A method and device for predicting oxygen content in flue gas
By obtaining the characteristic data and model training data of the target boiler and reference boiler, determining the error weight and performing model training, the problem of reducing model migration accuracy between different boilers is solved, and high accuracy of flue gas oxygen content prediction is achieved.
Patent Information
- Application Number
- CN202011012359.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2020-09-23
- Publication Date
- 2024-12-13
- Estimated Expiration
- 2040-09-23
AI Technical Summary
The historical operation data distribution of different boilers is different, resulting in a reduced accuracy when used on another boiler when the flue gas oxygen content prediction model learned on one boiler.
By obtaining the characteristic data and model training data of the target boiler and the reference boiler, the error weights are determined, and model training is performed based on these weights, a prediction model of the flue gas oxygen content for the target boiler is established.
The accuracy of the flue gas oxygen content prediction model is improved, and the accuracy of the flue gas oxygen content prediction results on the target boiler is ensured.
Smart Images
Figure CN114254792B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of energy technology, and in particular to a method and device for predicting oxygen content in flue gas. Background Art
[0002] Soft measurement is a technical solution that uses software technology instead of sensors to measure during the monitoring process of industrial equipment. Specifically, it uses sensors to measure variables at some measurement points, and infers or estimates variables at other measurement points through the association between algorithms and variables. Soft measurement can effectively reduce hardware costs and improve the robustness of measurement and control. Soft measurement of flue gas oxygen content refers to the use of historical operating data of the boiler, which includes gas flow, gas temperature, air temperature, exhaust temperature, flue gas flow, etc. The relationship between these historical operating data and flue gas oxygen content is established through machine learning methods, thereby predicting the flue gas oxygen content of the boiler.
[0003] However, the distribution of historical operating data of different boilers cannot be exactly the same. Therefore, using the flue gas oxygen content prediction model learned on one boiler for another boiler may result in a decrease in the accuracy of the flue gas oxygen content prediction model. Summary of the invention
[0004] The present invention provides a method and device for predicting the oxygen content in flue gas, a computer-readable storage medium and an electronic device, which can improve the accuracy of a flue gas oxygen content prediction model.
[0005] In a first aspect, the present invention provides a method for predicting oxygen content in flue gas, comprising:
[0006] Acquire at least two feature data corresponding to the target boiler, at least two model training data corresponding to the reference boiler, and historical flue gas oxygen contents corresponding to each of the model training data;
[0007] Determining the error weight corresponding to the model training data according to each of the feature data;
[0008] Model training is performed according to each of the model training data, the historical flue gas oxygen content corresponding to each of the model training data, and the error weight corresponding to each of the model training data to determine a flue gas oxygen content prediction model, which is used to predict the flue gas oxygen content of the target boiler.
[0009] In a second aspect, the present invention provides a flue gas oxygen content prediction device, comprising:
[0010] An acquisition module, used to acquire at least two characteristic data corresponding to the target boiler, at least two model training data corresponding to the reference boiler, and the historical flue gas oxygen content corresponding to each of the model training data;
[0011] A weight determination module, used to determine the error weight corresponding to the model training data according to each of the feature data;
[0012] A training module is used to perform model training according to each of the model training data, the historical flue gas oxygen content corresponding to each of the model training data, and the error weight corresponding to each of the model training data, so as to determine a flue gas oxygen content prediction model, wherein the flue gas oxygen content prediction model is used to predict the flue gas oxygen content of the target boiler.
[0013] In a third aspect, the present invention provides a computer-readable storage medium, comprising execution instructions. When a processor of an electronic device executes the execution instructions, the processor executes any method described in the first aspect.
[0014] In a fourth aspect, the present invention provides an electronic device, comprising a processor and a memory storing execution instructions, wherein when the processor executes the execution instructions stored in the memory, the processor executes any method described in the first aspect.
[0015] The present invention provides a flue gas oxygen content prediction method, device, computer-readable storage medium and electronic device. The method obtains multiple feature data corresponding to a target boiler, multiple model training data corresponding to a reference boiler and the historical flue gas oxygen content corresponding to each model training data. Then, for each model training data, according to each feature data, the error weight corresponding to the model training data is determined. The obtained error weight indicates the importance of the error generated in the model training process corresponding to the target boiler using the model training data, thereby achieving the goal of migrating the model training data of the reference boiler to the target boiler. Afterwards, model training is performed according to each model training data, the historical flue gas oxygen content corresponding to each model training data and the error weight corresponding to each model training data to determine the flue gas oxygen content prediction model. The obtained flue gas oxygen content prediction model takes into account the error weight of the model training data of the reference boiler relative to the target boiler, thereby ensuring the model accuracy. When the flue gas oxygen content prediction model is used to predict the flue gas oxygen content of the target boiler, the accuracy of the predicted flue gas oxygen content can be ensured. In summary, the technical solution of the present invention can improve the accuracy of the flue gas oxygen content prediction model, and when the flue gas oxygen content of the target boiler is subsequently predicted based on the flue gas oxygen content prediction model, the accuracy of the predicted flue gas oxygen content can be ensured.
[0016] The further effects of the above-mentioned non-conventional preferred manner will be described below in conjunction with specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS
[0017] In order to more clearly illustrate the embodiments of the present invention or the existing technical solutions, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.
[0018] Figure 1 A schematic flow chart of a method for predicting oxygen content in flue gas provided by one embodiment of the present invention;
[0019] Figure 2 A schematic flow chart of another method for predicting flue gas oxygen content provided by one embodiment of the present invention;
[0020] Figure 3 A schematic diagram of the structure of a flue gas oxygen content prediction device provided by one embodiment of the present invention;
[0021] Figure 4 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0022] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments 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 creative work are within the scope of protection of the present invention.
[0023] like Figure 1 The present invention provides a method for predicting oxygen content in flue gas, comprising the following steps:
[0024] Step 101: Obtain at least two feature data corresponding to a target boiler, at least two model training data corresponding to a reference boiler, and historical flue gas oxygen contents corresponding to each of the model training data.
[0025] Specifically, the operating data of the reference boiler can be obtained according to the sensor installed on the reference boiler connected to the network, and the operating data of the reference boiler can be processed to determine a plurality of model training data corresponding to the reference boiler. It should be noted that the process of determining the model training data based on the data collected by the sensor is a prior art, and the embodiments of the present invention will not be described in detail.
[0026] Specifically, the operating data of the target boiler can be obtained according to the sensor installed on the target boiler connected to the network, and the operating data of the target boiler can be processed to determine a plurality of characteristic data corresponding to the target boiler. It should be noted that determining characteristic data based on data collected by the sensor is a prior art, and the embodiments of the present invention will not be elaborated on this.
[0027] Specifically, the model training data includes several features and each feature corresponds to one or more characteristic values of the reference boiler, wherein several features are influencing factors that affect the flue gas oxygen content, for example, gas flow, gas temperature, air temperature, exhaust temperature, flue gas flow. Here, the model training data corresponds to a historical flue gas oxygen content. It should be noted that considering that the amount of model training data for a reference boiler is large, the number of reference boilers is one to ensure the accuracy of the flue gas oxygen content prediction model. It should be noted that the model training data and the feature data correspond to several features that are the same.
[0028] It should be noted that the embodiments of the present invention do not intend to limit the method for obtaining the historical flue gas oxygen content corresponding to each model training data. It can also be manual labeling, rule labeling, or cluster labeling. The above labeling methods are all existing technologies and will not be elaborated here.
[0029] Step 102: Determine the error weight corresponding to the model training data according to each of the feature data.
[0030] In one embodiment, the error weight corresponding to the model training data can be determined by the following method:
[0031] Determine the replacement data corresponding to the model training data from each feature data; determine the first distribution information of the model training data in each model training data and the second distribution information of the replacement data in each feature data; for each model training data, determine the error weight corresponding to the model training data according to the first distribution information of the model training data and the second distribution information of the replacement data corresponding to the model training data.
[0032] Specifically, the error weight can achieve the purpose of migrating the data of the reference boiler to the target boiler, in other words, establish a data relationship between the reference boiler and the target boiler. Specifically, by finding the replacement data corresponding to the model training data from the multiple feature data of the target boiler, here, the replacement data is preferably the same data as the model training data in the multiple feature data. If there is no data that is the same as the model training data in the multiple feature data, at this time, the replacement data is the data that is most similar to the model training data in the multiple feature data, and then, the first distribution information of the model training data in all the model training data of the reference boiler is determined, so as to understand the distribution of the model training data corresponding to the reference boiler, and the second distribution information of the replacement data in all the feature data corresponding to the target boiler is determined, so as to understand the distribution of the model training data corresponding to the target boiler. After that, for each model training data, according to the second distribution information of the model training data corresponding to the target boiler and the first distribution information corresponding to the reference boiler, the error weight of the model training data is determined, so as to achieve data association between the reference boiler and the target boiler.
[0033] Optionally, when the first distribution information and the second distribution information are both distribution probabilities, the error weight may be determined by the following method:
[0034] Determine the probability ratio of the distribution probability of the model training data and the distribution probability of the replacement data corresponding to the model training data, and determine the probability ratio as the error weight corresponding to the model training data.
[0035] Optionally, when the first distribution information and the second distribution information are both occurrence frequencies, the error weight may be determined by the following method:
[0036] Determine the frequency of occurrence of the model training data and the frequency ratio of the occurrence of the replacement data corresponding to the model training data, and determine the frequency ratio as the error weight corresponding to the model training data.
[0037] Optionally, when there are two or more replacement data in the multiple feature data, at this time, one replacement data may be randomly selected as the replacement data corresponding to the model training data.
[0038] Optionally, when there is only one replacement data among the multiple feature data, the replacement data is used as the replacement data corresponding to the model training data.
[0039] Considering that the non-parametric estimation method is not restricted by the overall distribution, does not assume the specific form of the overall distribution, and tries to obtain the required information from the data or sample itself, obtaining the distribution structure through estimation is extremely important for reducing bias, improving prediction accuracy, and understanding the dynamic structure of data. Especially when the data distribution is very irregular, the non-parametric estimation method is more accurate than the parameter estimation. Therefore, the non-parametric estimation method is used to determine the data distribution in the embodiment of the present invention.
[0040] Optionally, the first distribution information may be determined by the following method:
[0041] Based on a non-parametric estimation method, the distribution probability of the model training data in each model training data is determined, and the distribution probability is determined as the first distribution information.
[0042] The embodiment of the present invention is not intended to limit the non-parametric estimation method, and the kernel density estimation method is preferred.
[0043] Optionally, the first distribution information may be determined by the following method:
[0044] Obtain the data volume of each model training data; select a target non-parametric estimation method from at least two non-parametric estimation methods based on the data volume; determine the distribution probability of the model training data in each model training data based on the target non-parametric estimation method, and determine the distribution probability as the first distribution information.
[0045] By considering the amount of model training data, a suitable target nonparametric estimation method is selected from multiple nonparametric estimation methods, and based on the target nonparametric estimation method, the distribution probability of the model training data in each model training data is determined, thereby ensuring the reference value of the distribution probability, and then ensuring the reference value of the error weight.
[0046] Considering that the accuracy of distribution estimation of the same data by different non-parametric estimation methods is not exactly the same, in order to ensure the accuracy of the distribution probability, the data volume of each model training data corresponding to the reference boiler can be combined to automatically select a more appropriate non-parametric estimation method for distribution estimation. For example, a data volume interval can be determined in advance for each non-parametric estimation method, and the data volume interval to which the data volume of each model training data corresponding to the reference boiler belongs can be determined, thereby determining the target non-parametric estimation method.
[0047] Specifically, the at least two non-parametric estimation methods include but are not limited to any two or more of a kernel function method, a nearest neighbor function method, a spline function method, and a wavelet function method.
[0048] In one embodiment, the first distribution information may be determined by the following method:
[0049] Acquire at least two non-parametric estimation methods; for each non-parametric estimation method, determine the candidate distribution probability of the model training data in each model training data based on the non-parametric estimation method; fuse the candidate distribution probabilities to determine the target distribution probability; determine the target distribution probability as the distribution probability of the model training data in each model training data, and determine the distribution probability as the first distribution information.
[0050] By fusing the candidate distribution probabilities obtained by two or more non-parametric estimation methods, the accuracy of the obtained first distribution probability is ensured. It should be noted that the fusion of the candidate distribution probabilities can be an average or a weighted average, which is not specifically limited here.
[0051] Specifically, the at least two non-parametric estimation methods include but are not limited to any two or more of a kernel function method, a nearest neighbor function method, a spline function method, and a wavelet function method.
[0052] It should be noted that the method for determining the second distribution information and the first distribution information may be the same or different, preferably the same.
[0053] Step 103: Perform model training according to each of the model training data, the historical flue gas oxygen content corresponding to each of the model training data, and the error weight corresponding to each of the model training data to determine a flue gas oxygen content prediction model, wherein the flue gas oxygen content prediction model is used to predict the flue gas oxygen content of the target boiler.
[0054] In one embodiment, the flue gas oxygen content prediction model may be a neural network model or a regression model, which is not specifically limited here.
[0055] In one embodiment, the flue gas oxygen content prediction method provided by the embodiment of the present invention is applied to an edge server or a local server corresponding to a target boiler, thereby reducing the computing pressure of a cloud server.
[0056] Specifically, the flue gas oxygen content prediction model uses multiple model training data and the historical flue gas oxygen content corresponding to each model training data as training data, and adjusts the model parameters of the flue gas oxygen content prediction model through the error weights corresponding to each model training data, so that the adjusted model parameters can reflect the relationship between the characteristic data of the target boiler and the flue gas oxygen content, thereby ensuring the model accuracy of the flue gas oxygen content prediction model.
[0057] In one embodiment, the flue gas oxygen content prediction model can be determined by the following method:
[0058] A1. Determine the first errors corresponding to the model training data respectively according to the prediction results of substituting the model training data into the model to be predicted and the historical flue gas oxygen content corresponding to the model training data, and determine the second error corresponding to the model training data according to the first error corresponding to the model training data and the error weight;
[0059] A2, judging whether the number of iterations or the second errors corresponding to the training data of each model meet the preset conditions, if yes, determining the model to be predicted as the flue gas oxygen content prediction model, if no, executing A3;
[0060] A3. According to the second errors corresponding to the training data of each model, the model parameters in the model to be predicted are adjusted to determine the adjusted model parameters, and the model parameters in the model to be predicted are replaced with the adjusted model parameters, and A1 is executed.
[0061] Here, whether the second errors corresponding to each model training data meet the preset conditions can be whether the mean of the second errors corresponding to each model training data is less than a threshold, or whether the standard deviation of the second errors corresponding to each model training data is less than a threshold, which is not specifically limited here.
[0062] Here, the first error may be the difference between the predicted flue gas oxygen content and the historical flue gas oxygen content.
[0063] It can be seen from the above technical scheme that the beneficial effect of the embodiment of the present invention is: by determining the error weights of multiple model training data corresponding to a reference boiler relative to the target boiler, the multiple model training data corresponding to the reference boiler are migrated to the target boiler. Subsequently, the multiple model training data corresponding to the reference boiler, the historical flue gas oxygen content corresponding to each model training data, and the weight error corresponding to each model training data are used to obtain a flue gas oxygen content prediction model for predicting the flue gas oxygen content of the target boiler. The obtained flue gas oxygen content prediction model has relatively high accuracy because it comprehensively considers the error weight of the model training data for the target boiler. When the flue gas oxygen content prediction model of the target boiler is subsequently performed based on the flue gas oxygen content prediction model, a relatively accurate flue gas oxygen content of the target boiler can be obtained.
[0064] Figure 1 What is shown is only a basic embodiment of the method of the present invention. By performing certain optimization and expansion on this basis, other preferred embodiments of the method can be obtained.
[0065] like Figure 2 FIG. 1 is another specific embodiment of the method for predicting the oxygen content in flue gas according to the present invention. The embodiment of the present invention is described in more detail based on the above-mentioned embodiment and combined with the application scenario. The method specifically includes the following steps:
[0066] Step 201: Obtain at least two feature data corresponding to a target boiler, at least two model training data corresponding to a reference boiler, and historical flue gas oxygen contents corresponding to each of the model training data.
[0067] Assume that the reference boiler is boiler A, the target boiler is boiler B, and boiler A has N A model training data, each model training data corresponds to a historical flue gas oxygen content, boiler B corresponds to multiple N B Feature data. Feature data and model training data have multiple identical features.
[0068] Step 202, obtaining at least two non-parametric estimation methods; for each of the non-parametric estimation methods, determining the candidate distribution probability of the model training data in each of the model training data based on the non-parametric estimation method; fusing the respective candidate distribution probabilities to determine the target distribution probability, and determining the target distribution probability as the distribution probability of the model training data in each of the model training data.
[0069] Assume that there are four non-parametric estimation methods, set as C1, C2, C3, C4 respectively. For each model training data, determine the candidate distribution probabilities P1, P2, P3, P4 after C1, C2, C3, C4 respectively estimate their distributions. The average or weighted average of P1, P2, P3, P4 is determined as the training parameter of the model in N. A The distribution probability P in the model training data A .
[0070] Step 203: Determine the replacement data corresponding to the model training data from each of the replacement data; and determine the distribution probability of the replacement data in each of the feature data.
[0071] According to step 202, the distribution probability P is determined A In a similar way, determine the replacement data in N B The distribution probability P in the feature data B .
[0072] Step 204: for each of the model training data, determine the probability ratio of the distribution probability of the model training data and the distribution probability of the replacement data corresponding to the model training data, and determine the probability ratio as the error weight corresponding to the model training data.
[0073] For each model training data, the second distribution probability P of the replacement data corresponding to the model training data is B And the first distribution probability P of the model training data A The ratio S = PB / P A Determine the error weights corresponding to the training data for the model.
[0074] Step 205: Perform model training according to each of the model training data, the historical flue gas oxygen content corresponding to each of the model training data, and the error weight corresponding to each of the model training data to determine a flue gas oxygen content prediction model, wherein the flue gas oxygen content prediction model is used to predict the flue gas oxygen content of the target boiler.
[0075] Based on N A model training data and N A The historical flue gas oxygen content corresponding to each model training data is used to train the preset flue gas oxygen content prediction model. A The training data of each model corresponds to S = P B / P A , adjust the model parameters in the preset flue gas oxygen content prediction model, and finally obtain the flue gas oxygen content prediction model.
[0076] Based on the same concept as the embodiment of the method of the present invention, please refer to Figure 3 The embodiment of the present invention further provides a flue gas oxygen content prediction device, comprising:
[0077] An acquisition module 301 is used to acquire at least two feature data corresponding to a target boiler, at least two model training data corresponding to a reference boiler, and historical flue gas oxygen contents corresponding to each of the model training data;
[0078] A weight determination module 302 is used to determine the error weight corresponding to the model training data according to each of the feature data;
[0079] The training module 303 is used to perform model training according to each of the model training data, the historical flue gas oxygen content corresponding to each of the model training data, and the error weight corresponding to each of the model training data, so as to determine a flue gas oxygen content prediction model, wherein the flue gas oxygen content prediction model is used to predict the flue gas oxygen content of the target boiler.
[0080] In one embodiment of the present invention, the weight determination module 302 includes: a replacement data determination unit, a probability determination unit and a weight determination unit; wherein,
[0081] The replacement data determination unit is used to determine the replacement data corresponding to the model training data from each of the feature data;
[0082] The probability determination unit is used to determine the first distribution information of the model training data in each of the model training data and the second distribution information of the replacement data in each of the feature data;
[0083] The weight determination unit is used to determine, for each of the model training data, the error weight corresponding to the model training data according to the first distribution information of the model training data and the second distribution information of the replacement data corresponding to the model training data.
[0084] In one embodiment of the present invention, the weight determination unit is used to determine the probability ratio of the distribution probability of the model training data and the distribution probability of the replacement data corresponding to the model training data when the first distribution information and the second distribution information are distribution probabilities respectively, and determine the probability ratio as the error weight corresponding to the model training data.
[0085] In one embodiment of the present invention, the weight determination unit is used to determine the frequency ratio of the occurrence frequency of the model training data and the occurrence frequency of the replacement data corresponding to the model training data when the first distribution information and the second distribution information are occurrence frequencies respectively, and determine the frequency ratio as the error weight corresponding to the model training data.
[0086] In one embodiment of the present invention, the weight determination unit includes: a first acquisition subunit, a selection subunit and a first probability determination subunit; wherein,
[0087] The first acquisition subunit is used to acquire the data volume of each model training data;
[0088] The selection subunit is used to select a target non-parametric estimation method from at least two non-parametric estimation methods according to the data amount;
[0089] The first probability determination subunit is used to determine the distribution probability of the model training data in each of the model training data according to the target non-parametric estimation method, and determine the distribution probability as the first distribution information.
[0090] In one embodiment of the present invention, the weight determination unit includes: a second acquisition subunit, a distribution estimation subunit, a fusion unit and a second probability determination subunit; wherein,
[0091] The second acquisition subunit is used to acquire at least two non-parametric estimation methods;
[0092] The distribution estimation subunit is used to determine, for each of the non-parametric estimation methods, the candidate distribution probability of the model training data in each of the model training data based on the non-parametric estimation method;
[0093] The fusion subunit is used to fuse the candidate distribution probabilities to determine the target distribution probability;
[0094] The second probability determination subunit is used to determine the target distribution probability as the distribution probability of the model training data in each of the model training data, and determine the distribution probability as the first distribution information.
[0095] In one embodiment of the present invention, the probability determination unit is used to determine the distribution probability of the replacement data in each of the feature data based on a non-parametric estimation method, and determine the distribution probability as the second distribution information.
[0096] In one embodiment, the training module 303 includes: a training unit, a judgment unit and an adjustment unit; wherein:
[0097] The training unit is used to determine the first errors corresponding to the model training data respectively according to the prediction results of substituting the model training data into the model to be predicted and the historical flue gas oxygen content corresponding to the model training data, and to determine the second error corresponding to the model training data according to the first error corresponding to the model training data and the error weight;
[0098] The judging unit is used to judge whether the number of iterations or the second errors corresponding to each of the model training data meet the preset conditions, and if so, determine the model to be predicted as the flue gas oxygen content prediction model; if not, trigger the adjustment unit;
[0099] The adjustment unit is used to adjust the model parameters in the model to be predicted according to the second errors corresponding to each of the model training data to determine the adjusted model parameters, and replace the model parameters in the model to be predicted with the adjusted model parameters to trigger the training unit.
[0100] Figure 4It is a structural diagram of an electronic device provided by an embodiment of the present invention. At the hardware level, the electronic device includes a processor 401 and a memory 402 storing execution instructions, and optionally also includes an internal bus 403 and a network interface 404. Among them, the memory 402 may include a memory 4021, such as a high-speed random access memory (Random-Access Memory, RAM), and may also include a non-volatile memory 4022 (non-volatile memory), such as at least one disk storage, etc.; the processor 401, the network interface 404 and the memory 402 can be interconnected through an internal bus 403, and the internal bus 403 can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc.; the internal bus 403 can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 4 Only one bidirectional arrow is used in the figure, but it does not mean that there is only one bus or one type of bus. Of course, the electronic device may also include hardware required for other services. When the processor 401 executes the execution instruction stored in the memory 402, the processor 401 executes the method in any embodiment of the present invention, and is at least used to execute the following steps: Figure 1 or Figure 2 The method shown.
[0101] In a possible implementation, the processor reads the corresponding execution instruction from the non-volatile memory into the memory and then runs it, and can also obtain the corresponding execution instruction from other devices to form a flue gas oxygen content prediction device at the logical level. The processor executes the execution instruction stored in the memory to implement a flue gas oxygen content prediction method provided in any embodiment of the present invention through the execution of the execution instruction.
[0102] The processor may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by an integrated logic circuit of hardware in the processor or instructions in the form of software. The above processor may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it may also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The disclosed methods, steps and logic block diagrams in the embodiments of the present invention can be implemented or executed. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.
[0103] The embodiment of the present invention further provides a computer-readable storage medium, including an execution instruction. When a processor of an electronic device executes the execution instruction, the processor executes the method provided in any embodiment of the present invention. The electronic device may be specifically as follows: Figure 4 The electronic device shown; the execution instruction is a computer program corresponding to a flue gas oxygen content prediction device.
[0104] Those skilled in the art will appreciate that the embodiments of the present invention may be provided as methods or computer program products. Therefore, the present invention may be implemented in the form of a complete hardware embodiment, a complete software embodiment, or a combination of software and hardware.
[0105] The various embodiments of the present invention are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.
[0106] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.
[0107] The above description is only an embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent substitution, improvement, etc. made within the spirit and principle of the present invention should be included in the scope of the claims of the present invention.
Claims
1. A method for predicting flue gas oxygen content, characterized in that: include: Acquire at least two feature data corresponding to the target boiler, at least two model training data corresponding to the reference boiler, and historical flue gas oxygen contents corresponding to each of the model training data; Determining the error weight corresponding to the model training data according to each of the feature data; Performing model training according to each of the model training data, the historical flue gas oxygen content corresponding to each of the model training data, and the error weight corresponding to each of the model training data, so as to determine a flue gas oxygen content prediction model, wherein the flue gas oxygen content prediction model is used for flue gas oxygen content prediction of the target boiler; The error weight is used to achieve data migration from the reference boiler to the target boiler to establish a data relationship between the reference boiler and the target boiler, and the error weight corresponding to the model training data is determined according to each of the feature data, including: Determine replacement data corresponding to the model training data from each of the feature data; the replacement data is data identical to the model training data among the multiple feature data, and if the multiple feature data do not contain data identical to the model training data, the replacement data is data most similar to the model training data among the multiple feature data; Determine first distribution information of the model training data in each of the model training data and second distribution information of the replacement data in each of the feature data; For each of the model training data, determining an error weight corresponding to the model training data according to the first distribution information of the model training data and the second distribution information of the replacement data corresponding to the model training data; The model training is performed according to each of the model training data, the historical flue gas oxygen content corresponding to each of the model training data, and the error weight corresponding to each of the model training data to determine the flue gas oxygen content prediction model, including: A1. Determine the first errors corresponding to the model training data respectively according to the prediction result of substituting the model training data into the model to be predicted and the historical flue gas oxygen content corresponding to the model training data, and determine the second error corresponding to the model training data according to the first error corresponding to the model training data and the error weight; A2, judging whether the number of iterations or the second errors corresponding to each of the model training data meet the preset conditions, if yes, determining the model to be predicted as the flue gas oxygen content prediction model, if no, executing A3; A3. According to the second errors corresponding to the model training data, the model parameters in the model to be predicted are adjusted to determine the adjusted model parameters, and the model parameters in the model to be predicted are replaced with the adjusted model parameters, and A1 is executed.
2. The method according to claim 1, characterized in that The step of determining the error weight corresponding to the model training data according to the first distribution information of the model training data and the second distribution information of the replacement data corresponding to the model training data comprises: When the first distribution information and the second distribution information are distribution probabilities respectively, determining a probability ratio of the distribution probability of the model training data and the distribution probability of the replacement data corresponding to the model training data, and determining the probability ratio as the error weight corresponding to the model training data; or, When the first distribution information and the second distribution information are occurrence frequencies respectively, determine the frequency ratio of the occurrence frequency of the model training data and the occurrence frequency of the replacement data corresponding to the model training data, and determine the frequency ratio as the error weight corresponding to the model training data.
3. The method according to claim 2, characterized in that The determining the first distribution information of the model training data in each of the model training data includes: Obtaining the data volume of each model training data; According to the data amount, selecting a target non-parametric estimation method from at least two non-parametric estimation methods; According to the target non-parametric estimation method, the distribution probability of the model training data in each of the model training data is determined, and the distribution probability is determined as the first distribution information.
4. The method according to claim 2, characterized in that: The determining the first distribution information of the model training data in each of the model training data includes: Obtain at least two nonparametric estimation methods; For each of the non-parametric estimation methods, determining the candidate distribution probability of the model training data in each of the model training data based on the non-parametric estimation method; Fusion of the candidate distribution probabilities to determine the target distribution probability; The target distribution probability is determined as the distribution probability of the model training data in each of the model training data, and the distribution probability is determined as the first distribution information.
5. The method according to claim 2, characterized in that: The determining the second distribution information of the replacement data in each of the feature data includes: Based on a non-parametric estimation method, the distribution probability of the replacement data in each of the feature data is determined, and the distribution probability is determined as the second distribution information.
6. A flue gas oxygen content prediction device, characterized in that: include: An acquisition module, used to acquire at least two characteristic data corresponding to the target boiler, at least two model training data corresponding to the reference boiler, and the historical flue gas oxygen content corresponding to each of the model training data; A weight determination module, used to determine the error weight corresponding to the model training data according to each of the feature data; A training module, used for performing model training according to each of the model training data, the historical flue gas oxygen content corresponding to each of the model training data, and the error weight corresponding to each of the model training data, so as to determine a flue gas oxygen content prediction model, wherein the flue gas oxygen content prediction model is used for predicting the flue gas oxygen content of the target boiler; The error weight is used to realize the data migration of the reference boiler to the target boiler to establish the data relationship between the reference boiler and the target boiler, and the weight determination module is specifically used to: determine the replacement data corresponding to the model training data from each of the feature data; the replacement data is the data that is the same as the model training data in the multiple feature data, if there is no data that is the same as the model training data in the multiple feature data, then the replacement data is the data that is most similar to the model training data in the multiple feature data; determine the first distribution information of the model training data in each of the model training data and the second distribution information of the replacement data in each of the feature data; For each of the model training data, determining an error weight corresponding to the model training data according to the first distribution information of the model training data and the second distribution information of the replacement data corresponding to the model training data; The training module is specifically used to: A1, determine the first errors corresponding to the model training data respectively according to the prediction results of substituting the model training data into the model to be predicted and the historical flue gas oxygen content corresponding to the model training data, and determine the second errors corresponding to the model training data according to the first errors corresponding to the model training data and the error weight; A2, determine whether the number of iterations is met or whether the second errors corresponding to each of the model training data meet the preset conditions, if yes, determine the model to be predicted as the flue gas oxygen content prediction model, if not, execute A3; A3. According to the second errors corresponding to the model training data, the model parameters in the model to be predicted are adjusted to determine the adjusted model parameters, and the model parameters in the model to be predicted are replaced with the adjusted model parameters, and A1 is executed.
7. A computer-readable storage medium, comprising an execution instruction. When a processor of an electronic device executes the execution instruction, the processor executes the method according to any one of claims 1 to 5.
8. An electronic device, comprising a processor and a memory storing execution instructions, wherein when the processor executes the execution instructions stored in the memory, the processor executes the method according to any one of claims 1 to 5.