Boiler flue gas emission level evaluation method based on sample generation and ensemble learning

Abstract

The invention discloses a boiler flue gas emission level evaluation method based on sample generation and ensemble learning. The method comprises the following steps: determining flue gas component data of to-be-detected exhaust flue gas of a target boiler, the flue gas component data comprising a plurality of target components and a component value of each target component; determining a target feature vector corresponding to the smoke component data based on the target component and the component value; the input target feature vector is detected through an integrated learning model, a preliminary detection result corresponding to the to-be-detected emission flue gas is determined, and the integrated learning model comprises at least two of a random forest classifier, a logic review classifier and a decision tree classifier; and determining a target detection result of the to-be-detected discharged flue gas according to the preliminary detection result. According to the invention, the accuracy of the determined detection result of the discharged flue gas detection can be improved.

Description

Technical Field

[0001] The present invention relates to the field of computer application technology, and in particular to a boiler flue gas emission level assessment method based on sample generation and ensemble learning. Background Art

[0002] Boilers are indispensable equipment in industrial production and civil heating. At present, the application of boiler exhaust gas detection is very widespread. For example, exhaust gas detection is used to determine whether the exhaust gas meets environmental protection requirements, and exhaust gas detection is used to detect the combustion of fuel in the boiler.

[0003] In related technologies, the composition of exhaust gas is usually measured by a continuous emission monitoring system (CEMS), and the detection results are determined by judging whether the measured data of exhaust gas exceeds the standard. However, the detection standards are relatively simple, and the accuracy of exhaust gas detection is difficult to meet user requirements. Summary of the Invention

[0004] The present invention provides a boiler flue gas emission level assessment method based on sample generation and ensemble learning to solve the technical problem of poor accuracy in exhaust flue gas detection.

[0005] According to one aspect of the present invention, a method for evaluating boiler flue gas emission levels based on sample generation and ensemble learning is provided, wherein the method comprises:

[0006] Determining flue gas composition data of the flue gas to be tested emitted by the target boiler, wherein the flue gas composition data includes a plurality of target components and a component value of each target component;

[0007] determining a target feature vector corresponding to the smoke component data based on the target component and the component value;

[0008] Detecting the input target feature vector through an ensemble learning model to determine a preliminary detection result corresponding to the exhaust gas to be detected, wherein the ensemble learning model includes at least two of a random forest classifier, a logistic review classifier, and a decision tree classifier;

[0009] A target detection result of the exhaust gas to be detected is determined based on the preliminary detection result.

[0010] According to another aspect of the present invention, a device for evaluating boiler flue gas emission levels based on sample generation and ensemble learning is provided, wherein the device comprises:

[0011] a flue gas data acquisition module, configured to determine flue gas composition data of the flue gas to be detected emitted from a target boiler, wherein the flue gas composition data includes a plurality of target components and a component value of each target component;

[0012] a feature vector determination module, configured to determine a target feature vector corresponding to the smoke component data based on the target component and the component value;

[0013] a preliminary result determination module, configured to detect the input target feature vector using an ensemble learning model to determine a preliminary detection result corresponding to the exhaust gas to be detected, wherein the ensemble learning model includes at least two of a random forest classifier, a logistic regression classifier, and a decision tree classifier;

[0014] The target result determination module is used to determine the target detection result of the exhaust gas to be detected based on the preliminary detection result.

[0015] According to another aspect of the present invention, an electronic device is provided, comprising:

[0016] at least one processor; and

[0017] a memory communicatively connected to the at least one processor; wherein,

[0018] The memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the boiler flue gas emission level assessment method based on sample generation and integrated learning as described in any embodiment of the present invention.

[0019] According to another aspect of the present invention, a computer-readable storage medium is provided, which stores computer instructions, and the computer instructions are used to enable a processor to implement the boiler flue gas emission level assessment method based on sample generation and integrated learning as described in any embodiment of the present invention when executed.

[0020] The technical solution of the embodiment of the invention is to determine the flue gas composition data of the flue gas to be detected of the target boiler, wherein the flue gas composition data includes multiple target components and the component value of each target component; determine the target feature vector corresponding to the flue gas composition data based on the target components and the component values; detect the input target feature vector through an integrated learning model to determine the preliminary detection result corresponding to the flue gas to be detected, wherein the integrated learning model includes at least two of a random forest classifier, a logical review classifier, and a decision tree classifier; determine the target detection result of the flue gas to be detected based on the preliminary detection result. The present invention adopts the method of first determining the feature vector corresponding to the component in the flue gas, and then classifying the feature vector through a model integrating multiple different classifiers to perform exhaust gas detection, thereby improving the accuracy of the detection result of the determined exhaust gas detection.

[0021] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present invention, nor is it intended to limit the scope of the present invention. Other features of the present invention will become readily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0023] Figure 1 is a flow chart of a boiler flue gas emission level assessment method based on sample generation and ensemble learning according to the first embodiment of the present invention;

[0024] Figure 2 is a sorting data graph of key features provided according to an embodiment of the present invention;

[0025] Figure 3 is a result display diagram of a target detection result provided by an embodiment of the present invention;

[0026] Figure 4 is a flow chart of a boiler flue gas emission level assessment method based on sample generation and ensemble learning according to a second embodiment of the present invention;

[0027] Figure 5 is a data graph for generating a sample provided by an embodiment of the present invention;

[0028] Figure 6 is a comparative data diagram of an abnormal sample and a generated sample provided according to an embodiment of the present invention;

[0029] Figure 7 This is an overall flow chart of a flue gas emission detection method provided according to an embodiment of the present invention;

[0030] Figure 8 2 is a schematic structural diagram of a boiler flue gas emission level assessment device based on sample generation and ensemble learning according to a third embodiment of the present invention;

[0031] Figure 9 It is a structural diagram of an electronic device for implementing the boiler flue gas emission level assessment method based on sample generation and ensemble learning according to an embodiment of the present invention. DETAILED DESCRIPTION

[0032] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0033] It should be noted that the terms "first", "second", etc. in the description and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the numbers used in this way can be interchanged where appropriate so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0034] Example 1

[0035] Figure 1 A flowchart of a boiler flue gas emission level assessment method based on sample generation and integrated learning is provided for the first embodiment of the present invention. This embodiment is applicable to the case of abnormal detection of boiler flue gas emissions. The method can be performed by a boiler flue gas emission level assessment device based on sample generation and integrated learning. The boiler flue gas emission level assessment device based on sample generation and integrated learning can be implemented in the form of hardware and / or software. The boiler flue gas emission level assessment device based on sample generation and integrated learning can be configured in a computer. Figure 1 As shown, the method includes:

[0036] S110 : Determine flue gas composition data of the flue gas to be detected emitted from the target boiler, wherein the flue gas composition data includes a plurality of target components and a component value of each target component.

[0037] The target boiler can be understood as the boiler for which flue gas emission testing is to be performed. In the embodiments of the present invention, the target boiler can be preset based on the scenario requirements and is not specifically limited here. Optionally, the target boiler can include a domestic boiler, an industrial boiler, a power station boiler, or a vehicle or ship boiler.

[0038] The exhaust gas to be detected may be understood as the exhaust gas of the target boiler, or may be understood as the exhaust gas to be detected for abnormality.

[0039] The smoke composition data may be understood as the composition data of the exhaust smoke to be detected. In an embodiment of the present invention, the smoke composition data may include the target components contained in the exhaust smoke to be detected, and the component value corresponding to each target component.

[0040] In the embodiment of the present invention, the target component is related to the actual detected exhaust gas to be detected, and is not specifically limited herein. For example, the target component may include one or more of O2, NOx, SO2, smoke, flue gas flow rate, flue gas pressure, flue gas humidity, and flue gas temperature.

[0041] Correspondingly, the component value may be one or more of a concentration value, a flow rate value, a pressure value, a humidity value, and a temperature value.

[0042] Specifically, the flue gas composition data of the flue gas to be tested emitted by the target boiler may be obtained through a continuous emission monitoring system (CEMS) corresponding to the target boiler.

[0043] S120. Determine a target feature vector corresponding to the flue gas component data based on the target component and the component value.

[0044] The target feature vector may be understood as a vector including the features of the target components in the smoke composition data and the component values ​​corresponding to each target component.

[0045] Optionally, determining a target feature vector corresponding to the smoke component data based on the target component and the component value includes:

[0046] determining a plurality of preliminary features corresponding to the smoke composition data based on the target component and the component value;

[0047] determining a preset number of key features from the plurality of preliminary features by a recursive feature elimination algorithm and a random forest classifier;

[0048] A target feature vector corresponding to the smoke composition data is determined according to the key feature.

[0049] The preliminary features may be understood as the features corresponding to the component values ​​of each target component. The key features may be understood as the features among the preliminary features that have a greater impact on exhaust gas detection. In an embodiment of the present invention, the key features may be some of the preliminary features. The preset number of key features determined may be preset based on exhaust gas detection requirements and is not specifically limited herein.

[0050] Figure 2 is a sorting data diagram of key features provided according to an embodiment of the present invention. Figure 2 As shown, for example, the key features may include O2 concentration value, NOx concentration value and SO2 concentration value.

[0051] The Recursive Feature Elimination (RFE) algorithm can be understood as a model-based feature selection algorithm. Specifically, by constructing a random forest model (i.e., a random forest classifier) ​​and repeatedly training the model to evaluate the importance of each preliminary feature, it recursively eliminates unimportant features to obtain the optimal feature subset, i.e., the subset of key features.

[0052] S130. Detect the input target feature vector through an ensemble learning model to determine a preliminary detection result corresponding to the exhaust gas to be detected, wherein the ensemble learning model includes at least two of a random forest classifier, a logical review classifier, and a decision tree classifier.

[0053] The ensemble learning model may be understood as a learning model that integrates multiple classifiers. In an embodiment of the present invention, the ensemble learning model may include at least two of a random forest classifier, a logic review classifier, and a decision tree classifier.

[0054] The random forest classifier can be understood as a classifier that integrates multiple trees.

[0055] The logistic regression classifier can be understood as a linear regression classifier.

[0056] The decision tree classifier can be understood as a classifier that classifies data based on preset decision rules.

[0057] In the embodiment of the present invention, the classification logics corresponding to the random forest classifier, the logic review classifier, and the decision tree classifier may be different or the same.

[0058] The preliminary detection result can be understood as the vector classification result determined by each classifier in the ensemble learning model, wherein the vector classification result is the classification result of the target feature vector.

[0059] Optionally, the preliminary detection result includes at least two of the first detection result of the random forest classifier, the second detection result of the logical review classifier, and the third detection result of the decision tree classifier.

[0060] Optionally, detecting the input target feature vector by using an integrated learning model to determine a preliminary detection result corresponding to the exhaust gas to be detected includes:

[0061] The target feature vector is input into the random forest classifier, the logic review classifier and the decision tree classifier respectively, and the first detection result, the second detection result and the third detection result are obtained in sequence.

[0062] The first test result may include normal and / or abnormal. The second test result may include normal and / or abnormal. The third test result may include normal and / or abnormal.

[0063] In the embodiment of the present invention, the first detection result, the second detection result, and the third detection result may be the same or different.

[0064] S140: Determine a target detection result of the exhaust gas to be detected based on the preliminary detection result.

[0065] The target detection result may be understood as the target result of the exhaust smoke detection. Optionally, the exhaust smoke detection may include normal smoke detection and / or abnormal smoke detection.

[0066] Optionally, determining a target detection result of the exhaust gas to be detected based on the preliminary detection result includes:

[0067] Determine a first weight corresponding to the first detection result, a second weight corresponding to the second detection result, and a third weight corresponding to the third detection result;

[0068] A target detection result is determined according to the first detection result, the first weight, the second detection result, the second weight, the third detection result, and the third weight.

[0069] The first weight, the second weight, and the third weight can be understood as the calculation weights corresponding to the first detection result, the second detection result, and the third detection result, respectively. The calculation weights represent the importance of each preliminary detection result in the process of calculating the target detection result.

[0070] In the embodiment of the present invention, the first weight, the second weight, and the third weight can be preset according to the actual needs of exhaust gas detection, and are not specifically limited here. The first weight, the second weight, and the third weight can be different or the same.

[0071] Optionally, after determining the target detection result of the exhaust gas to be detected according to the preliminary detection result, the method further includes:

[0072] When the target detection result indicates that the smoke detection is abnormal, generating an abnormality prompt signal;

[0073] The abnormal prompt signal is sent to an abnormal alarm device installed on the target boiler, so that the abnormal alarm device issues a flue gas abnormality alarm.

[0074] The abnormal prompt signal may be understood as a signal indicating abnormality in exhaust gas detection.

[0075] The abnormal alarm device can be understood as a device with a smoke abnormality alarm function. In the embodiment of the present invention, the abnormal alarm device can be set and installed according to the scene requirements, which will not be detailed here.

[0076] Optionally, after determining the target detection result of the exhaust gas to be detected according to the preliminary detection result, the method further includes:

[0077] The target detection result is displayed in a target display mode.

[0078] In the embodiment of the present invention, the target display mode can be preset according to the scene requirements and is not specifically limited here. Figure 3 . Figure 3 This is a result display diagram of a target detection result provided according to an embodiment of the present invention.

[0079] The technical solution of the embodiment of the present invention is to determine the flue gas composition data of the flue gas to be detected of the target boiler, wherein the flue gas composition data includes multiple target components and the component value of each target component; determine the target feature vector corresponding to the flue gas composition data based on the target components and the component values; detect the input target feature vector through an integrated learning model to determine the preliminary detection result corresponding to the flue gas to be detected, wherein the integrated learning model includes at least two of a random forest classifier, a logical review classifier, and a decision tree classifier; determine the target detection result of the flue gas to be detected based on the preliminary detection result. The present invention adopts the method of first determining the feature vector corresponding to the component in the flue gas, and then classifying the feature vector through a model integrating multiple different classifiers to perform exhaust gas detection, thereby improving the accuracy of the detection result of the determined exhaust gas detection.

[0080] Example 2

[0081] Figure 4 This is a flow chart of a boiler flue gas emission level assessment method based on sample generation and ensemble learning provided in the second embodiment of the present invention. This embodiment is an addition to the detection of the input target feature vector by the ensemble learning model described in the above embodiment. Figure 4 As shown, the method includes:

[0082] S210 . Determine a flue gas abnormality sample set, wherein the flue gas abnormality sample set includes abnormal sample subsets of the target boiler in a startup state, a high load state, a low load state, and a shutdown state, respectively.

[0083] The abnormal smoke sample set may be understood as a collection of abnormal smoke samples.

[0084] In an embodiment of the present invention, the flue gas abnormality sample set includes subsets of abnormal samples of the target boiler in four states: startup, high load, low load, and shutdown. In other words, the flue gas abnormality sample set includes a subset of abnormal samples of the target boiler in the startup state, a subset of abnormal samples of the target boiler in the high load state, a subset of abnormal samples of the target boiler in the low load state, and a subset of abnormal samples of the target boiler in the shutdown state.

[0085] S220: For each of the abnormal sample subsets, determine a generated sample subset corresponding to the abnormal sample subset through a generative adversarial network.

[0086] The generative adversarial network can be used to generate samples with the same characteristics as abnormal samples.

[0087] The generated sample subset can be understood as a subset of generated samples (refer to Figure 5 and Figure 6 ), Figure 5 It is a data graph for generating samples provided according to an embodiment of the present invention. Figure 6 This is a comparative data diagram of an abnormal sample and a generated sample provided according to an embodiment of the present invention.

[0088] In an embodiment of the present invention, the generated sample subset may include a generated sample subset corresponding to an abnormal sample subset in a startup state, a generated sample subset corresponding to an abnormal sample subset in a high load state, a generated sample subset corresponding to an abnormal sample subset in a low load state, or a generated sample subset corresponding to an abnormal sample subset in a shutdown state.

[0089] S230: Determine a training abnormal sample set based on the multiple generated sample subsets and the abnormal smoke sample set.

[0090] The training abnormal sample set may be understood as the total set of the generated sample subset and the flue gas abnormal sample set.

[0091] S240 , determining a smoke normal sample set, taking the training abnormality sample set and the smoke normal sample set as a target training sample set, and determining a feature vector set corresponding to the target training sample set.

[0092] The smoke normal sample set may be understood as a set of smoke normal samples, or as a set of smoke normal samples.

[0093] The target training sample set can be understood as the total set of the flue gas normal sample set and the training abnormal sample set. The feature vector set can be understood as the set of feature vectors used to train the ensemble learning model. The feature vectors in the feature vector set correspond to the training samples in the target training sample set.

[0094] Optionally, determining the feature vector set corresponding to the target training sample set includes:

[0095] For each target training sample in the target training sample set, determining a plurality of sample features corresponding to the target training sample;

[0096] Determining a preset number of training features from the plurality of sample features by using a recursive feature elimination algorithm and a random forest classifier;

[0097] Determining a target training vector corresponding to the target training sample according to the training feature;

[0098] A feature vector set is determined according to the plurality of target training vectors.

[0099] The target training sample may include multiple sample components and sample values ​​corresponding to each sample component. The sample feature may be understood as a feature corresponding to the sample value of each sample component.

[0100] The training features can be understood as key features among the sample features.

[0101] The target training vector can be understood as a vector containing the features of the sample components in the target training sample and the sample values ​​corresponding to each sample component, and can also be understood as a training vector used to train the ensemble learning model.

[0102] The feature vector set can be understood as a set of target training vectors.

[0103] S250: Train a preliminary ensemble model using the target training sample set to obtain an ensemble learning model.

[0104] The preliminary integrated model can be understood as a preliminary constructed integrated model, and can include at least two of a preliminary constructed random forest classifier, a preliminary constructed logical review classifier, and a preliminary constructed decision tree classifier.

[0105] S260: Determine flue gas composition data of the flue gas to be detected emitted from the target boiler, wherein the flue gas composition data includes multiple target components and a component value of each target component.

[0106] S270. Determine a target feature vector corresponding to the flue gas component data based on the target component and the component value.

[0107] S280. Detect the input target feature vector through an integrated learning model to determine a preliminary detection result corresponding to the exhaust gas to be detected, wherein the integrated learning model includes at least two of a random forest classifier, a logical review classifier, and a decision tree classifier.

[0108] S290: Determine a target detection result of the exhaust gas to be detected based on the preliminary detection result.

[0109] The technical solution of the embodiment of the present invention is to determine a flue gas abnormality sample set, wherein the flue gas abnormality sample set includes abnormal sample subsets of the target boiler in the startup state, high load state, low load state, and shutdown state respectively; for each abnormal sample subset, determine a generated sample subset corresponding to the abnormal sample subset through a generative adversarial network; determine a training abnormal sample set based on multiple generated sample subsets and the flue gas abnormality sample set; determine a flue gas non-abnormality sample set, use the training abnormality sample set and the flue gas non-abnormality samples as the target training sample set, and determine a feature vector set corresponding to the target training sample set; train a preliminary ensemble model with the target training sample set to obtain an ensemble learning model. The present invention adopts a method of generating abnormal samples of the target boiler in multiple states, such as the startup state, high load state, low load state, and shutdown state, through a generative adversarial network, thereby improving the richness of abnormal samples and further improving the accuracy of the trained ensemble learning model.

[0110] Figure 7 This is an overall flow chart of a flue gas emission detection method provided according to an embodiment of the present invention.

[0111] like Figure 7 As shown, the overall process of the exhaust gas detection method can be:

[0112] 1. Generate abnormal flue gas sample data in four working states: startup, high load, low load, and shutdown using a generative adversarial network;

[0113] 2. Construct training sample data based on abnormal smoke sample data and normal smoke sample data;

[0114] 3. Use recursive feature elimination algorithm and random forest classifier to select key features that affect smoke emissions;

[0115] 4. Use the vectors corresponding to the selected important key features to train an ensemble learning model, where the ensemble learning model includes a random forest classifier, a logistic regression classifier, and a decision tree classifier;

[0116] 5. Obtain the actual flue gas emission data to be tested and determine the eigenvector corresponding to the data;

[0117] 6. Classify the input feature vector through the trained ensemble learning model to obtain three classification results;

[0118] 7. Use soft voting to determine the target results of exhaust gas detection corresponding to the three classification results.

[0119] The technical solution of the present invention solves the problem of scarcity of abnormal samples, enriches abnormal samples, and further improves the accuracy of the trained ensemble learning model; it adopts a method of first performing multi-branch classification through the ensemble learning model and then using soft voting to determine the target result of emission smoke detection, thereby improving the accuracy of the determined emission smoke detection result.

[0120] Example 3

[0121] Figure 8 This is a schematic diagram of the structure of a boiler flue gas emission level assessment device based on sample generation and integrated learning provided in the third embodiment of the present invention. Figure 8 As shown, the device includes: a smoke data acquisition module 310 , a feature vector determination module 320 , a preliminary result determination module 330 and a target result determination module 340 .

[0122] Among them, the flue gas data acquisition module 310 is used to determine the flue gas composition data of the flue gas to be detected of the target boiler, wherein the flue gas composition data includes multiple target components and the component value of each target component; the feature vector determination module 320 is used to determine the target feature vector corresponding to the flue gas composition data based on the target component and the component value; the preliminary result determination module 330 is used to detect the input target feature vector through an integrated learning model to determine the preliminary detection result corresponding to the flue gas to be detected, wherein the integrated learning model includes at least two of a random forest classifier, a logical review classifier and a decision tree classifier; the target result determination module 340 is used to determine the target detection result of the flue gas to be detected based on the preliminary detection result.

[0123] The technical solution of the embodiment of the invention is to determine the flue gas composition data of the flue gas to be detected of the target boiler, wherein the flue gas composition data includes multiple target components and the component value of each target component; determine the target feature vector corresponding to the flue gas composition data based on the target components and the component values; detect the input target feature vector through an integrated learning model to determine the preliminary detection result corresponding to the flue gas to be detected, wherein the integrated learning model includes at least two of a random forest classifier, a logical review classifier, and a decision tree classifier; determine the target detection result of the flue gas to be detected based on the preliminary detection result. The present invention adopts the method of first determining the feature vector corresponding to the component in the flue gas, and then classifying the feature vector through a model integrating multiple different classifiers to perform exhaust gas detection, thereby improving the accuracy of the detection result of the determined exhaust gas detection.

[0124] Optionally, the preliminary detection result includes at least two of the first detection result of the random forest classifier, the second detection result of the logical review classifier, and the third detection result of the decision tree classifier; the preliminary result determination module 330 is specifically configured to:

[0125] The target feature vector is input into the random forest classifier, the logic review classifier and the decision tree classifier respectively, and the first detection result, the second detection result and the third detection result are obtained in sequence.

[0126] Optionally, the target result determination module 340 is specifically configured to:

[0127] Determine a first weight corresponding to the first detection result, a second weight corresponding to the second detection result, and a third weight corresponding to the third detection result;

[0128] A target detection result is determined according to the first detection result, the first weight, the second detection result, the second weight, the third detection result, and the third weight.

[0129] Optionally, the feature vector determination module 320 is specifically configured to:

[0130] determining a plurality of preliminary features corresponding to the smoke composition data based on the target component and the component value;

[0131] determining a preset number of key features from the plurality of preliminary features by a recursive feature elimination algorithm and a random forest classifier;

[0132] A target feature vector corresponding to the smoke composition data is determined according to the key feature.

[0133] Optionally, the boiler flue gas emission level assessment device based on sample generation and integrated learning further includes: an abnormal sample determination module, an abnormal sample generation module, a training sample determination module, a training feature determination module, and a model training module;

[0134] The abnormal sample determination module is configured to determine a flue gas abnormal sample set before detecting the input target feature vector using the integrated learning model, wherein the flue gas abnormal sample set includes abnormal sample subsets of the target boiler in the startup state, high load state, low load state, and shutdown state respectively;

[0135] The abnormal sample generation module is configured to determine, for each abnormal sample subset, a generated sample subset corresponding to the abnormal sample subset through a generative adversarial network;

[0136] The training sample determination module is configured to determine a training abnormal sample set based on the plurality of generated sample subsets and the abnormal smoke sample set;

[0137] The training feature determination module is used to determine a smoke normal sample set, use the training abnormality sample set and the smoke normal sample set as a target training sample set, and determine a feature vector set corresponding to the target training sample set;

[0138] The model training module is used to train a preliminary integrated model using the target training sample set to obtain an integrated learning model.

[0139] Optionally, the training feature determination module is specifically configured to:

[0140] For each target training sample in the target training sample set, determining a plurality of sample features corresponding to the target training sample;

[0141] Determining a preset number of training features from the plurality of sample features by using a recursive feature elimination algorithm and a random forest classifier;

[0142] Determining a target training vector corresponding to the target training sample according to the training feature;

[0143] A feature vector set is determined according to the plurality of target training vectors.

[0144] Optionally, the target detection result includes normal flue gas detection and / or abnormal flue gas detection; the boiler flue gas emission level assessment device based on sample generation and integrated learning further includes: an abnormality detection module and an abnormality alarm module;

[0145] The abnormality detection module is configured to generate an abnormality prompt signal if the target detection result of the flue gas to be detected is abnormal after the target detection result of the flue gas to be detected is determined according to the preliminary detection result;

[0146] The abnormality alarm module is used to send the abnormality prompt signal to the abnormality alarm device installed on the target boiler, so that the abnormality alarm device can issue a flue gas abnormality alarm.

[0147] The boiler flue gas emission level assessment device based on sample generation and integrated learning provided in an embodiment of the present invention can execute the boiler flue gas emission level assessment method based on sample generation and integrated learning provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0148] Example 4

[0149] Figure 9A schematic diagram of the structure of an electronic device 10 that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or claimed herein.

[0150] like Figure 9 As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., which is communicatively connected to the at least one processor 11. The memory stores a computer program that can be executed by the at least one processor. The processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 into the random access memory (RAM) 13. Various programs and data required for the operation of the electronic device 10 can also be stored in the RAM 13. The processor 11, ROM 12, and RAM 13 are connected to each other via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0151] Multiple components in the electronic device 10 are connected to the I / O interface 15, including an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a magnetic disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices via a computer network such as the Internet and / or various telecommunication networks.

[0152] The processor 11 can be any general-purpose and / or specialized processing component with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various specialized artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any other suitable processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as the boiler flue gas emission level assessment method based on sample generation and ensemble learning.

[0153] In some embodiments, the boiler flue gas emission level assessment method based on sample generation and ensemble learning can be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as the storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the boiler flue gas emission level assessment method based on sample generation and ensemble learning described above can be performed. Alternatively, in other embodiments, the processor 11 can be configured to execute the boiler flue gas emission level assessment method based on sample generation and ensemble learning by any other appropriate means (for example, by means of firmware).

[0154] Various embodiments of the systems and techniques described herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), system-on-chip systems (SOCs), programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include being implemented in one or more computer programs that are executable and / or interpreted on a programmable system that includes at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0155] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when the computer program is executed by the processor, the functions / operations specified in the flowcharts and / or block diagrams are implemented. The computer program may be executed entirely on the machine, partially on the machine, as a stand-alone software package, partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0156] In the context of the present invention, computer-readable storage media can be tangible media that can contain or store a computer program for use with an instruction execution system, device or equipment or used in combination with an instruction execution system, device or equipment. Computer-readable storage media can include but are not limited to electronic, magnetic, optical, electromagnetic, infrared or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, computer-readable storage media can be machine-readable signal media. More specific examples of machine-readable storage media can include electrical connections based on one or more lines, portable computer disks, hard disks, random access memories (RAM), read-only memories (ROM), erasable programmable read-only memories (EPROM or flash memory), optical fibers, portable compact disk read-only memories (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0157] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0158] The systems and techniques described herein can be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system can be interconnected by 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), a blockchain network, and the Internet.

[0159] A computing system may include clients and servers. The clients and servers are typically remote from each other and typically interact via a communication network. This client-server relationship arises through computer programs running on the respective computers, creating a client-server relationship. The server may be a cloud server, also known as a cloud computing server or cloud host. This server is a hosting product within the cloud computing service ecosystem that addresses the management difficulties and limited scalability of traditional physical hosting and VPS services.

[0160] It should be understood that the various forms of the processes shown above can be used to reorder, add, or delete steps. For example, the steps described in the present invention can be performed in parallel, sequentially, or in a different order, as long as the desired results of the technical solution of the present invention can be achieved. This is not limited herein.

[0161] The above specific embodiments do not limit the scope of protection of the present invention. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions may be made based on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention are intended to be included within the scope of protection of the present invention.

Claims

1. A boiler flue gas emission level assessment method based on sample generation and ensemble learning, characterized in that: include: Determine the flue gas composition data of the flue gas to be detected from the target boiler, wherein the flue gas composition data includes a plurality of target components and a component value of each of the target components; Determining a target feature vector corresponding to the smoke component data based on the target component and the component value; The input target feature vector is detected by an integrated learning model to determine a preliminary detection result corresponding to the exhaust gas to be detected, wherein the integrated learning model includes at least two of a random forest classifier, a logical review classifier, and a decision tree classifier; The target detection result of the exhaust gas to be detected is determined according to the preliminary detection result.

2. The method according to claim 1, characterized in that: The preliminary detection result includes at least two of the first detection result of the random forest classifier, the second detection result of the logical review classifier, and the third detection result of the decision tree classifier; The step of detecting the input target feature vector by using an integrated learning model to determine a preliminary detection result corresponding to the exhaust gas to be detected includes: The target feature vector is respectively input into the random forest classifier, the logical review classifier and the decision tree classifier to obtain the first detection result, the second detection result and the third detection result in sequence.

3. The method according to claim 2, characterized in that Determining the target detection result of the exhaust gas to be detected according to the preliminary detection result includes: Determine a first weight corresponding to the first detection result, a second weight corresponding to the second detection result, and a third weight corresponding to the third detection result; A target detection result is determined according to the first detection result, the first weight, the second detection result, the second weight, the third detection result, and the third weight.

4. The method according to claim 1, characterized in that: The determining the target feature vector corresponding to the smoke component data based on the target component and the component value includes: Determining a plurality of preliminary features corresponding to the smoke component data based on the target component and the component value; Determining a preset number of key features from the plurality of preliminary features by a recursive feature elimination algorithm and a random forest classifier; A target feature vector corresponding to the smoke composition data is determined according to the key feature.

5. The method according to claim 1, characterized in that Before detecting the input target feature vector by using the integrated learning model, the method further includes: Determine a flue gas abnormal sample set, wherein the flue gas abnormal sample set includes abnormal sample subsets of the target boiler in a startup state, a high load state, a low load state, and a shutdown state; For each of the abnormal sample subsets, determining a generated sample subset corresponding to the abnormal sample subset by using a generative adversarial network; Determine a training abnormal sample set based on the plurality of generated sample subsets and the abnormal smoke sample set; Determine a smoke normal sample set, use the training abnormal sample set and the smoke normal sample set as a target training sample set, and determine a feature vector set corresponding to the target training sample set; The target training sample set is used to train a preliminary integrated model to obtain an integrated learning model.

6. The method according to claim 5, characterized in that The step of determining a feature vector set corresponding to the target training sample set includes: For each target training sample in the target training sample set, determining a plurality of sample features corresponding to the target training sample; Determining a preset number of training features from the plurality of sample features by a recursive feature elimination algorithm and a random forest classifier; Determine a target training vector corresponding to the target training sample according to the training feature; A feature vector set is determined according to the plurality of target training vectors.

7. The method according to claim 1, characterized in that The target detection result includes normal smoke detection and / or abnormal smoke detection; After determining the target detection result of the exhaust gas to be detected according to the preliminary detection result, the method further includes: When the target detection result is that the smoke detection is abnormal, generating an abnormality prompt signal; The abnormal prompt signal is sent to an abnormal alarm device installed on the target boiler, so that the abnormal alarm device issues a flue gas abnormality alarm.

8. A boiler flue gas emission level assessment device based on sample generation and ensemble learning, characterized in that: include: A flue gas data acquisition module, used to determine the flue gas composition data of the flue gas to be detected from the target boiler, wherein the flue gas composition data includes a plurality of target components and a component value of each of the target components; A feature vector determination module, used to determine a target feature vector corresponding to the smoke component data based on the target component and the component value; A preliminary result determination module, used to detect the input target feature vector through an integrated learning model to determine the preliminary detection result corresponding to the exhaust gas to be detected, wherein the integrated learning model includes at least two of a random forest classifier, a logical review classifier, and a decision tree classifier; The target result determination module is used to determine the target detection result of the exhaust gas to be detected according to the preliminary detection result.

9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can perform the boiler flue gas emission level assessment method based on sample generation and integrated learning as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the boiler flue gas emission level assessment method based on sample generation and integrated learning as described in any one of claims 1 to 7 when executed.

Images