a deep neural network to extract features, and then calculates the heat recovery effect anchor point and cosine similarity to obtain the recovery effect time series fluctuation feature to determine whether there is an abnormality in the operation state of the waste heat recovery system. In this way, real-time monitoring and abnormality judgment of the waste heat recovery system can be achieved, so as to timely discover problems and abnormalities in the system, which is conducive to taking timely measures to repair and avoid system performance degradation or even safety accidents.

[0041] The specific embodiments of the present disclosure are described in detail below with reference to the accompanying drawings.

[0042] Figure 1 is a flow chart showing a real-time monitoring method of a boiler flue gas waste heat recovery system according to an exemplary embodiment. Figure 1 As shown, the method includes:

[0043] Step 101, obtaining a time series of flue gas inlet temperature and a time series of flue gas outlet temperature collected by a first temperature sensor and a second temperature sensor deployed at the inlet and outlet of a flue gas cooler;

[0044] Step 102: performing sequence segmentation and data regularization on the time series of the smoke inlet temperature and the time series of the smoke outlet temperature at the same time scale to obtain a sequence of local time series input vectors of the smoke inlet temperature and a sequence of local time series input vectors of the smoke outlet temperature;

[0045] Step 103: extracting features from the sequence of local time series input vectors of the flue gas inlet temperature and the sequence of local time series input vectors of the flue gas outlet temperature respectively by using a temperature time series pattern feature extractor based on a deep neural network to obtain a sequence of local time series associated feature vectors of the flue gas inlet temperature and a sequence of local time series associated feature vectors of the flue gas outlet temperature;

[0046] Step 104: input each group of corresponding local time series correlation feature vectors of flue gas inlet temperature and local time series correlation feature vectors of flue gas outlet temperature in the sequence of local time series correlation feature vectors of flue gas inlet temperature and the sequence of local time series correlation feature vectors of flue gas outlet temperature into a heat recovery effect time series feature extractor to obtain a sequence of heat recovery local time series effect feature vectors;

[0047] Step 105: Calculate the heat recovery effect anchor point of the sequence of the heat recovery local time series effect feature vectors, and calculate the cosine similarity of each heat recovery local time series effect feature vector in the sequence of the heat recovery local time series effect feature vectors relative to the heat recovery effect anchor point to obtain the recovery effect time series fluctuation characteristics;

[0048] Step 106: Determine whether there is any abnormality in the operation state of the waste heat recovery system based on the time series fluctuation characteristics of the recovery effect.

[0049] Among them, the temperature time series pattern feature extractor based on deep neural network is a temperature time series pattern feature extractor based on a one-dimensional extended convolutional layer.

[0050] Specifically, in the technical solution of the present application, first, the time series of the smoke inlet temperature and the time series of the smoke outlet temperature collected by the first temperature sensor and the second temperature sensor deployed at the inlet and outlet of the smoke cooler are obtained. Next, considering that the smoke inlet temperature and the smoke outlet temperature have dynamic time series changes in the time dimension, that is, the smoke inlet temperature and the smoke outlet temperature will continue to change with the passage of time. Therefore, in order to better capture the time series change characteristics of the smoke temperature, including the time series pattern and change trend information of the temperature, so as to more accurately analyze the operating state of the system, in the technical solution of the present application, it is necessary to perform sequence segmentation and data regularization on the time series of the smoke inlet temperature and the time series of the smoke outlet temperature at the same time scale to obtain a sequence of local time series input vectors of the smoke inlet temperature and a sequence of local time series input vectors of the smoke outlet temperature. Slicing the time series data of the smoke inlet temperature and the smoke outlet temperature according to the same time scale and regularizing them to the temperature local time series input vectors of the same time scale can ensure that the data is aligned in time, which is convenient for subsequent feature extraction and analysis.

[0051] Furthermore, a one-dimensional convolution layer is used to perform feature analysis and extraction of the sequence of local time series input vectors of the flue gas inlet temperature and the sequence of local time series input vectors of the flue gas outlet temperature, so as to capture the time series characteristics of the flue gas inlet temperature and the flue gas outlet temperature in each local time period, including the local time series pattern and characteristics of the temperature. In particular, considering that due to the limitations of the one-dimensional convolution layer, it is not possible to pay attention to the local time series correlation characteristics of the flue gas temperature at different step lengths between each local time period, however, when the actual system operation status detection is performed, the flue gas inlet temperature and the flue gas outlet temperature may present different periodic time series patterns and correlation relationships at different time spans. Therefore, in order to better perform time series analysis on the flue gas inlet temperature and the flue gas outlet temperature, it is necessary to establish the correlation relationship between these local time series characteristics of temperature at different step lengths, so as to better understand and utilize the time series information of the temperature data, so as to better detect the heat recovery effect and judge the system operation status. Based on this, in the technical solution of the present application, the sequence of the local time series input vectors of the flue gas inlet temperature and the sequence of the local time series input vectors of the flue gas outlet temperature are further respectively passed through a temperature time series pattern feature extractor based on a one-dimensional extended convolution layer to obtain a sequence of local time series correlation feature vectors of the flue gas inlet temperature and a sequence of local time series correlation feature vectors of the flue gas outlet temperature. It should be understood that compared with the one-dimensional convolution layer, the one-dimensional extended convolution layer introduces a hyperparameter called the expansion rate to control the number of null values ​​0 in the convolution kernel. The expansion rate defines the spacing between each time point when the convolution kernel processes the temperature time series data. After the extended convolution operation, the temperature time series correlation characteristics at different step lengths and at a long distance can be captured. In this way, the time series variation law of the flue gas temperature data can be better understood, which helps to identify important patterns and trends in the temperature data, thereby better understanding the operating status of the system, which helps to detect abnormal situations in a timely manner and take corresponding measures.

[0052] In one embodiment of the present disclosure, a temperature time series pattern feature extractor based on a deep neural network is used to perform feature extraction on a sequence of local time series input vectors of the smoke inlet temperature and a sequence of local time series input vectors of the smoke outlet temperature to obtain a sequence of local time series associated feature vectors of the smoke inlet temperature and a sequence of local time series associated feature vectors of the smoke outlet temperature, including: processing the sequence of local time series input vectors of the smoke inlet temperature by the temperature time series pattern feature extractor based on a one-dimensional extended convolution layer using the following time series pattern feature extraction formula to obtain a sequence of local time series associated feature vectors of the smoke inlet temperature; wherein the time series pattern feature extraction formula is:

[0053]

[0054] C i =f(ω·Xi:i+j-l +b)

[0055] Among them, X1,X2,...,X n represents each local time series input vector of the flue gas inlet temperature in the sequence of the local time series input vector of the flue gas inlet temperature, X 1:n A cascade vector representing a sequence of local time series input vectors of the flue gas inlet temperature, Indicates the connection operation, X i:i+j-l X in the sequence representing the local time series input vector of the flue gas inlet temperature i ,X i+1 ,...,X i+j-1 , ω and b represent the weight matrix and bias vector respectively, f(·) represents the convolution operation, C i The i-th local time series correlation feature vector of the flue gas inlet temperature in the sequence of local time series correlation feature vectors of the flue gas inlet temperature is represented.

[0056] It should be understood that, since there is a time-series correlation relationship between the flue gas inlet temperature and the outlet temperature, in order to better understand the heat transfer and energy transfer of the flue gas inside the system, and thus accurately analyze the timing effect of heat recovery, so as to improve the accuracy of the system's operating status detection, in the technical solution of the present application, each group of corresponding flue gas inlet temperature local time-series correlation feature vectors and flue gas outlet temperature local time-series correlation feature vectors in the sequence of the flue gas inlet temperature local time-series correlation feature vectors and the sequence of the flue gas outlet temperature local time-series correlation feature vectors are further input into the heat recovery effect timing feature extractor to obtain a sequence of heat recovery local time-series effect feature vectors. By inputting each group of corresponding flue gas inlet and outlet temperature timing correlation feature vectors into the heat recovery effect timing feature extractor, the temperature difference interaction information between the flue gas passing through the inlet and outlet in the waste heat recovery system can be analyzed, thereby obtaining the local timing effect characteristics of the heat recovery system. This helps to evaluate the performance of the heat recovery system, understand the impact of temperature changes on energy recovery during the heat recovery process, and thus infer the operating effect of the heat recovery system.

[0057] In one embodiment of the present disclosure, each group of corresponding local time series correlation feature vectors of flue gas inlet temperature and local time series correlation feature vectors of flue gas outlet temperature in the sequence of local time series correlation feature vectors of flue gas inlet temperature and the sequence of local time series correlation feature vectors of flue gas outlet temperature is input into a heat recovery effect time series feature extractor to obtain a sequence of local time series effect feature vectors of heat recovery, including: each group of corresponding local time series correlation feature vectors of flue gas inlet temperature and local time series correlation feature vectors of flue gas outlet temperature in the sequence of local time series correlation feature vectors of flue gas inlet temperature and the sequence of local time series correlation feature vectors of flue gas outlet temperature is input into the heat recovery effect time series feature extractor and processed with the following heat recovery effect feature extraction formula to obtain a sequence of local time series effect feature vectors of heat recovery; wherein, the heat recovery effect feature extraction formula is:

[0058]

[0059] in, and are the local time series correlation feature vectors of the smoke inlet temperature and the local time series correlation feature vectors of the smoke outlet temperature corresponding to each group in the sequence of the local time series correlation feature vectors of the smoke inlet temperature and the sequence of the local time series correlation feature vectors of the smoke outlet temperature, respectively. is the square of the L2 norm of the vector, L2 is a hyperparameter, v c It is each heat recovery local timing effect feature vector in the sequence of the heat recovery local timing effect feature vectors.

[0060] Then, considering that in the actual process of detecting abnormal operation status of the waste heat recovery system, the heat recovery timing effect of the waste heat recovery system under abnormal operation status will obviously deviate from the normal situation. Based on this, in the technical solution of the present application, the global mean vector of the sequence of the heat recovery local timing effect feature vector is further calculated as the heat recovery effect anchor point, and the cosine similarity of each heat recovery local timing effect feature vector in the sequence of the heat recovery local timing effect feature vector relative to the heat recovery effect anchor point is calculated to obtain a heat recovery effect timing fluctuation feature vector composed of multiple cosine similarities. It should be understood that by taking the global mean vector as the heat recovery effect anchor point, the overall average heat recovery effect can be represented. At the same time, by calculating the cosine similarity between each heat recovery local timing effect feature vector and the heat recovery effect anchor point, the similarity between each heat recovery local timing effect feature vector and the overall average heat recovery effect can be quantified, thereby capturing the timing fluctuation characteristics of the heat recovery effect in the time dimension. Specifically, the heat recovery local time series effect feature vector with high similarity indicates that the effect is relatively stable, while the heat recovery local time series effect feature vector with low similarity indicates that the effect fluctuates greatly, which helps to analyze the stability and fluctuation of the heat recovery system. When the cosine similarity between a heat recovery local time series effect feature vector and the anchor point significantly deviates from the normal range, it may indicate that the system is abnormal, which is conducive to improving the accuracy of detecting abnormal operation status of the waste heat recovery system.

[0061] In one embodiment of the present disclosure, a heat recovery effect anchor point of the sequence of heat recovery local timing effect feature vectors is calculated, and the cosine similarity of each heat recovery local timing effect feature vector in the sequence of heat recovery local timing effect feature vectors relative to the heat recovery effect anchor point is calculated to obtain a recovery effect timing fluctuation feature, including: calculating a global mean vector of the sequence of heat recovery local timing effect feature vectors as the heat recovery effect anchor point; calculating the cosine similarity of each heat recovery local timing effect feature vector in the sequence of heat recovery local timing effect feature vectors relative to the heat recovery effect anchor point to obtain a heat recovery effect timing fluctuation feature vector composed of multiple cosine similarities as the heat recovery effect timing fluctuation feature.

[0062] Then, the time series fluctuation feature vector of the heat recovery effect is passed through a classifier-based monitor to obtain a monitoring result, which is used to indicate whether there is an abnormality in the operation state of the waste heat recovery system. In other words, the time series fluctuation feature information of the heat recovery effect is used for classification processing, so as to detect abnormal operation state of the waste heat recovery system based on the time series fluctuation semantics of the heat recovery effect. In this way, real-time monitoring and abnormality judgment of the waste heat recovery system can be achieved, so as to timely discover problems and abnormalities in the system, such as pipeline blockage, heat exchanger leakage, etc., which is conducive to taking timely measures to repair and avoid system performance degradation or even safety accidents.

[0063] In one embodiment of the present disclosure, based on the time-series fluctuation characteristics of the recovery effect, it is determined whether there is any abnormality in the operating state of the waste heat recovery system, including: passing the time-series fluctuation characteristic vector of the heat recovery effect through a classifier-based monitor to obtain a monitoring result, and the monitoring result is used to indicate whether there is any abnormality in the operating state of the waste heat recovery system.

[0064] In one embodiment of the present disclosure, the real-time monitoring method of the boiler flue gas waste heat recovery system also includes a training step: for training the temperature time series pattern feature extractor based on the one-dimensional extended convolution layer, the heat recovery effect time series feature extractor and the classifier-based monitor. The training step includes: obtaining a time series of training flue gas inlet temperature and a time series of training flue gas outlet temperature collected by a first temperature sensor and a second temperature sensor deployed at the inlet and outlet of a flue gas cooler, as well as a real value of whether the operating state of a waste heat recovery system is abnormal; performing sequence segmentation and data regularization on the time series of the training flue gas inlet temperature and the time series of the training flue gas outlet temperature at the same time scale to obtain a sequence of local time series input vectors of the training flue gas inlet temperature and a sequence of local time series input vectors of the training flue gas outlet temperature; performing feature extraction on the sequence of local time series input vectors of the training flue gas inlet temperature and the sequence of local time series input vectors of the training flue gas outlet temperature respectively through the temperature time series pattern feature extractor based on a deep neural network to obtain a sequence of local time series associated feature vectors of the training flue gas inlet temperature and a sequence of local time series associated feature vectors of the training flue gas outlet temperature; and combining the sequence of local time series associated feature vectors of the training flue gas inlet temperature and the sequence of local time series associated feature vectors of the training flue gas outlet temperature. Each group of corresponding training flue gas inlet temperature local time series association feature vectors and training flue gas outlet temperature local time series association feature vectors in the sequence is input into the heat recovery effect time series feature extractor to obtain a sequence of training heat recovery local time series effect feature vectors; the global mean vector of the sequence of training heat recovery local time series effect feature vectors is calculated as the training heat recovery effect anchor point; the training cosine similarity of each training heat recovery local time series effect feature vector in the sequence of training heat recovery local time series effect feature vectors relative to the training heat recovery effect anchor point is calculated to obtain a training heat recovery effect time series fluctuation feature vector composed of multiple training cosine similarities; the training heat recovery effect time series fluctuation feature vector is passed through the classifier-based monitor to obtain a classification loss function value; based on the classification loss function value, the temperature time series pattern feature extractor based on the one-dimensional expanded convolutional layer, the heat recovery effect time series feature extractor and the classifier-based monitor are trained, and in each iteration, the training heat recovery effect time series fluctuation feature vector is corrected.

[0065] Further, the training heat recovery effect time series fluctuation feature vector is passed through the classifier-based monitor to obtain a classification loss function value, including: using the classifier-based monitor to process the training heat recovery effect time series fluctuation feature vector with the following training classification formula to generate a training classification result, wherein the training classification formula is: softmax{(W n ,B n):…:(W1,B1)|X}, where X represents the time series fluctuation feature vector of the training heat recovery effect, W1 to W n is the weight matrix, B1 to B n represents a bias matrix; and calculating a cross entropy value between the training classification result and the true value as the classification loss function value.

[0066] In the technical solution described above, the sequence of local temporal association feature vectors of the training flue gas inlet temperature and the sequence of local temporal association feature vectors of the training flue gas outlet temperature respectively express the temporal association characteristics of the training flue gas inlet temperature and the training flue gas outlet temperature in the local time domain, thereby, the sequence of training heat recovery local temporal effect feature vectors expresses the high-order temporal association characteristics based on the heat recovery effect in the local time domain of the training flue gas inlet temperature and the training flue gas outlet temperature, so that each training heat recovery local temporal effect feature vector in the sequence of training heat recovery local temporal effect feature vectors has the temporal association feature expression complexity under the high-order temporal structure related to the local-global time domain relative to the training cosine similarity of the training heat recovery effect anchor point, that is, the training heat recovery effect temporal fluctuation feature vector will have rich geometric shape details in the high-dimensional temporal feature space, so that there is category representation distortion when it is classified by the classifier to map to the class probability domain, that is, it affects the category representation of the training heat recovery effect temporal fluctuation feature vector for the classifier predetermined category, thereby reducing the accuracy of the classification result.

[0067] Based on this, the applicant of the present application corrects the training heat recovery effect time series fluctuation feature vector each time the training heat recovery effect time series fluctuation feature vector is classified and iterated by the classifier, which is expressed as: the training heat recovery effect time series fluctuation feature vector is corrected by the following optimization formula to obtain the corrected training heat recovery effect time series fluctuation feature vector; wherein the optimization formula is:

[0068]

[0069] Where p is the class probability value obtained by the classifier of the training heat recovery effect time series fluctuation feature vector V, v i is the i-th eigenvalue of the training heat recovery effect time series fluctuation eigenvector V, is the mean of all eigenvalues ​​of the training heat recovery effect time series fluctuation feature vector V, ||V||1 represents the norm of the training heat recovery effect time series fluctuation feature vector V, and α is the weight hyperparameter, v' i is the i-th eigenvalue of the time series fluctuation eigenvector of the training heat recovery effect after correction, and exp(·) represents the calculation of the natural exponential function value with the value as the power.

[0070] Specifically, while maintaining the geometric shape details of the high-dimensional feature manifold under the class probability mapping of the training heat recovery effect time series fluctuation feature vector V, the shape attribute editing is performed on the feature value details relative to the variational representation of the feature set of the training heat recovery effect time series fluctuation feature vector V by means of class probability distortion consultation, thereby mapping the class probability distortion of the training heat recovery effect time series fluctuation feature vector V to the latent class space feature representation, and then consulting fusion is performed by supplementing the basic low-rank constrained latent inversion representation to bridge the gap between the edited consulting representation and the original geometric shape details, thereby improving the class representation of the high-dimensional features of the training heat recovery effect time series fluctuation feature vector V, thereby improving the accuracy of the classification results. In this way, the heat recovery effect can be detected and evaluated based on the time series collaborative correlation information of the flue gas inlet temperature and the flue gas outlet temperature, so as to more accurately detect and judge the abnormal operation status of the waste heat recovery system based on the time series fluctuation semantics of the heat recovery effect, so as to timely discover problems in the system.

[0071] In summary, the above scheme is adopted to determine whether there is any abnormality in the operation status of the waste heat recovery system by real-time monitoring of the flue gas temperature. Specifically, the first temperature sensor and the second temperature sensor deployed at the inlet and outlet of the flue gas cooler monitor and collect the flue gas inlet temperature and the flue gas outlet temperature data in real time, and introduce artificial intelligence-based data processing and analysis algorithms at the back end to perform time series collaborative correlation analysis of the flue gas inlet temperature and the flue gas outlet temperature, so as to detect and evaluate the heat recovery effect, thereby detecting abnormal operation status of the waste heat recovery system. In this way, real-time monitoring and abnormal judgment of the waste heat recovery system can be achieved, so as to timely discover problems and abnormalities in the system, such as pipe blockage, heat exchanger leakage, etc., which is conducive to taking timely measures to repair and avoid system performance degradation or even safety accidents.

[0072] Figure 2 FIG. 1 is a block diagram of a real-time monitoring system for a boiler flue gas waste heat recovery system according to an exemplary embodiment. Figure 2 As shown, the system 200 includes:

[0073] The smoke inlet and outlet temperature acquisition module 201 is used to acquire the time series of the smoke inlet temperature and the time series of the smoke outlet temperature acquired by the first temperature sensor and the second temperature sensor deployed at the inlet and outlet of the smoke cooler;

[0074] A sequence segmentation and data regularization module 202 is used to perform sequence segmentation and data regularization on the time series of the smoke inlet temperature and the time series of the smoke outlet temperature at the same time scale to obtain a sequence of local time series input vectors of the smoke inlet temperature and a sequence of local time series input vectors of the smoke outlet temperature;

[0075] The temperature time series pattern feature extraction module 203 is used to extract features from the sequence of the local time series input vectors of the smoke inlet temperature and the sequence of the local time series input vectors of the smoke outlet temperature respectively through a temperature time series pattern feature extractor based on a deep neural network to obtain a sequence of local time series associated feature vectors of the smoke inlet temperature and a sequence of local time series associated feature vectors of the smoke outlet temperature;

[0076] The heat recovery effect time series feature extraction module 204 is used to input each group of corresponding smoke inlet temperature local time series correlation feature vectors and smoke outlet temperature local time series correlation feature vectors in the sequence of the smoke inlet temperature local time series correlation feature vectors and the sequence of the smoke outlet temperature local time series correlation feature vectors into the heat recovery effect time series feature extractor to obtain a sequence of heat recovery local time series effect feature vectors;

[0077] A recovery effect time series fluctuation feature generation module 205 is used to calculate the heat recovery effect anchor point of the sequence of the heat recovery local time series effect feature vectors, and calculate the cosine similarity of each heat recovery local time series effect feature vector in the sequence of the heat recovery local time series effect feature vectors relative to the heat recovery effect anchor point to obtain the recovery effect time series fluctuation feature;

[0078] The operation status abnormality judgment module 206 is used to determine whether the operation status of the waste heat recovery system is abnormal based on the time series fluctuation characteristics of the recovery effect.

[0079] Reference below Figure 3 , which shows a schematic diagram of the structure of an electronic device 600 suitable for implementing the embodiment of the present disclosure. The terminal device in the embodiment of the present disclosure may include but is not limited to mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), vehicle terminals (such as vehicle navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 3 The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0080] like Figure 3As shown, the electronic device 600 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 601, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage device 608 into a random access memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the electronic device 600 are also stored. The processing device 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0081] Typically, the following devices may be connected to the I / O interface 605: an input device 606 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 607 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 608 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 609. The communication device 609 may allow the electronic device 600 to communicate with other devices wirelessly or by wire to exchange data. Although Figure 3 The electronic device 600 is shown with various devices, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead.

[0082] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program contains program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication device 609, or installed from the storage device 608, or installed from the ROM 602. When the computer program is executed by the processing device 601, the above-mentioned functions defined in the method of the embodiment of the present disclosure are executed.

[0083] It should be noted that the computer-readable medium disclosed above may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, device or device. In the present disclosure, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, in which a computer-readable program code is carried. This propagated data signal may take a variety of forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. The computer readable signal medium may also be any computer readable medium other than a computer readable storage medium, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer readable medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0084] In some embodiments, the client and the server may communicate using any currently known or future developed network protocol such as HTTP (HyperText Transfer Protocol), and may be interconnected with any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network ("LAN"), a wide area network ("WAN"), an internet (e.g., the Internet), and a peer-to-peer network (e.g., an ad hoc peer-to-peer network), as well as any currently known or future developed network.

[0085] The computer-readable medium may be included in the electronic device, or may exist independently without being incorporated into the electronic device.

[0086] Computer program code for performing the operations of the present disclosure may be written in one or more programming languages ​​or a combination thereof, including, but not limited to, object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0087] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present disclosure. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some implementations as replacements, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0088] The modules involved in the embodiments described in the present disclosure may be implemented by software or hardware. The name of a module does not limit the module itself in some cases. For example, a test parameter acquisition module may also be described as a "module for acquiring device test parameters corresponding to a target device".

[0089] The functions described above herein may be performed at least in part by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), and the like.

[0090] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0091] Figure 4 is an application scenario diagram of a real-time monitoring method for a boiler flue gas waste heat recovery system according to an exemplary embodiment. Figure 4 As shown, in this application scenario, first, the time series of the flue gas inlet temperature collected by the first temperature sensor and the second temperature sensor deployed at the inlet and outlet of the flue gas cooler is obtained (for example, Figure 4 C1) and the time series of flue gas outlet temperature (e.g., Figure 4 Then, the acquired time series of the flue gas inlet temperature and the time series of the flue gas outlet temperature are input to a server (for example, Figure 4 In S) shown in , the server is able to process the time series of the flue gas inlet temperature and the time series of the flue gas outlet temperature based on the real-time monitoring algorithm of the boiler flue gas waste heat recovery system to determine whether there is any abnormality in the operating status of the waste heat recovery system.

[0092] The above description is only a preferred embodiment of the present disclosure and an explanation of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in the present disclosure is not limited to the technical solutions formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above disclosed concept. For example, the above features are replaced with the technical features with similar functions disclosed in the present disclosure (but not limited to) by each other to form a technical solution.

[0093] In addition, although each operation is described in a specific order, this should not be understood as requiring these operations to be performed in the specific order shown or in a sequential order. Under certain circumstances, multitasking and parallel processing may be advantageous. Similarly, although some specific implementation details are included in the above discussion, these should not be interpreted as limiting the scope of the present disclosure. Some features described in the context of a separate embodiment can also be implemented in a single embodiment in combination. On the contrary, the various features described in the context of a single embodiment can also be implemented in multiple embodiments individually or in any suitable sub-combination mode.

[0094] Although the subject matter has been described in language specific to structural features and / or method logic actions, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. On the contrary, the specific features and actions described above are merely example forms of implementing the claims. Regarding the device in the above embodiment, the specific manner in which each module performs the operation has been described in detail in the embodiment related to the method, and will not be elaborated here.

Claims

1. A real-time monitoring method for a boiler flue gas waste heat recovery system, characterized in that: include: Acquire a time series of flue gas inlet temperature and a time series of flue gas outlet temperature collected by a first temperature sensor and a second temperature sensor deployed at an inlet and an outlet of a flue gas cooler; Performing sequence segmentation and data regularization on the time series of the smoke inlet temperature and the time series of the smoke outlet temperature at the same time scale to obtain a sequence of local time series input vectors of the smoke inlet temperature and a sequence of local time series input vectors of the smoke outlet temperature; The temperature time series pattern feature extractor based on a deep neural network is used to extract features from the sequence of local time series input vectors of the smoke inlet temperature and the sequence of local time series input vectors of the smoke outlet temperature to obtain a sequence of local time series associated feature vectors of the smoke inlet temperature and a sequence of local time series associated feature vectors of the smoke outlet temperature; Input each group of corresponding local time series correlation feature vectors of flue gas inlet temperature and local time series correlation feature vectors of flue gas outlet temperature in the sequence of local time series correlation feature vectors of flue gas inlet temperature and the sequence of local time series correlation feature vectors of flue gas outlet temperature into a heat recovery effect time series feature extractor to obtain a sequence of heat recovery local time series effect feature vectors, wherein the heat recovery effect time series feature extractor processes the sequence of heat recovery local time series effect feature vectors using the following heat recovery effect feature extraction formula; Wherein, the heat recovery effect feature extraction formula is: ； in, and are the local time series correlation feature vectors of the smoke inlet temperature and the local time series correlation feature vectors of the smoke outlet temperature corresponding to each group in the sequence of the local time series correlation feature vectors of the smoke inlet temperature and the sequence of the local time series correlation feature vectors of the smoke outlet temperature, respectively. For vector The square of the norm, is a hyperparameter, is each heat recovery local timing effect feature vector in the sequence of heat recovery local timing effect feature vectors; Calculating a heat recovery effect anchor point of the sequence of heat recovery local time series effect feature vectors, and calculating the cosine similarity of each heat recovery local time series effect feature vector in the sequence of heat recovery local time series effect feature vectors relative to the heat recovery effect anchor point to obtain a recovery effect time series fluctuation feature; Based on the time series fluctuation characteristics of the recovery effect, it is determined whether there is any abnormality in the operation state of the waste heat recovery system.

2. The real-time monitoring method of the boiler flue gas waste heat recovery system according to claim 1 is characterized in that: The temperature time series pattern feature extractor based on deep neural network is a temperature time series pattern feature extractor based on a one-dimensional extended convolutional layer.

3. The real-time monitoring method of the boiler flue gas waste heat recovery system according to claim 2 is characterized in that: The temperature time series pattern feature extractor based on a deep neural network is used to perform feature extraction on the sequence of the local time series input vectors of the smoke inlet temperature and the sequence of the local time series input vectors of the smoke outlet temperature to obtain a sequence of local time series associated feature vectors of the smoke inlet temperature and a sequence of local time series associated feature vectors of the smoke outlet temperature, including: the sequence of the local time series input vectors of the smoke inlet temperature is processed by the temperature time series pattern feature extractor based on the one-dimensional extended convolution layer according to the following time series pattern feature extraction formula to obtain a sequence of local time series associated feature vectors of the smoke inlet temperature; Among them, the temporal pattern feature extraction formula is: ； ； in, , ,..., respectively represent each local time series input vector of the flue gas inlet temperature in the sequence of the local time series input vector of the flue gas inlet temperature, A cascade vector representing a sequence of local time series input vectors of the flue gas inlet temperature, Indicates a connection operation. The sequence of local time series input vectors of the flue gas inlet temperature is represented by , ,..., The concatenated vector of and Represent the weight matrix and bias vector respectively, represents the convolution operation, The first in the sequence of local time series correlation feature vectors of the flue gas inlet temperature The local time series correlation feature vector of the flue gas inlet temperature.

4. The real-time monitoring method of the boiler flue gas waste heat recovery system according to claim 3 is characterized in that: Calculating the heat recovery effect anchor point of the sequence of the heat recovery local time series effect feature vectors, and calculating the cosine similarity of each heat recovery local time series effect feature vector in the sequence of the heat recovery local time series effect feature vectors relative to the heat recovery effect anchor point to obtain the recovery effect time series fluctuation feature, including: Calculating a global mean vector of a sequence of the heat recovery local time series effect feature vectors as the heat recovery effect anchor point; The cosine similarity of each heat recovery local time series effect feature vector in the sequence of the heat recovery local time series effect feature vectors relative to the heat recovery effect anchor point is calculated to obtain a heat recovery effect time series fluctuation feature vector composed of multiple cosine similarities as the heat recovery effect time series fluctuation feature.

5. The real-time monitoring method of the boiler flue gas waste heat recovery system according to claim 4 is characterized in that: Based on the time series fluctuation characteristics of the recovery effect, determine whether there is an abnormality in the operating state of the waste heat recovery system, including: passing the heat recovery effect time series fluctuation characteristic vector through a classifier-based monitor to obtain a monitoring result, and the monitoring result is used to indicate whether there is an abnormality in the operating state of the waste heat recovery system.

6. The real-time monitoring method of the boiler flue gas waste heat recovery system according to claim 5 is characterized in that: It also includes a training step: used to train the temperature time series pattern feature extractor based on the one-dimensional extended convolution layer, the heat recovery effect time series feature extractor and the classifier-based monitor.

7. The real-time monitoring method of the boiler flue gas waste heat recovery system according to claim 6 is characterized in that: The training step comprises: Obtaining a time series of training flue gas inlet temperature and a time series of training flue gas outlet temperature collected by a first temperature sensor and a second temperature sensor deployed at the inlet and outlet of the flue gas cooler, as well as a real value of whether there is an abnormality in the operation status of the waste heat recovery system; Performing sequence segmentation and data regularization on the time series of the training smoke inlet temperature and the time series of the training smoke outlet temperature at the same time scale to obtain a sequence of local time series input vectors of the training smoke inlet temperature and a sequence of local time series input vectors of the training smoke outlet temperature; The temperature time series pattern feature extractor based on the deep neural network is used to extract features from the sequence of the training smoke inlet temperature local time series input vectors and the sequence of the training smoke outlet temperature local time series input vectors to obtain a sequence of training smoke inlet temperature local time series associated feature vectors and a sequence of training smoke outlet temperature local time series associated feature vectors; Inputting each group of corresponding training smoke inlet temperature local time series association feature vectors and training smoke outlet temperature local time series association feature vectors in the sequence of training smoke inlet temperature local time series association feature vectors and the sequence of training smoke outlet temperature local time series association feature vectors into the heat recovery effect time series feature extractor to obtain a sequence of training heat recovery local time series effect feature vectors; Calculating the global mean vector of the sequence of the training heat recovery local temporal effect feature vectors as the training heat recovery effect anchor point; Calculating the training cosine similarity of each training heat recovery local time series effect feature vector in the sequence of the training heat recovery local time series effect feature vectors relative to the training heat recovery effect anchor point to obtain a training heat recovery effect time series fluctuation feature vector composed of a plurality of training cosine similarities; Passing the training heat recovery effect time series fluctuation feature vector through the classifier-based monitor to obtain a classification loss function value; The temperature time series pattern feature extractor based on the one-dimensional expanded convolutional layer, the heat recovery effect time series feature extractor and the classifier-based monitor are trained based on the classification loss function value, and in each iteration, the training heat recovery effect time series fluctuation feature vector is corrected.

8. The real-time monitoring method of the boiler flue gas waste heat recovery system according to claim 7 is characterized in that: Passing the training heat recovery effect time series fluctuation feature vector through the classifier-based monitor to obtain a classification loss function value includes: The classifier-based monitor is used to process the training heat recovery effect time series fluctuation feature vector using the following training classification formula to generate a training classification result, wherein the training classification formula is: ,in represents the time series fluctuation feature vector of the training heat recovery effect, to is the weight matrix, to represents the bias matrix; and The cross entropy value between the training classification result and the true value is calculated as the classification loss function value.

Images