Acid conversion section working condition matching method, device and system based on combination weight empowerment

Abstract

The application relates to the technical field of copper smelting acid-making conversion, and discloses a method, device and system for matching the working condition of an acid-making conversion section by means of combined weight assignment, the method comprising the following steps: collecting multi-source historical industrial data, processing the data, constructing core feature time sequence data, and constructing a historical working condition sample library through sliding time window processing; respectively adopting a subjective weighting method and an objective weighting method to combine weighting and construct a weighted similarity calculation model; performing the same processing on multi-source real-time industrial data to obtain a current working condition sample; calculating the weighted similarity between the current working condition sample and each historical working condition sample in the historical working condition sample library by means of the weighted similarity calculation model, and outputting a similar working condition result according to the weighted similarity ranking. The application comprehensively characterizes the working condition by fusing multi-source data, combines weighting to take into account process experience and data rules, has high matching accuracy and strong real-time performance, supports interactive visual application, and effectively solves the problems of one-sided characterization, extensive measurement and poor practicability in the prior art.

Description

Technical Field

[0001] This application relates to the field of copper smelting acid conversion technology, specifically to a method, apparatus and system for matching the operating conditions of acid conversion process by combining weights. Background Technology

[0002] The copper smelting flue gas sulfuric acid conversion process is a complex industrial process involving gas-solid catalytic reactions, multi-stage heat exchange, and fluctuations in upstream smelting load. Its operating conditions are determined by dozens of coupled variables. Achieving stable, efficient, and optimized operation of this process has always been a core challenge for the industry. Traditional operations rely on human experience, which is highly subjective and makes it difficult to pass on best practices.

[0003] With the development of the Industrial Internet and data acquisition technologies, utilizing historical data-driven methods to assist or replace manual experience-based judgment has become an important direction for intelligent upgrading. Among these, "operating condition matching" technology has attracted significant attention due to its intuitiveness. Its core idea is to compare the current operating conditions with massive records in a historical database to quickly identify the most similar historical operating conditions, thereby referencing the corresponding operational strategies or results. However, existing operating condition matching technologies have significant shortcomings: First, the representation of operating conditions is limited in scope and lacks completeness. Existing methods typically select only one or a very few key parameters for matching, failing to fully utilize the comprehensive, multi-source data collected by modern industrial control systems, which covers equipment status, raw material input, process variables, and actuator actions. This one-sided representation is like "the blind men and the elephant," unable to fully and accurately depict complex operating conditions, resulting in a weak foundation for the reliability of matching results and making them prone to misleading.

[0004] Secondly, the similarity measurement methods are unscientific, resulting in low matching accuracy. Neither simple equal-weighted Euclidean distance nor traditional methods relying on subjective expert weighting scientifically reflect the true weight differences in the impact of different process variables on operating conditions. This crude or subjective measurement method causes similarity calculations to deviate from the essence of the process, and the retrieved "similar operating conditions" may not have reference value in actual production, making it difficult to guarantee matching accuracy.

[0005] Third, the system architecture is closed and rigid, lacking applicability and flexibility. Existing solutions often solidify feature sets and weights within the system, forming a "black box." Field engineers cannot dynamically adjust the features involved in matching based on real-time changes in production targets or personalized analytical needs, resulting in a disconnect between the system and actual production decision-making processes, poor interactivity, and limited practical value.

[0006] Therefore, there is an urgent need for an intelligent matching method for the working conditions of acid production and conversion processes that can comprehensively characterize the working conditions and scientifically quantify the similarity. Summary of the Invention

[0007] To address the aforementioned issues, this application provides a method, apparatus, and system for matching operating conditions in acid production and conversion processes using combined weighting, aiming to achieve precise matching of operating conditions to support operational decisions.

[0008] The technical solution adopted in this application is as follows: Firstly, this application provides a method for matching operating conditions of a sulfuric acid conversion unit using combined weighting, including: Multi-source historical industrial data of the acid production and conversion section were collected. After data preprocessing, initial screening and configuration based on process mechanism, and redundancy removal of core feature time series data construction and processing, and sliding time window processing, historical operating condition samples were formed and a historical operating condition sample library was constructed. The subjective and objective weights of each core feature were calculated using subjective and objective weighting methods respectively. The subjective and objective weights were then merged into a combined weight, and a weighted similarity calculation model was constructed based on the combined weight. Collect multi-source real-time industrial data from the acid production and conversion section, and process it in the same way as multi-source historical industrial data to form a current operating condition sample; The weighted similarity calculation model is used to calculate the weighted similarity between the current working condition sample and each historical working condition sample in the historical working condition sample library, and the similar working condition results are output according to the weighted similarity ranking.

[0009] Secondly, this application also provides a combined weighted acid conversion section operating condition matching device, comprising: The historical data processing unit is used to collect multi-source historical industrial data of the acid production and conversion section. After data preprocessing, initial screening and configuration in response to process mechanism and redundancy removal of core feature time series data construction and processing, and sliding time window processing, historical operating condition samples are formed and a historical operating condition sample library is built. The weighting unit is used to calculate the subjective weight and objective weight of each core feature using subjective weighting method and objective weighting method respectively, and to merge the subjective weight and objective weight into a combined weight. Based on the combined weight, a weighted similarity calculation model is constructed. The real-time data processing unit is used to collect multi-source real-time industrial data from the acid production and conversion section, and to form a current operating condition sample in the same way as the multi-source historical industrial data. The matching unit is used to calculate the weighted similarity between the current working condition sample and each historical working condition sample in the historical working condition sample library using a weighted similarity calculation model, and output the similar working condition results according to the weighted similarity ranking.

[0010] Thirdly, this application also provides a combined weighted acid conversion section condition matching system, including: a terminal and a server; The terminal is used to transmit the initial screening configuration of the process mechanism; The server-side is used to execute the steps of the above-mentioned combined weighting method for matching the operating conditions of the acid conversion section; The terminal is also used to receive and display results from similar operating conditions.

[0011] Fourthly, this application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-mentioned combined weighting method for matching the operating conditions of the acid production conversion section.

[0012] Fifthly, this application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the above-described combined weighting method for matching the operating conditions of the acid production conversion section.

[0013] The above-mentioned technical solution adopted in this application can achieve the following beneficial effects: For heterogeneous time-series data from multiple sources, such as DCS (Distributed Control System) and PLC (Programmable Logic Controller), with different sampling frequencies, including equipment status, raw material feeding, process variables, and actuator actions, data cleaning and time alignment based on process rules are performed. In response to the process mechanism, initial screening and redundancy removal are performed to select core characteristic time-series data in order to construct a low-dimensional feature vector that comprehensively represents the working conditions.

[0014] Subjective weighting is used to generate subjective weights that reflect expert experience, while objective weighting is used to generate objective weights based on data statistical characteristics. The subjective and objective weights are then merged to form a combined weight, and the distance of this combined weight is used as the core metric for the similarity of working conditions.

[0015] Multi-source data representation, weighted similarity calculation with combined weighting, and operating condition matching are sequentially integrated into a complete intelligent operating condition matching method and process, which is applied to the historical experience retrieval and operational decision support scenarios of the copper smelting acid conversion section.

[0016] The system provides a terminal interface that allows operators to dynamically select and configure features for weighted similarity calculation by checking boxes. The system dynamically calls the matching engine and displays the results based on user configuration. It also provides the function of visualizing the changing trends of key parameters of the current working condition and any matched historical working condition within the same time window. Attached Figure Description

[0017] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1A flowchart illustrating a combined weighting method for matching the operating conditions of an acid conversion unit according to an embodiment of this application is shown. Figure 2 A schematic diagram of a data preprocessing flow according to an embodiment of this application is shown; Figure 3 A schematic diagram illustrating the process of constructing and matching a weighted similarity calculation model according to an embodiment of this application is shown; Figure 4 A weight comparison diagram is shown for a feature weight rationality comparison analysis according to an embodiment of this application; Figure 5 A schematic diagram of a combined weighting acid conversion section operating condition matching device according to an embodiment of this application is shown; Figure 6 A schematic diagram of a combined weighted acid conversion section operating condition matching system according to an embodiment of this application is shown; Figure 7 A terminal page effect diagram according to an embodiment of this application is shown; Figure 8 A schematic diagram of the structure of an electronic device according to an embodiment of this application is shown. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0019] Figure 1 A flowchart illustrating a combined weighting method for matching operating conditions in an acid conversion unit according to an embodiment of this application is shown. Figure 1 As can be seen, the method provided in this embodiment includes steps S110 to S140: Step S110: Collect multi-source historical industrial data of the acid production and conversion section, and after data preprocessing, initial screening and configuration in response to process mechanism and redundancy removal of core feature time series data construction and processing, and sliding time window processing, form historical operating condition samples and build a historical operating condition sample library.

[0020] This step first involves collecting four types of multi-source historical industrial data from DCS and PLC: equipment operating status, raw material feeding information, process parameters and flue gas composition, and control variables and actuator status. This covers comprehensive information including smelting exhaust fans, matte feeding, flue gas concentration, catalyst bed temperature, and regulating valve opening. Of course, in addition to the four types of data mentioned above, multi-source historical industrial data may also include, but is not limited to: equipment vibration signals, equipment acoustic signals, real-time image data of key equipment components, and production technology and quality inspection data from Manufacturing Execution System (MES). In short, any industrial data source that can reflect or influence the operating conditions of the conversion section is within the scope of this application.

[0021] Subsequently, the original multi-source historical industrial data is preprocessed. Data preprocessing may include missing value interpolation and imputation, outlier identification and correction, unified resampling and time alignment, standardization, etc., to eliminate noise, temporal misalignment and dimensional effects.

[0022] Then, based on the initial screening configuration and statistical verification analysis of the process mechanism, redundant features are eliminated to obtain the time series data of the core features.

[0023] Finally, a sliding time window is used to segment the continuous core feature time series data into discrete historical working condition samples, thus completing the construction of the historical working condition sample library and providing a standardized and structured data foundation for subsequent weighted similarity calculation.

[0024] In some optional implementations, step S110 involves collecting multi-source historical industrial data for the acid production and conversion section, including: collecting historical industrial data on equipment operating status, raw material feeding information, process parameters and flue gas composition, and control variables and actuator status; wherein, the historical industrial data on equipment operating status includes: smelting exhaust fan speed, blowing exhaust fan speed, fan outlet pressure, and guide vane opening of the primary negative pressure fan; the historical industrial data on raw material feeding information includes: FCF matte feed rate, F... SF feed rate; historical industrial data on process parameters and flue gas composition, including: fan inlet flue gas temperature, fan inlet flue gas flow rate, fan inlet SO2 concentration, fan inlet O2 concentration, fan outlet flue gas temperature, fan outlet flue gas flow rate, fan outlet SO2 concentration, fan outlet O2 concentration, No. 1 waste heat boiler tube-side temperature, No. 2 waste heat boiler tube-side temperature, tube-side temperature of each stage of heat exchanger, and multi-stage catalyst bed temperature; historical industrial data on control variables and actuator status, including: large air valve opening, small air valve opening, and opening of temperature regulating valves for layers 1 to 5.

[0025] The multi-source historical industrial data collected in this embodiment comes from the DCS and PLC systems of the acid conversion section of copper smelting enterprises. The collection frequency ranges from seconds to minutes, and the data types are divided into the following four categories: (1) Historical industrial data on equipment operating status. This includes: smelting exhaust fan speed, blowing exhaust fan speed, fan outlet pressure, and guide vane opening of the primary negative pressure fan, etc. (2) Historical industrial data on raw material input. This includes: FCF (ferroalloy flux) input to the furnace, FSF (ferroalloy flux) input, etc. (3) Historical industrial data on process parameters and flue gas composition. This includes: flue gas temperature at the fan inlet, flue gas flow rate at the fan inlet, SO2 concentration at the fan inlet, O2 concentration at the fan inlet, flue gas temperature at the fan outlet, flue gas flow rate at the fan outlet, SO2 concentration at the fan outlet, O2 concentration at the fan outlet, tube-side temperature of waste heat boiler No. 1, tube-side temperature of waste heat boiler No. 2, tube-side temperature of each stage of heat exchangers (V heat exchanger, I heat exchanger, etc.), and temperature of multi-stage catalyst bed, etc. (4) Historical industrial data on control variables and actuator status. This includes: opening degree of large air valve K1, opening degree of small air valve K2, opening degree of temperature regulating valves (H4, A1, H5, H2, A5) in layers 1 to 5, etc.

[0026] These raw, multi-source historical industrial data collectively constitute a high-dimensional, heterogeneous, strongly correlated raw dataset containing time-series characteristics. Directly using the raw dataset for subsequent analysis would introduce significant noise; therefore, rigorous data preprocessing is performed on the raw dataset.

[0027] In some optional implementations, step S110, after data preprocessing, responding to the initial screening configuration of process mechanisms and constructing core feature time series data for redundancy removal, and sliding time window processing, forms historical operating condition samples and constructs a historical operating condition sample library, including: performing missing value interpolation and imputation, outlier identification and correction, unified resampling and time alignment, and standardization processing on multi-source historical industrial data; in response to the initial screening configuration of process mechanisms, selecting candidate feature time series data from each feature after data preprocessing, removing collinear redundant features in the candidate feature time series data through correlation analysis to form core feature time series data; using a fixed-length sliding window to segment the core feature time series data, generating discrete historical operating condition samples based on the feature values ​​of the core features in each segment, and constructing a historical operating condition sample library based on the historical operating condition samples.

[0028] Figure 2 A schematic flowchart of data preprocessing according to an embodiment of this application is shown. Figure 2 As shown, the data preprocessing process begins with data cleaning. For long-term missing values ​​caused by obvious sensor malfunctions, linear interpolation across preceding and following time windows is used to fill in the gaps. For extreme outliers caused by transient disturbances, identification and correction are performed using a process knowledge base (such as physical limits for valve opening and reasonable temperature ranges). Alternatively, automatic detection methods such as box plots and isolated forest algorithms can also be used for outliers to identify and correct them.

[0029] After data cleaning, a unified resampling and time alignment are performed. All data are uniformly resampled to a time granularity of 1 minute, and the nearest-neighbor timestamp matching method is used to ensure that data from different sources are strictly synchronized on the timeline. This is the basis for subsequent segmentation according to fixed time windows.

[0030] Finally, considering the differences in physical dimensions, Z-score standardization was applied to all numerical data to achieve a mean of 0 and a standard deviation of 1, thus eliminating the dominant influence of dimensions on subsequent weighted similarity calculations. This resulted in a standardized time-series dataset. Of course, besides Z-score standardization, methods such as Min-Max normalization and decimal scaling can also be used.

[0031] After data preprocessing, the core task is to select a set of core features from the vast amount of data that can refine and comprehensively characterize the operational status of the conversion process. This is not only a dimensionality reduction requirement, but also a crucial step in improving the performance and interpretability of subsequent matching results. In this embodiment, in response to the initial screening configuration based on process mechanism and combined with data statistical characteristics, the core features are systematically screened and constructed.

[0032] First, based on the role and controllability of each feature in the process flow, they are divided into three categories, and their definitions and examples are shown in Table 1.

[0033] Table 1 Feature Classification:

[0034] Input condition variables reflect boundary conditions from upstream processes that cannot be changed in real time within the current work section, and are the root cause of changes in operating conditions. Equipment status variables are actuator parameters that operators can directly control, representing the system's "operation gestures." Process status variables comprehensively reflect the system's internal thermal and reaction conditions under the current input and operation, and are a direct manifestation of the operating conditions.

[0035] Operators configure representative features from various features in the standardized time-series dataset based on the process mechanism, thereby initially screening out candidate feature time-series data. Then, by calculating the Pearson correlation coefficient of the candidate feature time-series data, highly collinear and redundant features are eliminated, ultimately determining the core feature time-series data. For example, the final determined core feature time-series data are: [FCF copper matte feed rate, FSF feed rate, SO2 concentration at the blower outlet, flue gas flow rate at the blower inlet, and temperature at the top of the first catalyst bed]. This core feature time-series dataset combines comprehensiveness, real-time performance, and simplicity.

[0036] Of course, the core feature time series data obtained may differ depending on the initial screening configuration performed by the operator based on the process mechanism. The core feature time series data that can jointly represent the input condition class, equipment state class, and process state class, selected from the features of the standardized time series dataset, all fall within the protection scope of this application.

[0037] Operating conditions are essentially the dynamic behavioral characteristics of a system over a time period. In order to transform continuous time series data into discrete samples that can be processed by matching algorithms, this embodiment can use a sliding time window method to segment the core feature time series data.

[0038] The choice of window length needs to balance two aspects: it should be long enough to encompass the complete dynamic response of the process (such as temperature propagation caused by load changes), and short enough to ensure the relative stability of the "operating condition" within the window. Based on the experience of process experts in the conversion section and actual data analysis, the basic window length T is set to 60 minutes, and the sliding step size is set to 1 minute to achieve dense sampling of the operating condition evolution. The basic window length T can also be a fixed duration such as 30 minutes. Of course, the sliding time window can also adopt a variable-length window triggered by process events (such as changes in feed rate, load switching, etc.), or use an adaptive algorithm to automatically divide the window according to data stability.

[0039] For core feature time-series data, continuous data segments of length T are sequentially extracted starting from the initial time. Each data segment represents a "historical operating condition sample." To form discretized historical operating condition samples, feature values ​​for each core feature are extracted. Feature values ​​are the instantaneous values ​​of each core feature time-series data within the 60-minute window. Instantaneous values ​​can be the final value, mean, median, standard deviation, trend slope, or the dominant frequency domain feature obtained through Fourier transform, or a large number of feature values ​​automatically generated from a time-series feature extraction library and then filtered.

[0040] Through the above steps, massive amounts of raw, multi-source historical industrial data were successfully transformed into a structured set of historical operating condition samples represented by low-dimensional feature vectors, forming a historical operating condition sample library.

[0041] Step S120: The subjective weight and objective weight of each core feature are calculated using the subjective weighting method and the objective weighting method respectively. The subjective weight and objective weight are then merged into a combined weight, and a weighted similarity calculation model is constructed based on the combined weight.

[0042] This step obtains the core feature weights through both subjective and objective weighting. Subjective weights can be, but are not limited to, the analytic hierarchy process (AHP), fuzzy comprehensive evaluation, Delphi method, etc.—that is, any method that can transform the qualitative judgment of domain experts on the importance of features into quantitative weights. Objective weights can be, but are not limited to, the CRITIC method, entropy weight method, standard deviation method, or other mathematical methods that calculate weights based on the dispersion and correlation of the data itself. The subjective and objective weights are then fused using methods such as linear weighting, multiplicative synthesis, or dynamically solving for the optimal fusion coefficient based on the optimization objective to obtain a combined weight. A weighted similarity calculation model is constructed based on the combined weight calculation, ensuring that the weighted similarity calculation aligns with both the process mechanism and data statistical laws, significantly improving the scientific accuracy of the matching. The weighted similarity calculation model can include, but is not limited to, weighted Euclidean distance, weighted Manhattan distance, weighted Minkowski distance, or first incorporating the combined weights into the feature space before calculating the distance. Alternatively, the weighted similarity calculation model can also use the combined weights as an attention mechanism, adjusting the feature differences with weights before inputting them into a simple neural network for weighted similarity scoring.

[0043] In some optional implementations, step S120 involves calculating the subjective and objective weights of each core feature using both subjective and objective weighting methods, fusing the subjective and objective weights into a combined weight, and constructing a weighted similarity calculation model based on the combined weight. This includes: using the Analytic Hierarchy Process (AHP) to obtain subjective weights by constructing a hierarchical structure model, building an expert judgment matrix, calculating feature vectors, and performing consistency checks; using the CRITIC method to calculate information content based on the contrast strength of core features and the conflict between core features, thereby obtaining objective weights; linearly weighting and fusing the subjective and objective weights based on a preset trade-off coefficient to obtain the combined weight, and constructing a weighted similarity calculation model based on the weighted Euclidean distance of the combined weight.

[0044] Figure 3 A schematic diagram illustrating the process of constructing and matching a weighted similarity calculation model according to an embodiment of this application is shown. Combined with... Figure 3 As shown, the Analytic Hierarchy Process (AHP) is a multi-criteria decision-making method that transforms semi-qualitative and semi-quantitative problems into quantitative ones. It is suitable for structuring and quantifying the experience-based judgments of domain experts. This embodiment applies the AHP to determine subjective weights. Specifically: Construct a hierarchical model. The top-level target layer is based on operating condition similarity judgment; the intermediate criterion layers are three feature categories (input condition, equipment status, and process status); and the bottom layer is based on core features (the number of core features is...). ) to build a hierarchical model for the underlying solution layer.

[0045] Construct an expert judgment matrix. Invite k (usually ≥3) process experts (including senior process engineers and operation experts, etc.) to compare the importance of elements in the same layer relative to the layer above using the "1~9 scale method". Combine the judgments of the k process experts and synthesize the judgment matrix A using the geometric mean method.

[0046] Calculate the eigenvectors and perform a consistency check. Find the largest eigenvalue of the judgment matrix A. The subjective weights are obtained by normalizing the corresponding feature vectors and their corresponding feature vectors. Calculate the consistency ratio CR; where CR = CI / RI, CI is the consistency index, and RI is the random consistency index. When CR < 0.1, the consistency of judgment matrix A is considered acceptable, and the obtained subjective weights are valid; otherwise, process experts need to readjust the judgment.

[0047] By using the analytic hierarchy process, difficult-to-quantify process experience is transformed into specific subjective weight values.

[0048] The CRITIC (Criteria Importance Through Intercriteria Correlation) method is an objective weighting method based on data variability and conflict. The principle is that the weight of a feature depends on two factors: the feature's own contrast strength (a larger standard deviation indicates richer information) and its conflict with other features (weaker correlation, stronger independence, and lower information redundancy). This embodiment applies the CRITIC method to determine objective weights. Specifically: Suppose that the historical operating condition sample database contains m samples, and each sample includes The feature values ​​of the standardized core features constitute a data matrix. .

[0049] Calculate the contrast intensity. Use the standard deviation. Indicates the first The core feature (i.e., the first core feature of the data matrix) The magnitude of the fluctuation in the value of the column. The larger the value, the stronger the ability of this core feature to distinguish between samples.

[0050] Calculate conflict: by calculating the first The core feature (i.e., the first core feature of the data matrix) (column) and the first The core feature (i.e., the first core feature of the data matrix) Correlation coefficient of columns To measure, The larger the value, the more significant the [value]. The stronger the independence of a core feature from all other core features, the less easily its information can be replaced.

[0051] Calculated information content: the first Information content of each core feature The product of contrast intensity and conflict is expressed by the following formula (1): ;Formula (1); Calculating objective weights: The information content of each core feature is normalized to obtain the objective weights. It can be expressed by the following formula (2): ;Formula (2); The CRITIC method is based entirely on the statistical characteristics of historical data, objectively uncovering the intrinsic relationships between core features and avoiding potential biases from human subjectivity.

[0052] To balance the guiding role of process experience with the objectivity of data patterns, this embodiment uses a linear weighting method to fuse subjective and objective weights, resulting in a combined weight. It can be expressed by the following formula (3): ;Formula (3); in, This represents the tradeoff coefficient, reflecting the degree of preference between subjective experience and objective data. The matching accuracy of subsequent working conditions is the highest. Indicates subjective weighting. Indicates objective weighting. .

[0053] Furthermore, the weighted similarity calculation model is defined as a monotonically decreasing function of the weighted Euclidean distance, which is mapped to the [0,1] interval for ease of understanding and comparison. That is, the weighted similarity calculation model is constructed using the following formulas (4) and (5): ;Formula (4); ;Formula (5); in, Indicates the weighted Euclidean distance. This represents the current operating condition sample. This represents historical operating condition samples. Indicates the core feature index. Indicates the number of core features. Indicates the first The combined weights of the core features and , This represents the first sample of the current operating condition. The values ​​of the core features, The first historical operating condition sample The values ​​of the core features; This represents the weighted similarity.

[0054] when and When they are exactly the same, ; and The greater the difference, The closer it gets to 0. During real-time matching, the weighted similarity between the current working condition sample and all historical working condition samples in the historical working condition sample library is calculated, and the preset number of historical working condition samples with the highest weighted similarity and their detailed information are returned.

[0055] Step S130: Collect multi-source real-time industrial data from the acid production and conversion section, and form a current operating condition sample using the same processing method as the multi-source historical industrial data.

[0056] This step involves real-time acquisition of multi-source industrial data from the acid production and conversion section, strictly following the processing flow of multi-source historical industrial data. It completes data preprocessing, core feature time-series data construction and processing, and sliding time window processing to obtain current operating condition samples with the same dimension and format as historical operating condition samples. This ensures that the current operating condition samples and historical operating condition samples are completely consistent in feature space and data format, thus ensuring the fairness and accuracy of subsequent weighted similarity calculations.

[0057] Step S140: Calculate the weighted similarity between the current working condition sample and each historical working condition sample in the historical working condition sample library using a weighted similarity calculation model, and output the similar working condition results according to the weighted similarity ranking.

[0058] In some optional implementations, step S140, which calculates the weighted similarity between the current working condition sample and each historical working condition sample in the historical working condition sample library using a weighted similarity calculation model, and outputs the similar working condition results according to the weighted similarity ranking, includes: traversing the historical working condition sample library, calculating the weighted similarity between the current working condition sample and each historical working condition sample respectively; sorting the calculated weighted similarity from high to low, and outputting the top preset number of historical working condition samples and their corresponding working condition information.

[0059] This step utilizes the established weighted similarity calculation model to traverse the historical operating condition sample database, calculating the weighted similarity between the current operating condition sample and each historical operating condition sample. The weighted similarities are then sorted from highest to lowest, and a predetermined number of historical operating condition samples are selected. The output includes the occurrence time, key parameters, and corresponding operational information. Simultaneously, the comparison results of the feature values ​​of each core characteristic between the historical and current operating condition samples are also output. This provides field operators with directly referable historical best operating condition experience, enabling data-driven intelligent decision support.

[0060] The matching effect of the above method will be verified and explained through specific embodiments below.

[0061] The experimental data came from the actual production data of a copper smelter in China throughout 2023. Nine months of data were randomly selected to build a historical operating condition sample library, and the remaining three months of data were used as a test set to simulate the current operating conditions. The experiment was conducted on a server configured with an Intel Xeon CPU and 64GB of memory, using Python 3.10 and MySQL.

[0062] To verify the superiority of the above method, the following comparison method is designed: M1: Equal weighting method. All core features have equal weights and are used as the baseline.

[0063] M2: AHP subjective weighting method. Only expert AHP subjective weights are used.

[0064] M3: CRITIC objective weighting method. Only CRITIC objective weights are used.

[0065] M4: The combined weighting method proposed in this embodiment. .

[0066] The evaluation is conducted using three dimensions: Matching accuracy (Hit Rate@K). A "correct match" is defined as follows: if at least one of the retrieved Top-K historical operating condition samples has a subsequent operation (such as adjusting the direction of a critical valve) that is consistent in trend with the optimal subsequent operation (post-annotated by experts) of the current operating condition sample, then it is counted as a hit. The hit rate is key to evaluating the value of matching in business guidance.

[0067] Reasonableness of feature weights. Through expert review, a qualitative assessment is conducted to determine whether the weight allocations obtained from each method align with the understanding of the process.

[0068] Response time. The average time taken to complete a matching calculation for the entire historical operating condition sample database, used to assess project feasibility.

[0069] I. Comparative Analysis of Matching Accuracy: Table 2 shows the hit rates of different methods in Top-1, Top-3, and Top-5.

[0070] Table 2 Comparison of hit rates (%) of different comparison methods:

[0071] As can be seen, method M4 provided in this embodiment achieved the highest hit rate across all Top-K scenarios. Compared to M1, the Top-1 hit rate is improved by an absolute increase of 11.3%, demonstrating that combined weighting significantly enhances the ability to accurately match optimal historical experience. The performance of M2 and M3 falls between M1 and M4, with M2 slightly outperforming M3, indicating that expert experience has extremely high value, but purely subjective weighting may ignore the objective relationships implicit in the data. Method M4 provided in this embodiment achieves performance improvement by fusing subjective and objective information.

[0072] II. Comparative Analysis of the Reasonableness of Feature Weights: Figure 4 A weight comparison chart is shown for a feature weight rationality comparison analysis according to an embodiment of this application. (Refer to...) Figure 4 As shown, the assumptions in M1 clearly do not align with process understanding. M2 assigns the highest weights to the SO2 concentration at the blower outlet (reaction driving force) and the temperature at the top of the first catalyst bed (critical for reaction activation), consistent with expert experience. M3, on the other hand, emphasizes the objective information of the FCF matte feed rate (load source) and flue gas flow rate (system throughput). The weight allocation of method M4 provided in this embodiment inherits M2's emphasis on core reaction state variables and adopts M3's focus on input load fluctuations, forming a more comprehensive and balanced weighting system, which explains its higher accuracy in principle.

[0073] III. Comparative Analysis of Response Time

[0074] The average response time of method M4 in this embodiment is 820 milliseconds per match, which fully meets the real-time requirement of "minute-level" response for half-hour window conditions in production sites. Although it is slightly slower than M1 (approximately 650 milliseconds) in calculating the weighted distance, this time cost is completely acceptable in exchange for its significant improvement in accuracy. The response time comparison of each method is shown in Table 3.

[0075] Table 3 Comparison of Response Times:

[0076] Note: The real-time threshold is set to 2 seconds based on the decision-making and response requirements of on-site operators.

[0077] This embodiment achieves the following beneficial effects through the above method: 1. Significantly improved matching accuracy and reliability. Compared to existing technologies that use only a single or few features, this application constructs a comprehensive operating condition feature vector by integrating multi-source data such as equipment, raw materials, processes, and controls. This overcomes the shortcomings of one-sided representation and provides a more complete and reliable information foundation for weighted similarity calculation, thereby effectively reducing the risk of "false matching" due to missing information. Furthermore, compared to methods using equal weighting or purely subjective weighting, this application uses a combined weighting model that integrates AHP subjective experience with CRITIC objective data for weighted similarity calculation. This ensures that the weight allocation aligns with process understanding and adheres faithfully to the inherent laws of the data, significantly improving the scientific rigor of similarity measurement and the accuracy of matching results.

[0078] 2. The system's practicality and ease of use are significantly enhanced. Compared to the closed and rigid system architecture of existing technologies, the system provided in this application, which supports operator-defined features and interactive comparisons, allows on-site operators to flexibly adjust matching strategies based on real-time production concerns and verify them through an intuitive trend comparison interface. This feature enables the method proposed in this application to be deeply integrated into the actual production decision-making process, transforming it from a "black box" algorithm tool into a "white box" decision support platform, greatly improving the technology's acceptability and practical value.

[0079] 3. Achieved precise adaptation and intelligent support for complex industrial scenarios. This application integrates multi-source data representation, combined weighted matching, and interactive applications into a complete methodology, systematically solving the problem of condition matching in the specific and complex industrial scenario of copper smelting and acid production conversion. It not only accurately locates similar historical conditions but also transforms "data" into actionable "insights" through visualization, providing operators with unprecedented data-driven decision support. This helps reduce over-reliance on individual experience, promotes the standardization and inheritance of excellent operational experience, and thus improves the overall stability and intelligence level of production operations.

[0080] Figure 5 A schematic diagram of a combined weighting and matching device for acid production conversion section operating conditions, according to an embodiment of this application, is shown. (Refer to...) Figure 5 As shown, the combined weighted acid conversion section condition matching device 500 provided in this embodiment includes: The historical data processing unit 510 is used to collect multi-source historical industrial data of the acid production and conversion section. After data preprocessing, initial screening and configuration in response to process mechanism and redundancy removal of core feature time series data construction and processing, and sliding time window processing, historical operating condition samples are formed and a historical operating condition sample library is constructed. The weighting unit 520 is used to calculate the subjective weight and objective weight of each core feature using subjective weighting method and objective weighting method respectively, and to integrate the subjective weight and objective weight into a combined weight. Based on the combined weight, a weighted similarity calculation model is constructed. The real-time data processing unit 530 is used to collect multi-source real-time industrial data from the acid production and conversion section, and to form a current operating condition sample in the same processing method as the multi-source historical industrial data. The matching unit 540 is used to calculate the weighted similarity between the current working condition sample and each historical working condition sample in the historical working condition sample library using a weighted similarity calculation model, and output the similar working condition results according to the weighted similarity ranking.

[0081] In some optional embodiments, in the above-mentioned apparatus, the historical data processing unit 510 is used to: collect historical industrial data on equipment operating status, historical industrial data on raw material feeding information, historical industrial data on process parameters and flue gas composition, and historical industrial data on control variables and actuator status of the acid conversion section; wherein, the historical industrial data on equipment operating status includes: smelting exhaust fan speed, blowing exhaust fan speed, fan outlet pressure, and guide vane opening of the primary negative pressure fan; the historical industrial data on raw material feeding information includes: FCF matte feed rate, FSF feed rate, etc. Material quantity; historical industrial data on process parameters and flue gas composition include: fan inlet flue gas temperature, fan inlet flue gas flow rate, fan inlet SO2 concentration, fan inlet O2 concentration, fan outlet flue gas temperature, fan outlet flue gas flow rate, fan outlet SO2 concentration, fan outlet O2 concentration, No. 1 waste heat boiler tube side temperature, No. 2 waste heat boiler tube side temperature, tube side temperature of each stage of heat exchanger, and multi-stage catalyst bed temperature; historical industrial data on control variables and actuator status include: large air valve opening degree, small air valve opening degree, and temperature regulating valve opening degrees of layers 1 to 5.

[0082] In some optional embodiments, in the above-described apparatus, the historical data processing unit 510 is used to: perform missing value interpolation and imputation, outlier identification and correction, unified resampling and time alignment, and standardization processing on multi-source historical industrial data; in response to the initial screening configuration of process mechanism, select candidate feature time series data from the features after data preprocessing, remove collinear redundant features in the candidate feature time series data through correlation analysis to form core feature time series data; segment the core feature time series data using a fixed-length sliding window, generate discrete historical operating condition samples based on the feature values ​​of the core features in each segment, and construct a historical operating condition sample library based on the historical operating condition samples.

[0083] In some optional embodiments, in the above apparatus, the weighting unit 520 is used to: employ the AHP (Analytic Hierarchy Process) to obtain subjective weights by constructing a hierarchical structure model, building an expert judgment matrix, calculating feature vectors, and performing consistency checks; employ the CRITIC method to calculate information content based on the contrast intensity of core features and the conflict between core features, thereby obtaining objective weights; perform linear weighted fusion of subjective and objective weights based on a preset trade-off coefficient to obtain combined weights; and construct a weighted similarity calculation model based on the weighted Euclidean distance of the combined weights.

[0084] In some optional embodiments, in the above-mentioned device, the weighting unit 520 is used to: construct a hierarchical structure model with the working condition similarity judgment as the top target layer, the input condition class, equipment state class, and process state class features as intermediate criteria layers, and each core feature as the bottom scheme layer; process experts compare the importance of elements in the same layer with those in the previous layer using the 1-9 scale method, and construct a judgment matrix using the geometric mean method; solve for the maximum eigenvalue and the corresponding eigenvector of the judgment matrix, normalize the eigenvector to obtain the subjective weight, and obtain the subjective weight when the consistency test is passed based on the consistency ratio.

[0085] In some optional embodiments, in the above apparatus, the weighting unit 520 is used to: calculate the standard deviation of each core feature as the contrast intensity; calculate the correlation coefficient of each core feature with other core features, and determine the conflict based on the correlation coefficient; multiply the contrast intensity and conflict of each core feature to obtain the information content; and normalize the information content of each core feature to obtain the objective weight.

[0086] In some alternative embodiments, in the above apparatus, the weighting unit 520 is used to: calculate the combined weights using the following formula: ;in, Indicates the portfolio weight. Indicates the trade-off coefficient. Indicates subjective weighting. Indicates objective weighting. The weighted similarity calculation model is constructed using the following formula: ; ;in, Indicates the weighted Euclidean distance. This represents the current operating condition sample. This represents historical operating condition samples. Indicates the core feature index. Indicates the number of core features. Indicates the first The combined weights of the core features and , This represents the first sample of the current operating condition. Feature values ​​of each core feature The first historical operating condition sample Feature values ​​of each core feature; This represents the weighted similarity.

[0087] In some optional embodiments, in the above apparatus, the matching unit 540 is used to: traverse the historical working condition sample library, calculate the weighted similarity between the current working condition sample and each historical working condition sample respectively; sort the calculated weighted similarity from high to low, and output the top preset number of historical working condition samples and their corresponding working condition information.

[0088] It should be noted that the acid production conversion section operating condition matching device 500 with the above-mentioned combined weighting can realize the aforementioned combined weighting acid production conversion section operating condition matching method one by one, which will not be elaborated further.

[0089] Figure 6 A schematic diagram of a combined weighted acid conversion section condition matching system according to an embodiment of this application is shown. (Refer to...) Figure 6 As shown, the system provided in this embodiment includes: a terminal 610 and a server 620; Terminal 610 is used for transmitting the initial screening configuration of the process mechanism; Server 620 is used to execute the above-mentioned combined weighting method for matching the operating conditions of the acid conversion section; Terminal 610 is also used to receive and display results of similar working conditions.

[0090] To apply the combined weighting method for matching the operating conditions of the acid production and conversion section to actual production decision-making, a combined weighting system for matching the operating conditions of the acid production and conversion section was developed. This system can be implemented using a B / S (browser / server) architecture, a C / S (client / server) architecture, or a cloud-edge collaborative architecture based on microservices.

[0091] Figure 7 A terminal page effect diagram according to an embodiment of this application is shown. (Refer to...) Figure 7 As shown, the terminal page is designed around the core workflow of "configuration-matching-analysis" and mainly includes two functional areas.

[0092] The first area is the configuration area. Operators can dynamically select features for matching via checkboxes, covering key measurement points such as fan status, raw material feed, flue gas parameters, and bed temperature. Redundancy is removed based on the selected feature subset, and the core features are finally displayed. A default 60-minute sliding window is used, and operators can trigger the matching process with one click. Of course, the system can also provide other interactive configuration functions such as configuration file editing, natural language command parsing, or drag-and-drop workflow creation.

[0093] The second area is the main panel. It centrally displays the matching results and analysis views, presenting the three historical work condition cards with the highest matching scores in real time. Each card clearly shows the occurrence time, the shift team, and a snapshot of key status parameters for that historical work condition. When the operator clicks on any historical work condition card, the system automatically activates the trend comparison function, visually displaying the key parameter curves of the current real-time work condition and the selected historical work condition side-by-side on the same time scale. This allows the operator to intuitively judge the similarities and differences between the two during the dynamic evolution process. The visualization is not limited to two-dimensional trend curve comparisons; it can also use parallel coordinate graphs to simultaneously display differences in multiple core features, or use heatmaps to display the difference matrix between multiple historical work conditions and the current work condition across various core feature dimensions.

[0094] In practical applications, the system provides direct support for core operations through a closed-loop function of "custom search - trend comparison". When a specific trend emerges in production, operators can quickly retrieve similar historical operating conditions and, by comparing the shape, amplitude, and rate of change of the curves, achieve a quantitative assessment and directional prediction of the current operating condition. Simultaneously, by accurately locating similar historical states and their corresponding shifts and time information, the system provides experienced operators with efficient navigation for inferring operational logic and accessing detailed historical records, realizing decision support from experience and intuition to data verification.

[0095] The methods, apparatus and systems proposed in this application are not limited to copper smelting acid conversion processes, but can also be applied to other complex industrial processes with multiple characteristics, strong coupling and reliance on historical experience, such as steel smelting, petrochemicals, cement production and other fields, for similar working condition matching and operation optimization scenarios.

[0096] Figure 8 This invention illustrates a schematic diagram of the structure of an electronic device according to an embodiment of the present application. Figure 8As shown, the electronic device includes a processor, memory, and a network interface connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile and / or volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The network interface is used for communication with external devices via a network connection. When executed by the processor, the computer program implements the functions or steps of the combined weighting method for matching the operating conditions of the acid conversion section.

[0097] In one embodiment, the electronic device provided in this application includes a memory and a processor. The memory stores a database and a computer program that can run on the processor. When the processor executes the computer program, it implements the steps of the combined weighting method for matching the operating conditions of the acid production conversion section.

[0098] The above is as stated in this application. Figure 5 The method for performing combined weighting and matching of acid conversion process conditions disclosed in the illustrated embodiment can be applied to a processor or implemented by a processor. During implementation, each step of the above method can be completed by integrated logic circuits in the processor's hardware or by instructions in software form. The processor can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it can 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 gate or transistor logic devices, or discrete hardware components. The steps of the method disclosed in the embodiments of this application can be directly embodied as being executed by a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software modules can reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. This storage medium is located in memory, and the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above method.

[0099] In one embodiment, a computer-readable storage medium is also provided, on which a computer program is stored, wherein when the computer program is executed by a processor, the steps of a method for matching the operating conditions of an acid production conversion section with combined weighting are implemented.

[0100] It should be noted that the functions or steps that the above-mentioned electronic devices or computer-readable storage media can achieve can be referred to the relevant descriptions in the foregoing method embodiments. To avoid repetition, they will not be described one by one here.

[0101] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), Synchlink, DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM), etc.

[0102] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0103] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A method for matching operating conditions in a sulfuric acid conversion process using combined weighting, characterized in that, include: Multi-source historical industrial data of the acid production and conversion section were collected. After data preprocessing, initial screening and configuration based on process mechanism, and redundancy removal of core feature time series data construction and processing, and sliding time window processing, historical operating condition samples were formed and a historical operating condition sample library was constructed. The subjective and objective weights of each core feature were calculated using subjective and objective weighting methods respectively. The subjective and objective weights were then merged into a combined weight, and a weighted similarity calculation model was constructed based on the combined weight. Collect multi-source real-time industrial data from the acid production and conversion section, and process it in the same way as multi-source historical industrial data to form a current operating condition sample; The weighted similarity calculation model is used to calculate the weighted similarity between the current working condition sample and each historical working condition sample in the historical working condition sample library, and the similar working condition results are output according to the weighted similarity ranking.

2. The method for matching operating conditions of acid production and conversion section by combined weighting according to claim 1, characterized in that, The collected historical industrial data from multiple sources for the acid production and conversion section includes: Collect historical industrial data on equipment operating status, raw material feeding information, process parameters and flue gas composition, and control variables and actuator status of the acid production and conversion section. Among them, the historical industrial data of equipment operation status includes: smelting exhaust fan speed, blowing exhaust fan speed, fan outlet pressure, and guide vane opening of the primary negative pressure fan; Historical industrial data on raw material input includes: FCF (copper ash) input to the furnace and FSF (ferrous metal ash) input. Historical industrial data on process parameters and flue gas composition include: inlet flue gas temperature of the fan, inlet flue gas flow rate of the fan, inlet SO2 concentration of the fan, inlet O2 concentration of the fan, outlet flue gas temperature of the fan, outlet flue gas flow rate of the fan, outlet SO2 concentration of the fan, outlet O2 concentration of the fan, tube-side temperature of waste heat boiler No. 1, tube-side temperature of waste heat boiler No. 2, tube-side temperature of each stage of heat exchanger, and temperature of multi-stage catalyst bed. Historical industrial data on control variables and actuator status includes: large air valve opening, small air valve opening, and temperature control valve openings for layers 1 to 5.

3. The method for matching operating conditions of acid production and conversion section by combined weighting according to claim 2, characterized in that, The process involves data preprocessing, initial screening and configuration based on process mechanisms, redundancy removal of core feature time-series data construction, and sliding time window processing to form historical operating condition samples and build a historical operating condition sample library, including: Perform missing value interpolation and imputation, outlier identification and correction, unified resampling and time alignment, and standardization processing on multi-source historical industrial data; In response to the initial screening configuration of the process mechanism, candidate feature time series data are selected from various features after data preprocessing. Collinearity redundancy features in the candidate feature time series data are eliminated through correlation analysis to form core feature time series data. A fixed-length sliding window is used to segment the time series data of core features. Discrete historical operating condition samples are generated based on the feature values ​​of the core features in each segment, and a historical operating condition sample library is constructed based on the historical operating condition samples.

4. The method for matching operating conditions of acid production and conversion section by combined weighting according to claim 1, characterized in that, The subjective and objective weights of each core feature are calculated using subjective and objective weighting methods, respectively. These subjective and objective weights are then merged into a combined weight. A weighted similarity calculation model is constructed based on this combined weight, including: The Analytic Hierarchy Process (AHP) is used to obtain subjective weights by constructing a hierarchical structure model, building an expert judgment matrix, calculating eigenvectors, and performing consistency checks. The CRITIC method is used to calculate the information content based on the contrast strength of core features and the conflict between core features, thereby obtaining objective weights; The subjective and objective weights are linearly weighted and fused based on a preset trade-off coefficient to obtain a combined weight. A weighted similarity calculation model is then constructed based on the weighted Euclidean distance of the combined weight.

5. The method for matching operating conditions of acid production and conversion section by combined weighting according to claim 4, characterized in that, The method employs the Analytic Hierarchy Process (AHP), which involves constructing a hierarchical model, building an expert judgment matrix, calculating eigenvectors, and performing consistency checks to obtain subjective weights, including: A hierarchical model is constructed with working condition similarity judgment as the top target layer, input condition class, equipment status and process status features as intermediate criteria layers, and each core feature as the bottom solution layer. Process experts compared the importance of elements in the same layer to those in the layer above using a 1-9 scale and constructed a judgment matrix using the geometric mean method. Find the largest eigenvalue and the corresponding eigenvector of the judgment matrix, normalize the eigenvector, and obtain the subjective weight when passing the consistency test based on the consistency ratio.

6. The method for matching operating conditions of acid production and conversion section by combined weighting according to claim 4, characterized in that, The CRITIC method is employed to calculate information content based on the contrast strength of core features and the conflict between core features, thereby obtaining objective weights, including: Calculate the standard deviation of each core feature as the contrast intensity; Calculate the correlation coefficient between each core feature and other core features, and determine the conflict based on the correlation coefficient; The information content is obtained by multiplying the contrast intensity and conflict of each core feature; The information content of each core feature is normalized to obtain objective weights.

7. The method for matching operating conditions of acid production and conversion section by combined weighting according to claim 4, characterized in that, The method involves linearly weighting and fusing subjective and objective weights based on a preset tradeoff coefficient to obtain a combined weight, and then constructing a weighted similarity calculation model based on the weighted Euclidean distance of the combined weight, including: The combined weights are calculated using the following formula: ； in, Indicates the portfolio weight. Indicates the trade-off coefficient. Indicates subjective weighting. Indicates objective weighting. ; The weighted similarity calculation model is constructed using the following formula: ； ； in, Indicates the weighted Euclidean distance. This represents the current operating condition sample. This represents historical operating condition samples. Indicates the core feature index. Indicates the number of core features. Indicates the first The combined weights of the core features and , This represents the first sample of the current operating condition. Feature values ​​of each core feature The first historical operating condition sample Feature values ​​of each core feature; This represents the weighted similarity.

8. The method for matching operating conditions of acid production and conversion section by combined weighting according to claim 1, characterized in that, The step involves calculating the weighted similarity between the current working condition sample and each historical working condition sample in the historical working condition sample database using a weighted similarity calculation model, and outputting the similar working condition results according to the weighted similarity ranking, including: Traverse the historical working condition sample database and calculate the weighted similarity between the current working condition sample and each historical working condition sample; The calculated weighted similarity scores are sorted from high to low, and the top preset number of historical working condition samples and their corresponding working condition information are output.

9. A combined weighted system for matching operating conditions in a sulfuric acid conversion process, characterized in that, include: The historical data processing unit is used to collect multi-source historical industrial data of the acid production and conversion section. After data preprocessing, initial screening and configuration in response to process mechanism and redundancy removal of core feature time series data construction and processing, and sliding time window processing, historical operating condition samples are formed and a historical operating condition sample library is built. The weighting unit is used to calculate the subjective weight and objective weight of each core feature using subjective weighting method and objective weighting method respectively, and to merge the subjective weight and objective weight into a combined weight. Based on the combined weight, a weighted similarity calculation model is constructed. The real-time data processing unit is used to collect multi-source real-time industrial data from the acid production and conversion section, and to form a current operating condition sample in the same way as the multi-source historical industrial data. The matching unit is used to calculate the weighted similarity between the current working condition sample and each historical working condition sample in the historical working condition sample library using a weighted similarity calculation model, and output the similar working condition results according to the weighted similarity ranking.

10. A combined weighted system for matching operating conditions in a sulfuric acid conversion process, characterized in that, include: Terminal and server; The terminal is used to transmit the initial screening configuration of the process mechanism; The server is used to execute the steps of the combined weighting method for matching the operating conditions of the acid conversion section as described in any one of claims 1 to 8; The terminal is also used to receive and display results from similar operating conditions.

Images