Intelligent measurement method and system for online component analyzer

Abstract

The invention discloses an intelligent measurement method and system for an online component analyzer, and relates to the technical field of intelligent measurement, and the method comprises the steps: extracting scene features and environment features, setting the number of initial points, and carrying out the random distribution of the initial points; performing clustering analysis on the initial points, establishing a clustering data set, taking a clustering center as an updating measurement point, and establishing a corresponding position weight; real-time environment data collection of the updated measurement points is carried out, correction compensation is carried out through an environment compensation network, and a compensated updated measurement point data set is established; and through the position weight, updating the measurement point data set for measurement analysis, and establishing a measurement analysis result. The technical problem that in the prior art, due to unreasonable measurement point distribution and environmental interference in the measurement process, the precision and stability of a measurement result are low is solved, and the technical effect that the measurement precision and stability of the online component analyzer are improved through measurement point distribution optimization and environmental influence compensation is achieved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent measurement technology, and in particular to an intelligent measurement method and system for an online component analyzer. Background Art

[0002] Online component analyzers, also known as process analysis instruments, are mainly used to be installed in industrial production processes or other source fluid sites to automatically and continuously measure the components or physical parameters of the measured medium. They are widely used in industrial production, environmental monitoring, quality control and other fields, especially in natural gas, petroleum, chemical, electric power and other industries, and play an important role in ensuring product quality, improving production efficiency and protecting the environment.

[0003] However, due to the diversity of measurement environments, the distribution of measurement points is unreasonable, which cannot fully cover the measurement scenarios. In addition, the measurement process may be interfered by various factors, such as temperature, pressure, flow, etc., which affects the measurement accuracy and stability of the online component analyzer. Summary of the invention

[0004] The present application provides an intelligent measurement method and system for an online component analyzer, which is used to solve the technical problem in the prior art of low accuracy and stability of measurement results due to unreasonable distribution of measurement points and environmental interference during the measurement process.

[0005] In a first aspect of the present application, there is provided an intelligent measurement method for an online component analyzer, the method comprising: performing pre-measurement detection to construct a pre-data set; extracting scene features and environmental features from the pre-data set, setting the number of initial points based on the scene features and environmental features, and performing random distribution of the initial points under the number of initial points; performing cluster analysis on the initial points, and performing position iteration, updating the position and number of the initial points according to the position iteration result and the clustering result, and establishing a cluster data set, wherein each cluster data in the cluster data set maps a cluster center, and the cluster center is used as an updated measurement point, and a position weight corresponding to the cluster center is established through the cluster data set; performing environmental impact compensation fitting based on an environmental impact database, and establishing an environmental compensation network; when performing intelligent measurement processing by updating the measurement point, synchronously performing real-time environmental data collection of the updated measurement point, performing correction compensation based on the real-time environmental data collection result and the environmental compensation network, and establishing a compensated updated measurement point data set; performing measurement analysis of the target measurement scene through the position weight and the updated measurement point data set, and establishing a measurement analysis result.

[0006] The second aspect of the present application provides an intelligent measurement system for an online component analyzer, the system comprising: a pre-measurement detection module, the pre-measurement detection module is used to perform pre-measurement detection and construct a pre-data set; an initial point setting module, the initial point setting module is used to extract scene features and environmental features from the pre-data set, set the number of initial points based on the scene features and environmental features, and perform random distribution of the initial points under the number of initial points; a position weight establishment module, the position weight establishment module is used to perform cluster analysis on the initial points, and perform position iteration, update the position and number of initial points according to the position iteration result and the clustering result, and establish a cluster data set, in which each cluster data in the cluster data set maps a cluster center, and the clusters are clustered. The center is used as the updated measurement point, and the position weight corresponding to the cluster center is established through the cluster data set; an environmental compensation network establishment module, the environmental compensation network establishment module is used to perform environmental impact compensation fitting based on the environmental impact database and establish an environmental compensation network; an updated measurement point data set establishment module, the updated measurement point data set establishment module is used to synchronously perform real-time environmental data collection of the updated measurement point when performing intelligent measurement processing through the updated measurement point, perform correction compensation based on the real-time environmental data collection result and the environmental compensation network, and establish a compensated updated measurement point data set; a measurement analysis module, the measurement analysis module is used to perform measurement analysis of the target measurement scene through the position weight and the updated measurement point data set, and establish a measurement analysis result.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages: The present application provides an intelligent measurement method for an online component analyzer, which relates to the field of intelligent measurement technology. The method extracts scene features and environmental features to set the number of initial points, randomly distributes the initial points, performs cluster analysis on the initial points, establishes updated measurement points and corresponding position weights, collects real-time environmental data, performs correction and compensation of the measurement point data set through an environmental compensation network, and establishes measurement analysis results in combination with the position weights. The method solves the technical problem of low accuracy and stability of measurement results due to unreasonable distribution of measurement points and environmental interference during the measurement process in the prior art, and achieves the technical effect of improving the measurement accuracy and stability of the online component analyzer by optimizing the distribution of measurement points and compensating for environmental impacts. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0009] Figure 1 A schematic flow chart of an intelligent measurement method for an online component analyzer provided in an embodiment of the present application; Figure 2 A schematic diagram of a flow chart of establishing an environmental compensation network in an intelligent measurement method for an online component analyzer provided in an embodiment of the present application; Figure 3 A schematic diagram of the structure of an intelligent measurement system for an online component analyzer provided in an embodiment of the present application.

[0010] Description of reference numerals: pre-measurement detection module 11 , initial point setting module 12 , position weight establishing module 13 , environment compensation network establishing module 14 , updated measurement point data set establishing module 15 , measurement analysis module 16 . DETAILED DESCRIPTION

[0011] The present application provides an intelligent measurement method for an online component analyzer, which is used to solve the technical problem in the prior art of low accuracy and stability of measurement results due to unreasonable distribution of measurement points and environmental interference during the measurement process.

[0012] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0013] It should be noted that the terms "first", "second", etc. in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products, or devices. Embodiment 1

[0014] like Figure 1 As shown, the present application provides an intelligent measurement method for an online component analyzer, the method comprising: P10: Perform pre-measurement inspection and build a pre-dataset.

[0015] Specifically, to ensure the accuracy and reliability of the measurement, a pre-measurement detection is performed, that is, scene data collection of the target measurement scene is performed through a preset collection device, which may include sensors, cameras, spectrometers, etc., which can capture various physical and chemical information of the target scene. By collecting and detecting the scene data of the target measurement scene, a pre-data set is constructed, covering various characteristics and information of the target measurement scene, including but not limited to environmental parameters, equipment status, physical and chemical properties of the target object, etc. Through the construction of the pre-data set, the characteristics and conditions of the target measurement scene can be more comprehensively understood, thereby improving the accuracy and reliability of the measurement, and providing key data support for subsequent analysis and measurement.

[0016] P20: extracting scene features and environmental features from the previous data set, setting the number of initial points based on the scene features and environmental features, and performing random distribution of the initial points under the number of initial points.

[0017] Optionally, scene features and environmental features are extracted from the previous data set. The scene feature extraction refers to identifying and extracting features directly related to the target measurement scene from the previous data set, including the layout, structure, lighting conditions, object distribution, etc. of the scene. The environmental feature extraction refers to extracting features related to the measurement environment from the previous data set, including temperature, humidity, pressure, gas concentration, etc.

[0018] Furthermore, after extracting the scene features and environmental features, it is necessary to set the number of initial points according to these features, that is, to set the number of initial measurement points. The number of initial points has an important impact on the subsequent cluster analysis and the selection of measurement points. Too many initial points may lead to excessive calculations, while too few initial points may not be able to fully cover the measurement scene. Therefore, it is necessary to reasonably set the number of initial points according to actual conditions and specific needs. After setting the number of initial points, it is necessary to randomly distribute the initial points in the measurement scene. Random distribution can ensure that the initial points are evenly distributed throughout the measurement scene, thereby more comprehensively reflecting the characteristics of the measurement environment, and at the same time avoiding measurement deviations caused by excessive concentration of initial points. Random distribution of initial points can provide strong support for subsequent cluster analysis and selection of measurement points.

[0019] P30: Perform cluster analysis on the initial points and perform position iteration. Update the positions and quantities of the initial points according to the position iteration results and clustering results, establish a cluster data set, map each cluster data in the cluster data set to a cluster center, use the cluster center as the updated measurement point, and establish the position weight of the corresponding cluster center through the cluster data set.

[0020] It should be understood that cluster analysis is performed on the initial points by K-means, hierarchical clustering, etc., that is, measuring points with similar characteristics or similar position distribution are grouped into the same category, so as to find points with similar distribution or similar characteristics in the measurement scene. Further, position iteration is performed, that is, the position of the cluster center (i.e., the center point of each cluster) is iteratively adjusted to minimize the distance sum of the data points within the cluster, ensuring that the cluster center more accurately represents the points in the cluster to which it belongs.

[0021] Further, the position and number of initial points are updated according to the position iteration results and the cluster analysis results. Exemplarily, if there are too many or too dense data points in a cluster, new initial points can be added; if there are too few or too sparse data points in a cluster, some initial points can be deleted or adjacent clusters can be merged. Then a cluster data set is generated, which contains information about each cluster, including the position of the cluster center, the number of data points contained in the cluster, etc. Moreover, each cluster data in the cluster data set is mapped to a cluster center, and this cluster center is used as the updated measurement point.

[0022] Furthermore, the position weights of the corresponding cluster centers are established through the clustering data set. The position weights of the cluster centers can be allocated according to the number of initial points corresponding to each cluster center and the importance or representativeness of the cluster center in the measurement scene, which can be used to guide the selection and weight allocation of measurement points in the future.

[0023] Furthermore, step P30 of the embodiment of the present application further includes: P31: Set a preset search range. After the initial point is distributed, perform a similarity search based on the initial point based on the preset search range to generate similar search results; P32: Perform fitness-based initial point position update within the similar search results, establish clusters based on the similar search results, call the search range update coefficient based on the size of the cluster, execute the preset search range update, and complete a round of search iterations; P33: Establish a cluster data set based on the results of multiple rounds of search iterations.

[0024] Specifically, the specific process of establishing the clustering data set may be, first, setting a preset search range. This range limits the boundary of the initial point search to ensure that the search process does not exceed the effective area of ​​the measurement scene. Further, after the initial point is distributed, a search for similar initial points is performed based on the initial point within the preset search range to generate similar search results. Then, within the similar search results, the initial point position is updated according to fitness, such as distance, density, etc., that is, these points are fine-tuned or moved to a new position in the hope of finding a better solution.

[0025] At the same time, a cluster is established based on the updated initial point, and the search range update coefficient is called according to the size of the cluster to perform the update of the preset search range. The search range update coefficient is a coefficient used to adjust the size of the search range. For example, if a cluster is large, it may mean that there are a large number of potential good solutions in the area. The search range update coefficient is used to increase the search range to explore more areas. Conversely, if a cluster is small, the search range is reduced to search the current area more finely. This process will be iterated for multiple rounds to optimize the clustering results and search range. Finally, based on the results of multiple rounds of search iterations, a clustering data set is established. The clustering data set contains the optimized clustering information and cluster centers, that is, the updated measurement points, which can be used for subsequent measurement analysis and decision-making processes.

[0026] Furthermore, step P33 of the embodiment of the present application also includes: P33-1: Determine whether there are iterative results without clustering cluster update in the results of multiple rounds of iterations; P33-2: If there are iterative results without clustering cluster update in the results of multiple rounds of iterations, generate new instructions, add new initial points in the non-clustered area according to the new instructions, and establish a cluster data set based on the new initial points.

[0027] Optionally, the process of establishing a clustering data set based on the results of multiple rounds of search iterations also includes, first, after multiple rounds of iterative search and cluster cluster update, it is necessary to determine whether there are iterative results without cluster cluster update, that is, whether there is a situation where all cluster clusters have not been updated, that is, all initial points have been clustered under the current number of cluster points.

[0028] If there are iterative results in which cluster clusters are not updated in multiple rounds of iteration results, it is necessary to generate new instructions, add new initial points in the non-clustered area according to the new instructions, and establish a cluster data set based on the newly added initial points. Specifically, new instructions can be generated according to the size and shape of the non-clustered area and the distribution characteristics of the entire measurement scene, and the new instructions include the location, quantity and other relevant parameters of the newly added initial points. Then, new initial points are added in the non-clustered area according to the new instructions, and cluster analysis is re-executed based on these newly added initial points to establish a new cluster data set. In this way, it is ensured that the entire measurement scene is effectively covered and clustered to avoid omissions or repeated measurements. At the same time, the newly added initial points can also improve the accuracy and efficiency of clustering, making the clustering results more in line with the actual situation.

[0029] Furthermore, the embodiment of the present application further includes step P33a, and step P33a further includes: P33-1a: Determine whether there is a merger between clusters in the results of multiple rounds of iterations.

[0030] P33-2a: If there is a merger between clusters in the results of multiple rounds of iterations, add a new initial point in the unclustered area and establish a clustering data set based on the newly added initial point.

[0031] Specifically, during the iterative optimization process, it is determined whether there is a merger between clusters in the results of multiple rounds of iterations. If there is such a merger between clusters, it indicates that the search has converged to a local optimal solution area and has failed to fully explore the entire search space. Therefore, in order to avoid falling into a local optimal solution, in the search space, in addition to the clustered areas, the unclustered areas that have not been covered or considered by the clustering algorithm are identified, and a certain number of initial points are added to the unclustered areas. The selection of these initial points should cover different positions of the unclustered areas as much as possible to increase the diversity of the search. The newly added initial points are used as new data points, and cluster analysis is performed together with the original data points to establish a new clustering data set. In this way, the search range is effectively expanded and the possibility of finding the global optimal solution is increased.

[0032] P40: Perform environmental impact compensation fitting based on the environmental impact database and establish an environmental compensation network.

[0033] Further, such as Figure 2 As shown, step P40 of the embodiment of the present application also includes: P41: Establish an environmental impact database for an online component analyzer, and perform data preprocessing on the environmental impact database to generate a preprocessed data set; P42: Sort the preprocessed data set in sequence according to the environmental status, and establish a sequential sorting result; P43: Perform a balanced analysis of the environmental impact value based on the sequential sorting result, and establish a balanced impact value; P44: Perform a abnormal analysis of the environmental impact value through the balanced impact value, and establish N abnormal nodes; P45: Perform data segmentation on the preprocessed data set through the N abnormal nodes, perform environmental impact compensation fitting based on the data segmentation result, and construct an environmental compensation network.

[0034] It should be understood that the environmental impact compensation fitting is performed based on the preprocessed data set. Specifically, first, in order to analyze the impact of environmental data on the measurement results of the online component analyzer, it is necessary to collect environmental data related to the measurement results of the online component analyzer through real-time monitoring equipment, historical data records, literature, etc., including temperature, humidity, pressure, gas concentration, wind speed, wind direction and other environmental factors, sort the collected data, and remove duplicate, erroneous or invalid data. And select an appropriate data storage method, such as a relational database or a distributed file system, store the sorted data in the database, and generate an environmental impact database.

[0035] Furthermore, the data in the environmental impact database are preprocessed, including data cleaning, removing duplicates and outliers in the data, and processing missing values ​​through interpolation, and then integrating the cleaned data to integrate data from different sources into a complete data set. In the integration process, data format conversion and data unification are performed to ensure data consistency and comparability. For large amounts of data, data reduction is performed to reduce the amount of calculation and improve analysis efficiency, and a preprocessed data set is generated to provide strong support for subsequent data analysis and modeling.

[0036] Next, the environmental status fields in the preprocessed data set are identified, such as temperature, humidity, pollution concentration, etc., and the data are sorted sequentially according to the values ​​or change trends of these environmental status fields to establish a sequential sorting result. The sequential sorting result is an ordered data set that reflects the changes in environmental status over time or space.

[0037] Furthermore, statistical methods such as regression analysis and variance analysis are used to conduct a balanced analysis of the environmental impact values ​​of the sorted data, calculate statistical indicators such as the mean value and standard deviation of the environmental impact values ​​under different environmental conditions, and establish a balanced impact value based on the calculation results, that is, the concentrated value of each environmental impact value. The balanced impact value can also be a mathematical expression or model that reflects the stable relationship between the environmental state and the environmental impact value.

[0038] Furthermore, an abnormal analysis of the environmental impact value is performed through the equilibrium impact value, and the equilibrium impact value is analyzed using data mining or machine learning algorithms, such as cluster analysis, anomaly detection algorithms, etc., and by setting thresholds or judgment conditions, data points that significantly deviate from the equilibrium impact value are identified as abnormal nodes. That is, abnormal points or mutation points in the data are identified through abnormal analysis, and N abnormal nodes are established. The abnormal nodes are nodes where sudden changes in environmental conditions or abnormal events have a significant impact on the environmental impact value.

[0039] Furthermore, the N abnormal nodes are used as segmentation points to divide the preprocessed data set into different data segments. By using appropriate mathematical models, such as neural networks, support vector machines, etc., environmental impact compensation fitting is performed on each data segment separately, that is, data supervised training is performed, and the fitting results of each data segment are combined to form a complete environmental compensation network. The environmental compensation network can reflect the comprehensive impact of changes in environmental conditions on environmental impact values, and can be used to predict or control the occurrence and extent of environmental impacts.

[0040] Furthermore, step P45 of the embodiment of the present application also includes: P45-1: Establish N+1 environmental compensation sub-networks corresponding to the data segmentation results, and map the N+1 environmental compensation sub-networks to the data segmentation results; P45-2: When training any environmental compensation sub-network P, use the corresponding P-th data segmentation result as the first weight data, and the P-1-th data segmentation result and the P+1-th data segmentation result as the second weight data, and perform fitting training; P45-3: When the training of all environmental compensation sub-networks is completed, construct an environmental compensation network based on all environmental compensation sub-networks.

[0041] Optionally, the specific process of constructing an environmental compensation network through data segmentation results can be, first, using an appropriate mathematical model, such as a neural network, a support vector machine, etc., according to the data segmentation results, to establish an environmental compensation sub-network for each data segment. Since there are N abnormal nodes, the data will be divided into N+1 parts, and N+1 sub-networks need to be established accordingly. At the same time, a mapping relationship is established between each environmental compensation sub-network and the corresponding data segmentation result, so that the data segment can be accurately extracted during training.

[0042] Furthermore, an environmental compensation subnetwork P is randomly selected for training, and the Pth data segmentation result corresponding to the subnetwork is used as the main training data, and a higher weight is assigned, namely, the first weight data. At the same time, in order to enhance the generalization ability of the subnetwork and take into account the continuity of the data, the adjacent P-1th and P+1th data segmentation results are also used as training data, but with a lower weight, namely, the second weight data. Fitting training is performed using this weighted method to optimize the performance of the subnetwork.

[0043] By analogy, all N+1 environmental compensation sub-networks are trained in sequence until all environmental compensation sub-networks have completed training, and all environmental compensation sub-networks are combined to form a complete environmental compensation network. The environmental compensation network can cover the entire environmental state space and provide corresponding compensation according to different environmental states.

[0044] P50: When intelligent measurement processing is performed by updating the measurement points, real-time environmental data collection of the updated measurement points is performed synchronously, correction compensation is performed based on the real-time environmental data collection results and the environmental compensation network, and a compensated updated measurement point data set is established.

[0045] Specifically, when intelligent measurement processing is performed through the above-mentioned updated measurement points, appropriate sensors or measuring devices are used to collect real-time environmental data of the updated measurement points, including collecting parameters related to the environment such as temperature, humidity, light, pressure, sound, etc., and the real-time environmental data is preprocessed as necessary, such as filtering, standardization or normalization, so as to be compatible with the environmental compensation network.

[0046] Furthermore, the preprocessed real-time environmental data is input into the environmental compensation network, which automatically calculates the corresponding correction compensation value based on the input real-time environmental data, and adjusts or corrects the measurement error of each measuring point caused by environmental factors through the correction compensation value, and establishes an updated measurement point data set based on the corrected and compensated measurement data, thereby ensuring the accuracy and reliability of the measurement results.

[0047] P60: Perform measurement analysis of the target measurement scene using the position weight and the updated measurement point data set, and establish a measurement analysis result.

[0048] Optionally, the position weight and the updated measurement point data set are used to perform measurement analysis of the target measurement scene, and a weighted multiple regression model can be designed for each measurement point, taking into account the influence of environmental factors, with the measurement value as the dependent variable and the environmental factors and other related variables as independent variables. Position weights are introduced, and multiple weighted multiple regression models are fitted to adjust the importance of the contribution of different measurement points to the model, and based on the fitted weighted multiple regression model, the measurement values ​​in the target measurement scene are corrected, the degree and direction of the influence of environmental factors on the measurement values, and the influence of position weights on the prediction results are analyzed, and the measurement analysis results are established to improve the accuracy and reliability of the measurement results.

[0049] Furthermore, the embodiment of the present application further includes step P60a, and step P60a further includes: P61a: Perform abnormality evaluation on the location points of the updated measurement point data set to generate abnormality evaluation results; P62a: Construct additional coefficients based on the abnormality evaluation results, and establish measurement analysis results after compensating the updated point data set with the additional coefficients.

[0050] Exemplarily, in order to further improve the accuracy of the measurement results, an abnormality detection algorithm is improved, such as a statistical method, a distance method, a density method, etc., to identify abnormal location points in the updated measurement point data set, which may be caused by measurement errors, equipment failures, environmental factors or other unknown reasons. And according to the detection results of the abnormal location points, each location point is evaluated for abnormality, for example, according to the proportion of the number of abnormal location points, a corresponding abnormality evaluation score or label is generated as the abnormality evaluation result.

[0051] Furthermore, based on the abnormality evaluation result, an additional coefficient is constructed for each location point. Exemplarily, the additional coefficient can be a value of 0~1, 1 or greater than 1, depending on the result of the abnormality evaluation. An additional coefficient greater than 1 is used to increase the weight of the location point, and an additional coefficient of 0~1 is used to reduce the weight, and 1 indicates no abnormality. The additional coefficient is used to update the point data set compensation, for example, the measurement value of each location point is multiplied by the corresponding additional coefficient to correct the measurement value, thereby establishing a more accurate and reliable measurement analysis result. It helps to improve the accuracy and practicality of data analysis and support more accurate decision-making and optimization operations.

[0052] In summary, the embodiments of the present application have at least the following technical effects: This application sets the number of initial points by extracting scene features and environmental features, and randomly distributes the initial points, performs cluster analysis on the initial points, establishes updated measurement points and corresponding position weights, collects real-time environmental data, corrects and compensates the measurement point data set through the environmental compensation network, and establishes measurement analysis results in combination with the position weights.

[0053] The technical effect of improving the measurement accuracy and stability of the online component analyzer by optimizing the distribution of measurement points and compensating for environmental influences is achieved. Embodiment 2

[0054] Based on the same inventive concept as the intelligent measurement method for an online component analyzer in the aforementioned embodiment, Figure 3 As shown, the present application provides an intelligent measurement system for an online component analyzer, and the system and method embodiments in the present application are based on the same inventive concept. The system includes: The pre-measurement detection module 11 is used to perform pre-measurement detection and construct a pre-data set.

[0055] The initial point setting module 12 is used to extract scene features and environmental features from the pre-set data set, set the number of initial points based on the scene features and environmental features, and perform random distribution of the initial points under the number of initial points.

[0056] A position weight establishing module 13 is used to perform cluster analysis on the initial points and execute position iteration, update the positions and quantities of the initial points according to the position iteration results and the clustering results, establish a cluster data set, each cluster data in the cluster data set maps a cluster center, use the cluster center as an updated measurement point, and establish the position weight of the corresponding cluster center through the cluster data set.

[0057] The environment compensation network establishment module 14 is used to perform environment impact compensation fitting based on the environment impact database to establish an environment compensation network.

[0058] The update measurement point data set establishment module 15 is used to synchronously collect real-time environmental data of the update measurement point when performing intelligent measurement processing through the update measurement point, perform correction compensation based on the real-time environmental data collection result and the environmental compensation network, and establish a compensated update measurement point data set.

[0059] The measurement analysis module 16 is used to perform measurement analysis of the target measurement scene through the position weight and the updated measurement point data set, and establish a measurement analysis result.

[0060] Furthermore, the position weight establishing module 13 is also used to perform the following steps: A preset search range is set, and after the initial point is distributed, a similar search based on the initial point is performed based on the preset search range to generate similar search results; the initial point position is updated based on fitness within the similar search results, and a cluster is established based on the similar search results, and a search range update coefficient is called based on the size of the cluster, and a preset search range update is executed to complete a round of search iteration; A clustering data set is established based on the results of multiple rounds of search iterations.

[0061] Furthermore, the position weight establishing module 13 is also used to perform the following steps: Determine whether there are any iteration results without clustering cluster update in multiple iteration results; If there are iteration results without clustering cluster update in multiple iteration results, a new instruction is generated, and a new initial point is added in the non-clustered area according to the new instruction, and a clustering data set is established according to the new initial point.

[0062] Furthermore, the position weight establishing module 13 is also used to perform the following steps: Determine whether there is a merger between clusters in the results of multiple rounds of iterations; if there is a merger between clusters in the results of multiple rounds of iterations, add an initial point in the non-clustered area and establish a clustering data set based on the newly added initial point.

[0063] Furthermore, the environment compensation network establishing module 14 is also used to perform the following steps: An environmental impact database of an online component analyzer is established, and data preprocessing is performed on the environmental impact database to generate a preprocessed data set; the preprocessed data set is sorted in sequence according to the environmental state to establish a sorting result; a balanced analysis of environmental impact values ​​is performed based on the sorting result to establish a balanced impact value; a deformed analysis of environmental impact values ​​is performed through the balanced impact value to establish N deformed nodes; data segmentation is performed on the preprocessed data set through the N deformed nodes, and environmental impact compensation fitting is performed based on the data segmentation result to construct an environmental compensation network.

[0064] Furthermore, the environment compensation network establishing module 14 is also used to perform the following steps: Establish N+1 environmental compensation sub-networks corresponding to the data segmentation results, and map the N+1 environmental compensation sub-networks to the data segmentation results; when training any environmental compensation sub-network P, use the corresponding P-th data segmentation result as the first weight data, and use the P-1-th data segmentation result and the P+1-th data segmentation result as the second weight data to perform fitting training; when the training of all environmental compensation sub-networks is completed, construct an environmental compensation network based on all environmental compensation sub-networks.

[0065] Furthermore, the measurement and analysis module 16 is further configured to perform the following steps: An abnormality evaluation is performed on the updated measurement point data set to generate an abnormality evaluation result; an additional coefficient is constructed according to the abnormality evaluation result, and after compensating the updated point data set by the additional coefficient, a measurement analysis result is established.

[0066] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0067] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

[0068] This specification and drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.

Claims

1. An intelligent measurement method for an online component analyzer, characterized in that: The method comprises: Perform pre-measurement inspection and build pre-dataset; Extracting scene features and environmental features from the pre-set data set, setting the number of initial points based on the scene features and environmental features, and performing random distribution of the initial points under the number of initial points; Performing cluster analysis on the initial points and performing position iteration, updating the positions and quantities of the initial points according to the position iteration results and the clustering results, establishing a cluster data set, wherein each cluster data in the cluster data set is mapped to a cluster center, the cluster center is used as an updated measurement point, and the position weight corresponding to the cluster center is established through the cluster data set; Perform environmental impact compensation fitting based on the environmental impact database and establish an environmental compensation network; When intelligent measurement processing is performed by updating the measurement point, real-time environmental data collection of the updated measurement point is synchronously performed, correction compensation is performed based on the real-time environmental data collection result and the environmental compensation network, and a compensated updated measurement point data set is established; The measurement analysis of the target measurement scene is performed by using the position weight and the updated measurement point data set to establish a measurement analysis result.

2. The intelligent measurement method for an online component analyzer according to claim 1, characterized in that: The performing of environmental impact compensation fitting based on the environmental impact database to establish an environmental compensation network also includes: Establishing an environmental impact database of an online component analyzer, and performing data preprocessing on the environmental impact database to generate a preprocessed data set; Sorting the preprocessed data set in sequence according to the environmental state, and establishing a sequence sorting result; Performing a balanced analysis of environmental impact values ​​based on the sequence ranking results to establish a balanced impact value; Performing abnormal analysis of environmental impact values ​​through the equilibrium impact values ​​to establish N abnormal nodes; The preprocessed data set is segmented by the N abnormal nodes, and environmental impact compensation fitting is performed based on the data segmentation result to construct an environmental compensation network.

3. An intelligent measurement method for an online component analyzer as claimed in claim 2, characterized in that: The step of performing data segmentation on the preprocessed data set by using the N abnormal nodes, performing environmental impact compensation fitting based on the data segmentation result, and constructing an environmental compensation network further includes: Establishing N+1 environmental compensation sub-networks corresponding to the data segmentation results, and mapping the N+1 environmental compensation sub-networks to the data segmentation results; When training any environment compensation sub-network P, the corresponding P-th data segmentation result is used as the first weight data, and the P-1-th data segmentation result and the P+1-th data segmentation result are used as the second weight data to perform fitting training; When all the environment compensation sub-networks are trained, an environment compensation network is constructed based on all the environment compensation sub-networks.

4. The intelligent measurement method for an online component analyzer according to claim 1, characterized in that: The clustering analysis is performed on the initial points, and position iteration is performed, and the position and quantity of the initial points are updated according to the position iteration result and the clustering result, and a clustering data set is established, including: Setting a preset search range, after the initial points are distributed, performing a similar search based on the initial points based on the preset search range, and generating similar search results; Performing fitness-based initial point position update within the similar search results, establishing clusters based on the similar search results, calling search range update coefficients based on the size of the clusters, executing preset search range updates, and completing a round of search iterations; A clustering data set is established based on the results of multiple rounds of search iterations.

5. The intelligent measurement method for an online component analyzer according to claim 4, characterized in that: The step of establishing a clustering data set according to the results of multiple rounds of search iterations includes: Determine whether there are any iteration results without clustering cluster update in multiple iteration results; If there are iteration results without clustering cluster update in multiple rounds of iteration results, a new instruction is generated, and a new initial point is added in the non-clustered area according to the new instruction, and a clustering data set is established according to the new initial point.

6. The intelligent measurement method for an online component analyzer according to claim 5, characterized in that: The method further comprises: Determine whether there is a merger between clusters in the results of multiple rounds of iterations; If there is a merger between clusters in the results of multiple rounds of iterations, new initial points are added in the unclustered area, and a clustering data set is established based on the new initial points.

7. The intelligent measurement method for an online component analyzer according to claim 1, characterized in that: The method further comprises: Performing abnormality evaluation on the location points of the updated measurement point data set to generate an abnormality evaluation result; An additional coefficient is constructed according to the abnormal evaluation result, and after the update point data set is compensated by the additional coefficient, the measurement analysis result is established.

8. An intelligent measurement system for an online component analyzer, characterized in that: The system comprises: A pre-measurement detection module, the pre-measurement detection module is used to perform pre-measurement detection and construct a pre-data set; An initial point setting module, the initial point setting module is used to extract scene features and environmental features from the pre-set data set, set the number of initial points based on the scene features and environmental features, and perform random distribution of the initial points under the number of initial points; A position weight establishment module, the position weight establishment module is used to perform cluster analysis on the initial points and perform position iteration, update the position and quantity of the initial points according to the position iteration result and the clustering result, establish a cluster data set, each cluster data in the cluster data set is mapped to a cluster center, the cluster center is used as an updated measurement point, and the position weight corresponding to the cluster center is established through the cluster data set; An environmental compensation network establishment module, the environmental compensation network establishment module is used to perform environmental impact compensation fitting based on the environmental impact database to establish an environmental compensation network; An update measurement point data set establishment module, wherein the update measurement point data set establishment module is used to synchronously collect real-time environmental data of the update measurement point when performing intelligent measurement processing through the update measurement point, perform correction compensation based on the real-time environmental data collection result and the environmental compensation network, and establish a compensated update measurement point data set; A measurement analysis module is used to perform measurement analysis of a target measurement scene through the position weight and the updated measurement point data set, and establish a measurement analysis result.

Images