Oxygen lance service life prediction method, system and device and readable storage medium

Abstract

The invention provides an oxygen lance service life prediction method, system and device and a readable storage medium, and belongs to the technical field of metallurgical industry automation. The method comprises the following steps: acquiring operation parameters of an oxygen lance, and generating corresponding sequence data according to the types of the operation parameters and acquisition time; performing preprocessing and data cleaning on the sequence data to generate standard sequence data; performing feature engineering on the cleaned sequence data according to domain knowledge, and converting the sequence data into sequence features; deploying a pre-trained sequence large model in the prediction scene, or deploying a sequence large model finely adjusted by using oxygen lance state historical sequence data; on the basis of the feature sequence, according to a selected prediction mode, utilizing a deployed sequence large model to perform prediction, and generating a prediction result feature sequence; and converting the prediction result feature sequence, restoring the prediction result feature sequence into a prediction value, and determining the remaining service life of the oxygen lance according to the prediction value. According to the invention, efficient and accurate prediction of the service life of the oxygen lance is realized.

Description

Technical Field

[0001] The present invention relates to the field of metallurgical industry automation technology, and more particularly to an oxygen lance service life prediction method, system, device and readable storage medium. Background Art

[0002] In the field of modern steel smelting and manufacturing, the oxygen lance is a core component of the steelmaking furnace, and its performance is directly related to production efficiency and product quality. The oxygen lance blows high-pressure and high-purity oxygen into the metal pool in the converter at high speed to assist combustion and accelerate the smelting process. However, the oxygen lance works under high temperature and high pressure for a long time, and its service life has become a key issue.

[0003] At present, oxygen lance life prediction technologies are mainly divided into two categories. The first category is based on finite element analysis and fatigue life simulation analysis methods. This method establishes a three-dimensional geometric model of the oxygen lance, inputs material properties, sets boundary conditions, and performs meshing and solving to evaluate the stress-strain state, temperature distribution and fatigue life of the oxygen lance under different working conditions. However, finite element analysis is essentially a simulation technology, and it is difficult to predict the remaining life of the oxygen lance in real time in a production environment, which limits its flexibility in practical applications.

[0004] The second method is to use data mining and machine learning technology to build algorithm modeling, and establish a prediction model to predict the service life of new oxygen guns by learning data such as past oxygen gun usage, maintenance records, and environmental conditions. This method has the advantage of real-time prediction, but the premise is that a large amount of high-quality historical data needs to be collected. However, in actual operations, many companies have not accumulated enough data, or the data quality is not high, resulting in insufficient model training and limited accuracy of prediction results. Summary of the invention

[0005] In view of the above problems, the purpose of the present invention is to provide a method, system, device and readable storage medium for predicting the service life of an oxygen lance. By comprehensively using advanced technical means such as data acquisition, preprocessing, feature engineering, large model prediction and reasoning, prediction result post-processing, and result interpretation and application, efficient and accurate prediction of the service life of the oxygen lance is achieved, providing strong decision-making support for enterprises.

[0006] In order to achieve the above object, the present invention is implemented through the following technical solutions: In a first aspect, the present invention discloses a method for predicting the service life of an oxygen lance, comprising: Collect the operating parameters of the oxygen gun and generate corresponding sequence data according to the type of operating parameters and the collection time; Preprocess and clean the sequence data to generate target sequence data; Perform feature engineering on the target sequence data based on domain knowledge information to convert the target sequence data into a feature sequence; Deploy the pre-trained sequence model in the current prediction scenario, and fine-tune the sequence model according to the oxygen lance status data in the current prediction scenario; Based on the feature sequence, the large sequence model is used to make predictions according to the selected prediction mode to generate a feature sequence of prediction results; The characteristic sequence of the prediction results is converted and restored to the prediction value, and the remaining service life of the oxygen lance is determined according to the prediction value.

[0007] Furthermore, the operating parameters of the oxygen lance include but are not limited to operating temperature, operating pressure, oxygen flow rate, oxygen purity, input voltage, and current.

[0008] Further, the preprocessing and data cleaning of the sequence data to generate target sequence data includes: Convert sequence data into a unified format; Check and delete outliers and duplicate values ​​in sequence data; Interpolation algorithms are used to fill in missing data points in the sequence data; Dimensionally align the sequence data to generate target sequence data.

[0009] Further, the feature engineering is performed on the target sequence data according to the domain knowledge information to convert the target sequence data into a feature sequence, including: Based on the domain knowledge information related to the oxygen lance, features related to the life of the oxygen lance are derived from the sequence data; Based on the target sequence data, the corresponding statistical features are calculated as the model input features; The statistical characteristics include but are not limited to: cycle mean value, cycle coefficient of variation, and cycle range.

[0010] Furthermore, the pre-trained sequence large model is deployed in the current prediction scenario, and the sequence large model is fine-tuned according to the oxygen lance state data in the current prediction scenario, including: In the current prediction scenario, the time series model and spatiotemporal data model based on the Transformer structure are used as the sequence model for predicting the service life of the oxygen lance; Obtain the status data of the oxygen lance throughout its life cycle in the current prediction scenario and determine the data volume M; Determine whether the data volume M is greater than the adjustment judgment value N; if so, use the state data of the entire life cycle of the oxygen gun to generate the input sequence and output sequence training sequence model to fine-tune the sequence model; if not, do not fine-tune the sequence model.

[0011] Further, the selected prediction mode includes a zero-sample prediction mode and a few-sample prediction mode; When the zero-sample prediction mode is selected, the large sequence model is directly used for prediction based on the feature sequence to generate a feature sequence of the prediction results; When the few-sample prediction mode is selected, A real oxygen lance state feature sequences are input into the large sequence model as part (or all) of the prompt words. When the large sequence model performs prediction extrapolation on the sequence to be predicted, the oxygen lance state feature sequence in the prompt words is synchronously entered into the inference operation and affects the final prediction result.

[0012] Further, the prediction result feature sequence is converted to a prediction value, and the remaining service life of the oxygen lance is determined according to the prediction value, including: According to the effective features, the prediction result feature sequence is converted and restored to the original data format to generate a prediction value; The predicted value is compared with the corresponding state threshold value when the design service life of the oxygen lance expires; Determine the remaining service life of the oxygen lance based on the comparison results.

[0013] In a second aspect, the present invention further discloses an oxygen lance service life prediction system, comprising: The data acquisition module is used to collect the operating parameters of the oxygen lance and generate corresponding sequence data according to the type of operating parameters and the acquisition time; A data preprocessing module is used to preprocess and clean the sequence data to generate target sequence data; The feature engineering module is used to perform feature engineering on the target sequence data according to the domain knowledge information and convert the target sequence data into a feature sequence; The model deployment module is used to deploy the pre-trained sequence model in the current prediction scenario and fine-tune the sequence model according to the oxygen lance status data in the current prediction scenario; The prediction and reasoning module is used to make predictions based on the feature sequence and the selected prediction mode using the sequence big model to generate a prediction result feature sequence; The prediction result processing module is used to convert the prediction result feature sequence, restore it to the prediction value, and determine the remaining service life of the oxygen gun according to the prediction value.

[0014] In a third aspect, the present invention further discloses an oxygen lance service life prediction device, comprising: A memory, used for storing an oxygen lance service life prediction program; A processor is used to implement the steps of the oxygen lance service life prediction method as described in any one of the above items when executing the oxygen lance service life prediction program.

[0015] In a fourth aspect, the present invention further discloses a readable storage medium, on which is stored an oxygen lance service life prediction program, which, when executed by a processor, implements the steps of the oxygen lance service life prediction method as described in any one of the above items.

[0016] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention collects the operating parameters of the oxygen lance and generates sequence data, and then performs cleaning, dimension alignment and other steps to ensure the quality and consistency of the data. This data standardization process not only improves the quality of the data, but also provides a reliable basis for subsequent model training and prediction. Through effective data preprocessing, the present invention can still achieve effective oxygen lance life prediction when the data volume is small or the data quality is poor.

[0017] 2. The present invention uses domain knowledge information to extract feature data from sequence data, and performs dimensionality reduction processing through principal component analysis to generate feature sequences. This feature extraction method not only reduces data redundancy, but also improves the expressiveness of features. At the same time, the present invention deploys a pre-trained large sequence model and fine-tunes the model according to the current prediction scenario. The strong adaptability of this model enables the technical solution to maintain a high prediction accuracy in different prediction scenarios, while reducing the modeling workload and time cost.

[0018] 3. The present invention provides two prediction modes: zero-sample prediction mode and few-sample prediction mode, to adapt to different data volumes and prediction requirements. When the amount of data is sufficient, the zero-sample prediction mode can be selected to directly use the pre-trained large model for prediction and quickly generate prediction results. When the amount of data is small, the few-sample prediction mode can be selected to train the model using some feature sequences, and then predict based on the remaining feature sequences. This diversified prediction mode not only improves the flexibility of prediction, but also maintains a high prediction accuracy when the amount of data is limited.

[0019] 4. The application of the present invention provides an intelligent solution for the maintenance and management of oxygen lances. By converting and comparing the prediction result feature sequence, the remaining service life of the oxygen lance can be determined, providing a scientific basis for the production plan and maintenance plan of the enterprise. This intelligent management method not only reduces the maintenance cost of the enterprise, but also improves production efficiency and extends the service life of the oxygen lance. In addition, the present invention can also monitor the operating status of the oxygen lance in real time, discover potential faults in time, avoid production interruptions caused by oxygen lance failures, and can bring significant economic and social benefits to the enterprise.

[0020] It can be seen that compared with the prior art, the present invention has outstanding substantive features and significant progress, and the beneficial effects of its implementation are also obvious. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, 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 embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0022] Figure 1 It is a method flow chart of a specific implementation mode of the present invention.

[0023] Figure 2 It is a system structure diagram of a specific implementation mode of the present invention.

[0024] In the figure, 1. Data acquisition module; 2. Data preprocessing module; 3. Feature engineering module; 4. Model deployment module; 5. Prediction and reasoning module; 6. Prediction result processing module. DETAILED DESCRIPTION

[0025] In order to enable those skilled in the art to better understand the scheme of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific implementation methods. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than 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 work are within the scope of protection of the present invention.

[0026] See also Figure 1 As shown, this embodiment provides a method for predicting the service life of an oxygen lance, comprising the following steps: S1: Collect the operating parameters of the oxygen gun and generate corresponding sequence data according to the type of the operating parameters and the collection time.

[0027] In a specific embodiment, the operating parameters of the oxygen lance are collected from the oxygen lance and its related systems, and the operating parameters include but are not limited to operating temperature, operating pressure, oxygen flow rate, oxygen purity, input voltage, current, etc., and the information of each maintenance and replacement is recorded. The collection can be performed at a time interval of minutes. After the collection is completed, the corresponding sequence data is generated according to the type of operating parameters and the collection time.

[0028] S2: Preprocess and clean the sequence data to generate target sequence data.

[0029] In a specific embodiment, basic cleaning and integration of sequence data is performed, including: Convert sequence data into a unified format; Check and remove outliers, duplicates, and missing values ​​in sequence data; Interpolation algorithms are used to fill in missing data points in the sequence data; Perform dimensionally aligned sequence data to generate normalized sequence data.

[0030] Through the above processing, the consistency and availability of data can be ensured.

[0031] S3: Perform feature engineering on the target sequence data based on domain knowledge information and convert the target sequence data into a feature sequence.

[0032] In a specific implementation, first, feature data is extracted from sequence data based on domain knowledge information.

[0033] Then, based on the characteristic data, the corresponding cycle average value or coefficient of variation is calculated, and the original sequence data is converted into more descriptive characteristic data, such as cycle average value, coefficient of variation, etc., as effective features to better reflect the operating status and life trend of the oxygen lance.

[0034] S4: Deploy the pre-trained sequence model in the current prediction scenario, and fine-tune the sequence model according to the oxygen gun status data in the current prediction scenario.

[0035] Specifically, a pre-trained large sequence model is deployed in the current prediction scenario, or a large sequence model fine-tuned using historical sequence data of oxygen gun states is deployed.

[0036] In a specific implementation, firstly, in the current prediction scenario, a Transformer-structured time series model (LM4TS) and a spatiotemporal data model (LM4STD) are used as sequence models for predicting the service life of oxygen lances.

[0037] After the model is deployed, the status data of the oxygen lance in the current prediction scenario for the entire life cycle is obtained, and the data volume M is determined. At this time, it is determined whether the data volume M is greater than the judgment value N; if so, the input sequence (independent variable x) and output sequence (prediction value y) are used to train the sequence model to fine-tune the sequence model; if not, the sequence model is not fine-tuned. N can be set to 10.

[0038] S5: Based on the feature sequence, the sequence big model is used to make predictions according to the selected prediction mode to generate a feature sequence of prediction results.

[0039] In a specific implementation, the selected prediction mode includes a zero-sample prediction mode and a few-sample prediction mode; 1. When the zero-sample prediction mode is selected, the large sequence model is directly used for prediction based on the feature sequence to generate a feature sequence of the prediction result. That is, when performing zero-sample prediction, the pre-trained large sequence model is directly used for prediction without the need for additional data or training process.

[0040] 2. When the few-sample prediction mode is selected, a real oxygen lance state feature sequence is input into the large sequence model as part (or all) of the prompt word. When the large sequence model predicts and extrapolates the sequence to be predicted, the oxygen lance state feature sequence in the prompt word is synchronously entered into the inference operation and affects the final prediction result. It can be seen that when performing a few-sample prediction, a small amount of actual sequence data is used as examples to let the large sequence model learn the patterns of these examples, and then predict other sequence data.

[0041] S6: Convert the prediction result feature sequence to restore it to the prediction value, and determine the remaining service life of the oxygen lance according to the prediction value.

[0042] In a specific implementation, first, the prediction result feature sequence is converted according to the effective features, restored to the original data format, and a prediction value is generated. Then the prediction value is compared with the corresponding state threshold when the designed service life of the oxygen lance expires; if the prediction value is lower than the state threshold, it indicates that the oxygen lance is about to reach the end of its service life and needs to be maintained or replaced as soon as possible. If the prediction value is higher than the state threshold, the remaining service life of the oxygen lance can be estimated based on the difference between the prediction value and the state threshold. Finally, the remaining service life of the oxygen lance is determined based on the comparison result.

[0043] It can be seen that in this step, the predicted feature sequence is first converted and restored to the original data format, and then the predicted value is compared with the state threshold when the designed service life of the oxygen lance expires, so as to obtain the remaining service life of the oxygen lance.

[0044] In a specific implementation, the method further includes: after the remaining service life of the oxygen lance is determined, the prediction results can be displayed through a visual interface to help technicians understand the prediction logic and formulate a reasonable maintenance plan accordingly.

[0045] See also Figure 2 As shown, the present invention also discloses an oxygen lance service life prediction system, including: a data acquisition module 1, a data preprocessing module 2, a feature engineering module 3, a model deployment module 4, a prediction reasoning module 5 and a prediction result processing module 6. The data acquisition module 1 is used to acquire the operating parameters of the oxygen lance and generate corresponding sequence data according to the type of the operating parameters and the acquisition time.

[0046] The data preprocessing module 2 is used to preprocess and clean the sequence data to generate target sequence data.

[0047] The feature engineering module 3 is used to perform feature engineering on the target sequence data according to the domain knowledge information, and convert the target sequence data into a feature sequence.

[0048] The model deployment module 4 is used to deploy the pre-trained sequence large model in the current prediction scenario and fine-tune the sequence large model according to the oxygen gun state data in the current prediction scenario.

[0049] The prediction and reasoning module 5 is used to make predictions based on the feature sequence and according to the selected prediction mode using the sequence big model to generate a prediction result feature sequence.

[0050] The prediction result processing module 6 is used to convert the prediction result feature sequence, restore it to the prediction value, and determine the remaining service life of the oxygen lance according to the prediction value.

[0051] The specific implementation of the oxygen lance service life prediction system of this embodiment is basically the same as the specific implementation of the above-mentioned oxygen lance service life prediction method, which will not be repeated here.

[0052] The present invention also discloses an oxygen lance service life prediction device, comprising a processor and a memory; wherein the processor implements the steps of the oxygen lance service life prediction method as described in any one of the above items when executing the oxygen lance service life prediction program stored in the memory.

[0053] Furthermore, the oxygen lance service life prediction device in this embodiment may also include: The input interface is used to obtain the oxygen lance service life prediction program imported from the outside, and save the obtained oxygen lance service life prediction program to the memory, and can also be used to obtain various instructions and parameters transmitted by the external terminal device, and transmit them to the processor, so that the processor can use the above various instructions and parameters to carry out corresponding processing. In this embodiment, the input interface can specifically include but is not limited to a USB interface, a serial interface, a voice input interface, a fingerprint input interface, a hard disk reading interface, etc.

[0054] The output interface is used to output various data generated by the processor to the terminal device connected thereto, so that other terminal devices connected to the output interface can obtain various data generated by the processor. In this embodiment, the output interface may specifically include but is not limited to a USB interface, a serial interface, etc.

[0055] The communication unit is used to establish a remote communication connection between the oxygen lance service life prediction device and an external server so that the oxygen lance service life prediction device can mount the image file to the external server. In this embodiment, the communication unit may specifically include but is not limited to a remote communication unit based on wireless communication technology or wired communication technology.

[0056] The keyboard is used to obtain various parameter data or instructions input by the user by tapping the keycaps in real time.

[0057] The display is used to display the relevant information of the oxygen gun service life prediction process in real time.

[0058] The mouse can be used to assist users in inputting data and simplify user operations.

[0059] The present invention also discloses a readable storage medium, wherein the readable storage medium includes a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable hard disk, a CD-ROM or any other form of storage medium known in the technical field. The readable storage medium stores an oxygen gun service life prediction program, and when the oxygen gun service life prediction program is executed by a processor, the steps of the oxygen gun service life prediction method described in any one of the above items are implemented.

[0060] In summary, the present invention realizes efficient and accurate prediction of the service life of oxygen lances by comprehensively using advanced technical means such as data collection, preprocessing, feature engineering, large model prediction and reasoning, prediction result post-processing, and result interpretation and application, providing powerful decision-making support for enterprises.

[0061] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. As for the method disclosed in the embodiment, since it corresponds to the system disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description.

[0062] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0063] In the several embodiments provided by the present invention, it should be understood that the disclosed systems, systems and methods can be implemented in other ways. For example, the system embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of systems or units, which can be electrical, mechanical or other forms.

[0064] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0065] In addition, each functional module in each embodiment of the present invention may be integrated into one processing unit, or each module may exist physically separately, or two or more modules may be integrated into one unit.

[0066] Similarly, each processing unit in each embodiment of the present invention may be integrated into one functional module, or each processing unit may exist physically, or two or more processing units may be integrated into one functional module.

[0067] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be implemented directly using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

[0068] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article 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, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0069] The above is a detailed introduction to the oxygen lance service life prediction method, system, device and readable storage medium provided by the present invention. This article uses specific examples to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. It should be pointed out that for ordinary technicians in this technical field, without departing from the principles of the present invention, the present invention can also be improved and modified in a number of ways, and these improvements and modifications also fall within the scope of protection of the present invention.

Claims

1. A method for predicting the service life of an oxygen lance, characterized in that: include: Collect the operating parameters of the oxygen gun and generate corresponding sequence data according to the type of operating parameters and the collection time; Preprocess and clean the sequence data to generate target sequence data; Perform feature engineering on the target sequence data based on domain knowledge information to convert the target sequence data into a feature sequence; Deploy the pre-trained sequence model in the current prediction scenario, and fine-tune the sequence model according to the oxygen lance status data in the current prediction scenario; Based on the feature sequence, the large sequence model is used to make predictions according to the selected prediction mode to generate a feature sequence of prediction results; The characteristic sequence of the prediction results is converted and restored to the prediction value, and the remaining service life of the oxygen lance is determined according to the prediction value.

2. The oxygen lance service life prediction method according to claim 1, characterized in that: The operating parameters of the oxygen lance include, but are not limited to, operating temperature, operating pressure, oxygen flow rate, oxygen purity, input voltage, and current.

3. The oxygen lance service life prediction method according to claim 1, characterized in that: The preprocessing and data cleaning of the sequence data to generate target sequence data includes: Convert sequence data into a unified format; Check and delete outliers and duplicate values ​​in sequence data; Interpolation algorithms are used to fill in missing data points in the sequence data; Dimensionally align the sequence data to generate target sequence data.

4. The oxygen lance service life prediction method according to claim 3, characterized in that: The feature engineering of the target sequence data according to the domain knowledge information to convert the target sequence data into a feature sequence includes: Based on the domain knowledge information related to the oxygen lance, features related to the life of the oxygen lance are derived from the sequence data; Based on the target sequence data, the corresponding statistical features are calculated as the model input features; The statistical characteristics include but are not limited to: cycle mean value, cycle coefficient of variation, and cycle range.

5. The oxygen lance service life prediction method according to claim 4, characterized in that: The pre-trained sequence large model is deployed in the current prediction scenario, and the sequence large model is fine-tuned according to the oxygen lance state data in the current prediction scenario, including: In the current prediction scenario, the time series model and spatiotemporal data model based on the Transformer structure are used as the sequence model for predicting the service life of the oxygen lance; Obtain the status data of the oxygen lance throughout its life cycle in the current prediction scenario and determine the data volume M; Determine whether the data volume M is greater than the adjustment judgment value N; if so, use the state data of the entire life cycle of the oxygen gun to generate the input sequence and output sequence training sequence model to fine-tune the sequence model; if not, do not fine-tune the sequence model.

6. The oxygen lance service life prediction method according to claim 5, characterized in that: The selected prediction mode includes a zero-sample prediction mode and a few-sample prediction mode; When the zero-sample prediction mode is selected, the large sequence model is directly used for prediction based on the feature sequence to generate a feature sequence of the prediction results; When the few-sample prediction mode is selected, a preset number of feature sequences of real oxygen lance states are input into the large sequence model as prompt words. When the large sequence model performs prediction extrapolation on the sequence to be predicted, the feature sequence of the oxygen lance state in the prompt word is synchronously added to the inference operation to generate the final prediction result.

7. The oxygen lance service life prediction method according to claim 6, characterized in that: The step of converting the prediction result feature sequence to restore it to a prediction value, and determining the remaining service life of the oxygen lance according to the prediction value includes: According to the effective features, the prediction result feature sequence is converted and restored to the original data format to generate a prediction value; The predicted value is compared with the corresponding state threshold value when the design service life of the oxygen lance expires; Determine the remaining service life of the oxygen lance based on the comparison results.

8. An oxygen lance service life prediction system, characterized in that: The system adopts the oxygen lance service life prediction method according to any one of claims 1 to 7, and the system comprises: The data acquisition module is used to collect the operating parameters of the oxygen lance and generate corresponding sequence data according to the type of operating parameters and the acquisition time; A data preprocessing module is used to preprocess and clean the sequence data to generate target sequence data; The feature engineering module is used to perform feature engineering on the target sequence data according to the domain knowledge information and convert the target sequence data into a feature sequence; The model deployment module is used to deploy the pre-trained sequence model in the current prediction scenario and fine-tune the sequence model according to the oxygen lance status data in the current prediction scenario; The prediction and reasoning module is used to make predictions based on the feature sequence and the selected prediction mode using the sequence big model to generate a prediction result feature sequence; The prediction result processing module is used to convert the prediction result feature sequence, restore it to the prediction value, and determine the remaining service life of the oxygen gun according to the prediction value.

9. An oxygen lance service life prediction device, characterized in that: include: A memory, used for storing an oxygen lance service life prediction program; A processor is used to implement the steps of the oxygen lance service life prediction method according to any one of claims 1 to 7 when executing the oxygen lance service life prediction program.

10. A readable storage medium, characterized in that: The readable storage medium stores an oxygen lance service life prediction program, and when the oxygen lance service life prediction program is executed by the processor, the steps of the oxygen lance service life prediction method according to any one of claims 1 to 7 are implemented.

Images