Multi-source heterogeneous data fusion and scheduling method based on industrial internet data center

Abstract

The invention provides a multi-source heterogeneous data fusion and scheduling method based on an industrial internet data center, and the method comprises the steps: carrying out the full-quality data sampling of an industrial internet of things, obtaining a plurality of pieces of object data, carrying out the key quality index detection, and obtaining an object index difference set; performing corresponding analysis on the data source entities to obtain a data source-scene type matrix, performing traversal to perform fusion quality evaluation, and positioning heterogeneous data source entities; performing scheduling hierarchical classification and multi-level to-be-scheduled entities on heterogeneous data source entities in a standard process of obtaining a data source line; multi-level progressive scheduling is performed on heterogeneous data source entities, and data source index scheduling of data source lines obtained by the industrial Internet of Things is performed, so that efficient data fusion of multi-source heterogeneous data is realized, the problem of large data difference in an industrial Internet environment is solved, potential data values are fully mined, tasks can be reasonably arranged, and the data fusion efficiency is improved. The data processing timeliness and the resource utilization rate are improved, and the industrial production cost is reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of industrial internet, and in particular to a multi-source heterogeneous data fusion and scheduling method based on an industrial internet data middle platform. BACKGROUND

[0002] With the development of science and technology, the application ability of big data is becoming more and more comprehensive. Among them, the application of big data in industrial internet is gradually popularized, and massive multi-source heterogeneous data is showing explosive growth.

[0003] However, the integration of multi-source heterogeneous data in industrial scenarios is difficult, and the traditional data processing and scheduling method cannot meet the complex and changeable business requirements. Specifically, the data format, semantics and structure differ greatly, making it difficult to integrate and utilize, and the processing efficiency is low and the data value is not fully mined.

[0004] Therefore, how to realize efficient data fusion of multi-source heterogeneous data, fully mine the potential value of data, reasonably arrange tasks, improve data processing timeliness and resource utilization, and reduce industrial production cost is a problem to be solved at present. SUMMARY

[0005] The present application provides a multi-source heterogeneous data fusion and scheduling method based on an industrial internet data middle platform, which realizes efficient data fusion of multi-source heterogeneous data, fully mines the potential value of data, reasonably arranges tasks, improves data processing timeliness and resource utilization, and reduces industrial production cost.

[0006] In one aspect, the present application provides a multi-source heterogeneous data fusion and scheduling method based on an industrial internet data middle platform, which includes:

[0007] In the finished product stage of the industrial internet data source line, the industrial internet full-quality data sampling is performed to obtain P object data; wherein the industrial internet data source line includes W data source entities;

[0008] The key quality indicators of each object data are detected to obtain the indicator difference set of each object data;

[0009] Based on the industrial internet fusion log, the W data source entities are analyzed to obtain a data source-scene type matrix;

[0010] According to the indicator difference set of each object data, the data source-scene type matrix is traversed for fusion quality evaluation to locate H heterogeneous data source entities; wherein H is less than W, and the H heterogeneous data source entities have H heterogeneous confidence degrees;

[0011] According to the H heterogeneous confidence and the standard process of the industrial internet of things data source line, the H heterogeneous data source entities are scheduled hierarchical, and a multi-level to-be-scheduled entity set is obtained;

[0012] The multi-level to-be-scheduled entity set is taken as a scheduling direction, and the H heterogeneous data source entities are executed multi-level progressive scheduling until the difference value of the key quality indicators of the industrial internet of things full-quality data is less than the preset difference value after the heterogeneous data source entities in the current level to-be-scheduled entity set are scheduled, and the scheduling is stopped.

[0013] According to the multi-source heterogeneous data fusion and scheduling method based on the industrial internet data middle platform provided by the application, the index difference set of each object data includes N quality index difference amounts;

[0014] According to the index difference set of each object data, the data source-scene type matrix is traversed for fusion quality evaluation, and H heterogeneous data source entities are located, including:

[0015] For the index difference set of the i-th object data, the j-th quality index difference amount is used to traverse the data source-scene type matrix, and the j-th deviation confidence of each of the W data source entities is obtained; i=1, 2, …, P; j=1, 2, …, N;

[0016] The j-th quality index difference amount is used to schedule the j-th deviation confidence of each of the W data source entities, and the j-th scheduling confidence of each of the W data source entities is obtained;

[0017] The all scheduling confidences of each of the W data source entities are weighted and summed to obtain the i-th heterogeneous confidence of each of the W data source entities;

[0018] The all heterogeneous confidences of each of the W data source entities are weighted and summed to obtain the total heterogeneous confidence of each of the W data source entities;

[0019] The H data source entities with the total heterogeneous confidence greater than the preset heterogeneous confidence are selected as the H heterogeneous data source entities.

[0020] According to the multi-source heterogeneous data fusion and scheduling method based on the industrial internet data middle platform provided by the application, based on the industrial internet fusion log, the W data source entities are analyzed, and a data source-scene type matrix is obtained, including:

[0021] From the industrial internet fusion log, the industrial internet obtained data source line is comprehensively collected, and the difference information set of the key quality indicators of each of the W data source entities in the history fusion is recorded;

[0022] The analysis is performed on each difference information set to obtain a difference frequency of each key quality indicator in each difference information set, a difference amplitude of each key quality indicator, and a single difference duration of each key quality indicator;

[0023] A product of a weight of the difference amplitude of the key quality indicator and the single difference duration of the key quality indicator is calculated as a weighted total duration of the key quality indicator;

[0024] A ratio of the weighted total duration to the difference frequency of the key quality indicator is taken as a confidence duration of the key quality indicator;

[0025] A data source-scene type matrix with a behavior data source entity, a column as a key quality indicator, and an element as a confidence duration of the key quality indicator is constructed.

[0026] According to the H heterogeneous confidences and the industrial Internet of Things, a standard process of a data source line is obtained, the H heterogeneous data source entities are hierarchically scheduled to obtain a multi-level to-be-scheduled entity set, including:

[0027] H engineering smooth periods, H data source scheduling resources, and H data source scheduling complexities of the H heterogeneous data source entities are obtained through interaction;

[0028] H data source execution bit positions of the H heterogeneous data source entities are obtained by calling the standard process;

[0029] A scheduling heterogeneous quantization function is pre-constructed, and the H engineering smooth periods, the H data source scheduling resources, the H data source scheduling complexities, the H data source execution bit positions, and the H heterogeneous confidences are loaded into the scheduling heterogeneous quantization function to obtain H scheduling heterogeneous quantization values;

[0030] A scheduling hierarchical scale is preset, and the H heterogeneous data source entities are hierarchically scheduled according to the scheduling hierarchical scale and the H scheduling heterogeneous quantization values to obtain a multi-level to-be-scheduled entity set.

[0031] According to the industrial Internet data middle platform-based multi-source heterogeneous data fusion and scheduling method provided by the application, the scheduling heterogeneous quantization function is:

[0032]

[0033] Wherein, F is a scheduling heterogeneous quantization value, T is an engineering smooth period, C is a data source scheduling resource, O is a data source scheduling complexity, S is a data source execution bit position, and R is a heterogeneous confidence.

[0034] According to the multi-source heterogeneous data fusion and scheduling method based on the industrial internet data middle platform provided by the application, the multi-level to-be-scheduled entity set is taken as the scheduling direction, and multi-level progressive scheduling is performed on the H heterogeneous data source entities until after the heterogeneous data source entities in the current level to-be-scheduled entity set are scheduled, the difference value of the key quality indicators of the industrial internet full-quality data is less than the preset difference value, and the scheduling is stopped, comprising:

[0035] Initializing the scheduling level: obtaining a multi-level to-be-scheduled entity set, starting from the first level to-be-scheduled entity set with the highest priority;

[0036] Performing current level scheduling: according to the bit order of the H data sources, optimizing the to-be-scheduled heterogeneous data source entities in the current level to-be-scheduled entity set one by one, and using a preset scheduling algorithm to determine the optimal data source indicators of each to-be-scheduled heterogeneous data source entity, and aggregating all the optimal data source indicators to obtain the control indicators of the current level to-be-scheduled entity set;

[0037] Scheduling effect verification: taking the control indicators of all the scheduled entity sets as the benchmark, and according to the control indicators of the current level to-be-scheduled entity set, detecting whether the difference of the key quality indicators of the scheduled industrial internet full-quality data is less than the preset difference value after the data source indicators of the industrial internet data source system are scheduled;

[0038] Decision-making for next scheduling: if the difference of the key quality indicators of the scheduled industrial internet full-quality data is less than the preset difference value, stop scheduling, and if the difference of the key quality indicators of the scheduled industrial internet full-quality data is not less than the preset difference value, enter the scheduling of the next level to-be-scheduled entity set.

[0039] According to the multi-source heterogeneous data fusion and scheduling method based on the industrial internet data middle platform provided by the application, industrial internet full-quality data sampling is performed to obtain a plurality of object data, comprising:

[0040] A preset sampling period window, a sampling confidence interval and a sampling confidence threshold;

[0041] Industrial internet full-quality data sampling is performed on the finished product stage of the industrial internet data source line in the sampling period window to obtain K backup data, wherein the K backup data has K sampling period labels;

[0042] The K backup data are evaluated to obtain a K data source indicator sequence;

[0043] The K backup data are classified into M backup data groups according to the K sampling period labels under the grouping constraint of the sampling confidence interval, wherein M is a positive integer less than K.

[0044] After the M standby data groups are evaluated according to the K data source index sequences, the sampling confidence threshold is used as a filtering environment to filter the P object data from the M standby data groups.

[0045] According to the multi-source heterogeneous data fusion and scheduling method based on the industrial internet data middle platform provided by the application, after the M standby data groups are evaluated according to the K data source index sequences, the sampling confidence threshold is used as a filtering environment to filter the P object data from the M standby data groups, which comprises the following steps:

[0046] According to the mapping relationship between the M standby data groups and the K standby data, the K data source index sequences are classified into M data source index sequence groups;

[0047] The data source criticality of the industrial internet data source line is evaluated to obtain W entity criticalities of the W data source entities, wherein W is a positive integer greater than or equal to H;

[0048] The first data source index sequence and the second data source index sequence are obtained from the mth data source index sequence group; m = 1, 2, … M;

[0049] After the W entity correlation degrees of the first data source index sequence and the second data source index sequence are calculated according to the cosine distance, the W entity correlation degrees are weighted and calculated according to the W entity criticalities to obtain a first data source correlation degree;

[0050] The mth data source index sequence group is combined and enumerated, and the data source correlation degree is calculated to obtain an mth data source correlation degree set of the mth data source index sequence group;

[0051] According to the mth data source correlation degree set, the mth object data whose data source correlation degrees all satisfy the sampling confidence threshold is filtered from the mth standby data group;

[0052] All the object data of the groups are combined to obtain the P object data.

[0053] On the other hand, the application further provides a multi-source heterogeneous data fusion and scheduling system based on the industrial internet data middle platform, which comprises:

[0054] The sampling module is used to sample the full quality data of the industrial internet in the product stage of the industrial internet data source line to obtain P object data; wherein the industrial internet data source line comprises W data source entities;

[0055] The detection module is configured to detect each object data based on a key quality indicator to obtain an indicator difference set of each object data.

[0056] The analysis module is configured to analyze the W data source entities based on the industrial internet fusion log to obtain a data source-scene type matrix.

[0057] The positioning module is configured to traverse the data source-scene type matrix based on the indicator difference set of each object data to perform fusion quality evaluation and locate H heterogeneous data source entities, wherein H is less than W, and the H heterogeneous data source entities have H heterogeneous confidences.

[0058] The hierarchical module is configured to perform hierarchical scheduling on the H heterogeneous data source entities based on the H heterogeneous confidences and a standard process of a data source line of the industrial internet to obtain a multi-level to-be-scheduled entity set.

[0059] The scheduling module is configured to perform multi-level progressive scheduling on the H heterogeneous data source entities by taking the multi-level to-be-scheduled entity set as a scheduling direction, and stop the scheduling when the industrial internet full-quality data is less than a preset difference value in the difference value of the key quality indicator after scheduling the heterogeneous data source entities in a current level to-be-scheduled entity set.

[0060] In another aspect, the present application also provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the above-mentioned multi-source heterogeneous data fusion and scheduling method based on the industrial internet data platform when executing the program.

[0061] In another aspect, the present application also provides a non-transitory computer readable storage medium, which stores a computer program executable by a processor to implement the above-mentioned multi-source heterogeneous data fusion and scheduling method based on the industrial internet data platform.

[0062] In another aspect, the present application also provides a computer program product, which includes a computer program executable by a processor to implement the above-mentioned multi-source heterogeneous data fusion and scheduling method based on the industrial internet data platform.

[0063] The application provides a multi-source heterogeneous data fusion and scheduling method based on an industrial internet data middle platform, performs industrial internet of things full-quality data sampling, obtains a plurality of object data for key quality index detection, and obtains an object index difference set; a data source entity corresponding analysis obtains a data source-scene type matrix, fusion quality evaluation is performed through iteration, and heterogeneous data source entities are located; a standard process of a data source line is obtained to perform scheduling level classification and multi-level to-be-scheduled entities on the heterogeneous data source entities; multi-level progressive scheduling is performed on the heterogeneous data source entities, and data source index scheduling of the data source line is obtained through the industrial internet of things, efficient multi-source heterogeneous data fusion is realized, the problem of large data difference in the industrial internet environment is solved, the potential value of data is fully tapped, tasks can be reasonably arranged, data processing timeliness and resource utilization rate are improved, and industrial production cost is reduced. BRIEF DESCRIPTION OF DRAWINGS

[0064] In order to more clearly illustrate the technical solutions in the present application or prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative effort.

[0065] Figure 1 is a flowchart of the multi-source heterogeneous data fusion and scheduling method based on the industrial internet data middle platform provided by the embodiment of the present application.

[0066] Figure 2 is a structural schematic diagram of the multi-source heterogeneous data fusion and scheduling system based on the industrial internet data middle platform provided by the embodiment of the present application.

[0067] Figure 3 is a structural schematic diagram of the electronic device provided by the embodiment of the present application. DETAILED DESCRIPTION

[0068] In order to make the purpose, technical solutions and advantages of the present application more clear, the technical solutions in the present application will be described clearly and completely in the following with reference to the drawings in the present application. Obviously, the described embodiments are some embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative effort belong to the protection scope of the present application.

[0069] Figure 1 is a flowchart of the multi-source heterogeneous data fusion and scheduling method based on the industrial internet data middle platform provided by the embodiment of the present application.

[0070] As Figure 1As shown, the execution subject of the multi-source heterogeneous data fusion and scheduling method based on the industrial internet data platform provided in this embodiment of the invention can be an electronic device. The method mainly includes the following steps:

[0071] 101. In the finished product stage of obtaining data source lines in the Industrial Internet of Things (IIoT), perform full-quality data sampling of the IIoT to obtain data on P objects.

[0072] In a specific implementation, a sampling period window, a sampling confidence interval, and a sampling confidence threshold can be preset. Within the sampling period window, full-quality data sampling of the industrial IoT data source line at the finished product stage is performed to obtain K backup data points, where each of the K backup data points has a sampling period label. The K backup data points are then evaluated using a fusion data source to obtain K data source indicator sequences. Using the sampling confidence interval as a grouping constraint, the K backup data points are categorized into M backup data groups based on the K sampling period labels, where M is a positive integer less than K. After evaluating the M backup data groups based on the K data source indicator sequences, the sampling confidence threshold is used as a filtering environment to filter and obtain the P object data points from the M backup data groups. The industrial IoT data source line includes W data source entities.

[0073] Specifically, in the finished product stage of the Industrial Internet of Things (IIoT) data acquisition line, an appropriate sampling period window should first be set. This window defines the rules for object extraction within a certain period, typically considering factors such as the fusion rhythm and data source fluctuations. The sampling period length can be set based on the speed of the fusion process and the magnitude of data source stability fluctuations. Based on the sampling period window, full-quality IIoT data sampling can be performed from the finished product stage of the IIoT data acquisition line, using methods such as random sampling or stratified sampling to ensure that objects at each stage and link are covered. K backup data points are obtained through sampling to ensure the representativeness of the objects and avoid discrepancies in results caused by sampling only a portion of the objects.

[0074] The sampling confidence interval can be determined based on the fusion scale, product quality requirements, and standard operating procedures (SOPs). For example, the larger the fusion scale, the larger the sampling confidence interval; the higher the product quality requirements, the smaller the sampling confidence interval; and the more SOPs, the smaller the sampling execution interval. K spare data points can be categorized into M spare data groups according to their sampling period labels to ensure that each group has a relatively similar fusion period. The sampling confidence threshold refers to the maximum allowable error range during quality control. By setting a reasonable sampling confidence threshold, product quality can be ensured to remain within acceptable limits. From the M spare data groups, object data that meets the sampling confidence threshold is filtered out, and data exceeding the range is removed, resulting in multiple sample data points. By sampling at the finished product stage, product quality fluctuations can be detected, potential quality problems can be identified in a timely manner, and the consistency of industrial IoT data quality in each batch of fusion can be guaranteed.

[0075] Specifically, in the quality assessment process, the first step is to set relevant sampling indicators, including the sampling period window, sampling confidence interval, and sampling confidence threshold. The sampling period window refers to a defined period range used to determine the timeframe within which object data can be collected. The sampling period window determines the sampling period interval; that is, sampling can only be conducted within this period. The sampling confidence interval is an indicator used to ensure data accuracy during sampling. It refers to the allowable error range between sampling points (e.g., within a sampling period window, the minimum interval between two adjacent samples is 15 minutes, and the maximum is 45 minutes). In other words, the setting of the sampling interval ensures a certain level of confidence to guarantee the reliability of the sampling. The sampling confidence threshold is used to filter and evaluate the quality of the sampled data. Only sampled data that meets certain confidence requirements will be included in the analysis.

[0076] Using a sampling period window as a sampling interval constraint, the period interval between samples is set according to the length of the sampling period window during the sampling process, ensuring that sampling activities do not overlap and meet sampling standards. Based on the sampling period window, full-quality data sampling of the Industrial Internet of Things (IIoT) data source line is performed at the finished product stage, resulting in K backup data points. Each data point corresponds to a sampling period label, used to label the specific period point of sample collection. The finished product stage of the IIoT data source line refers to the stage where, after acquisition, concentration, and other processing during the fusion process, the final full-quality IIoT data is formed. At this stage, the data is considered a semi-finished product or final product; only after necessary quality inspections can its compliance with standards be determined before further processing can proceed.

[0077] For each backup data set, a data source evaluation analysis can be performed to review and analyze all key data source entities the data encountered during the fusion process. The evaluation analysis examines the fusion records for each backup data set, including all involved data source metrics and equipment operating environments. After evaluating the fusion data sources for K backup data sets, a data source metric sequence is established for each backup data set. This sequence records the changes in data source metrics for each sample throughout the fusion process, reflecting every detail of the fusion process and aiding in the analysis of the relationship between these metrics and the final product quality. The data source metric sequence refers to the record of changes in all involved data source control metrics throughout the fusion process, consisting of a set of periodic data source data points, where each data point represents the data source environment at a specific moment or stage.

[0078] Based on the sampling confidence interval, the K backup data points can be divided into M backup data groups according to the sampling period label, ensuring that the data in each group have relatively similar fusion periods. Data source evaluation is then performed on the M backup data groups to assess the impact of different fusion environments on data quality. Data source evaluation refers to assessing the impact of different data source environments on the final product quality based on the collected data source indicator sequences. This is typically done by calculating the data source correlation degree of each combination in each backup data group, resulting in the M sets of data source correlation degrees corresponding to the M data source indicator sequence groups of the M backup data groups.

[0079] After evaluating the data sources, data meeting the quality standards are filtered out from M backup data groups based on the sampling confidence threshold. Specifically, data whose data source correlation meets the sampling confidence threshold is filtered from each data source correlation set. Only those backup data that meet the sampling confidence threshold requirements in the data source evaluation are ultimately selected as multiple object data, resulting in P object data. By setting the sampling period window, confidence interval, and confidence threshold, the collected data is ensured to have high accuracy and reliability, avoiding the impact of data differences on quality analysis. Data source evaluation and indicator sequence analysis of K backup data reveal the impact of different fusion environments on product quality, ensuring that the objects ultimately used for quality evaluation have sufficient confidence and quality standards, and preventing substandard objects from affecting the overall analysis results.

[0080] In a specific implementation, when performing the operation of "evaluating the M backup data groups based on the K data source indicator sequences, and using the sampling confidence threshold as a filtering environment to filter and obtain the P object data from the M backup data groups", it can be implemented as follows:

[0081] (1) Based on the mapping relationship between the M backup data groups and the K backup data, classify the K data source indicator sequences into M data source indicator sequence groups;

[0082] Specifically, the M backup data groups are formed by grouping the K backup data groups, with each group having similar sampling periods or data source environments. The mapping relationship is the correspondence between the K backup data groups and the M backup data groups. In this process, the K backup data groups are categorized based on similar data source indicators, forming M groups, with each group containing data that shares certain similarities. In other words, the K data source indicator sequences are classified into M data source indicator sequence groups.

[0083] (2) Assign data source criticality values ​​to the data source lines of the Industrial Internet of Things to obtain W entity criticality values ​​of the W data source entities, where W is a positive integer greater than or equal to H;

[0084] In a specific implementation, key values ​​can be assigned to each data source entity in the Industrial Internet of Things (IIoT) data acquisition data line. Based on the duration of each data source entity's impact on the final product quality, a weight value is assigned to each entity, resulting in W entity key values ​​corresponding to W data source entities. The key value of an entity typically depends on its confidence in the final product, such as the acquired temperature or acquisition cycle. W is a positive integer greater than or equal to H, meaning that the number of data source entities in the IIoT data acquisition data line is greater than or equal to the number of located heterogeneous data source entities.

[0085] (3) Obtain the first data source indicator sequence and the second data source indicator sequence from the m-th data source indicator sequence group; m = 1, 2, ... M;

[0086] (4) After calculating the correlation degree between the first data source indicator sequence and the second data source indicator sequence in the W entities of the W data source entities based on the cosine distance, the correlation degree of the W entities is calculated by weighting the W entities based on the keyness of the W entities to obtain the correlation degree of the first data source.

[0087] In a specific implementation, each time, a data source indicator sequence group is randomly or sequentially selected from M data source indicator sequence groups, denoted as the m-th data source indicator sequence group (m = 1, 2, ..., M). Two data source indicator sequences are arbitrarily selected from the selected data source indicator sequence group as the first data source indicator sequence and the second data source indicator sequence. Cosine distance is used to calculate the correlation degree between the first and second data source indicator sequences and each of the W data source entities. After calculating the correlation degree of the W data source entities, a weighted calculation is performed based on the W entity key scores. The correlation degree of each data source entity is multiplied by the key score of each entity, and all products are summed to obtain the correlation degree of the first data source.

[0088] The cosine distance is a statistic that measures the duration of a linear correlation between two variables. Its value ranges from -1 to 1, with values ​​close to 1 indicating a strong positive correlation, close to -1 indicating a strong negative correlation, and close to 0 indicating no correlation. The specific calculation process for the cosine distance is typically as follows: First, calculate the average of the indicator sequences from the first and second data sources. Then, calculate the product of the differences between the two data source indicator sequences and sum them to obtain A. Next, calculate the variance of each data source indicator sequence, and based on the variance, calculate the standard deviation of the two data source indicator sequences and multiply them to obtain B. The ratio of A to B is then the cosine distance.

[0089] (5) Perform combination enumeration and data source correlation calculation on the m-th data source indicator sequence group to obtain the m-th data source correlation set of the m-th data source indicator sequence group;

[0090] In a specific implementation process, based on processes (3) and (4), the m-th data source indicator sequence group can be combined and enumerated to calculate the data source correlation degree of each possible combination, thus forming the first set of data source correlation degrees for the first set of data source indicator sequences. Combination enumeration refers to combining and enumerating the m-th data source indicator sequence group, that is, calculating the correlation degree of all possible data source indicator sequences, and then using the calculated set of data source correlation degrees for the m-th group.

[0091] (6) Based on the set of data source correlation in the mth group, filter the mth backup data group to obtain the mth group of object data whose data source correlation all meet the sampling confidence threshold.

[0092] In a specific implementation, a sampling confidence threshold can be used to filter out a set of object data that meets the criteria from each backup data group. Only objects whose data source correlation reaches or exceeds the preset sampling confidence threshold are obtained. This allows for the acquisition of multiple objects with significant differences in data source indicators within an approximate period, reducing the resource consumption issues caused by fusing samples based on similar data sources for quality evaluation. By calculating and weighting the correlation of indicator sequences from different data sources, the most suitable data source environment can be accurately identified, thereby improving the quality and stability of the data.

[0093] (7) Merge the object data of all groups to obtain the P object data.

[0094] 102. Perform key quality indicator detection on each object's data to obtain the indicator difference set for each object's data;

[0095] In a specific implementation process, P object data points are multiple objects extracted from the finished product stage of the Industrial Internet of Things (IIoT) data source line. These objects are representative and can reflect the data quality of IIoT data across different fusion batches, different cycle points, or different data source environments. Multiple object data points are tested to select appropriate key quality indicators (KPIs). In the process of fusing full-quality IIoT data, KPIs are crucial indicators reflecting product quality.

[0096] Key quality indicators (KPIs) are tested on data from multiple objects to obtain KPI data for each object, forming an object KPI variance set. This variance set reflects the differences between each object and each KPI. The variance set refers to the difference between the actual measured value and the target standard value for each object in terms of KPIs. By comprehensively testing the KPIs of multiple objects, the quality stability during the fusion process can be accurately assessed, and differences can be identified in a timely manner. The KPI variance set for each object's data can include N KPI variance values.

[0097] 103. Based on the industrial IoT fusion logs, analyze the W data source entities to obtain the data source-scenario type matrix;

[0098] In a specific implementation process, comprehensive historical data can be collected from the industrial IoT fusion logs to obtain data from the industrial IoT data source lines, recording the difference information sets of key quality indicators for each of the W data source entities in historical fusion; each difference information set is analyzed to obtain the difference frequency, difference magnitude, and single difference duration of each key quality indicator in each difference information set; the product of the weight of the difference magnitude of the key quality indicator and the single difference duration of the key quality indicator is calculated as the weighted total duration of the key quality indicator; the ratio of the weighted total duration to the difference frequency of the key quality indicator is used as the confidence duration of the key quality indicator; a data source-scenario type matrix is ​​constructed with behavioral data source entities, columns representing key quality indicators, and elements representing the confidence duration of key quality indicators.

[0099] Specifically, the Industrial Internet of Things (IIoT) convergence log records information such as operations, equipment status, environment, and quality inspection results at each stage of the IIoT convergence process. This includes detailed descriptions, cycles, equipment settings, operators, and quality control data for each data source step. Operation records of each data source entity in the IIoT convergence log are collected, and a data source entity correspondence analysis is performed to reveal the relationship between different data source entities and key quality indicators. Based on the results of the correspondence analysis, a data source-scenario type matrix is ​​constructed. Rows in the matrix represent different data source entities, columns represent different quality indicators, and each element in the matrix represents the duration of an influence of a particular data source entity on a particular quality indicator.

[0100] 104. Based on the index difference set of each object's data, traverse the data source-scene type matrix to perform fusion quality evaluation and locate H heterogeneous data source entities;

[0101] In a specific implementation process, this step can be implemented in the following way:

[0102] (11) For the index difference set of the i-th object data, use the j-th quality index difference to traverse the data source-scenario type matrix to obtain the j-th deviation confidence of each of the W data source entities; i = 1, 2, ..., P; j = 1, 2, ..., N;

[0103] Specifically, each time, a set of indicator differences for an object data point can be selected, denoted as the indicator difference set for the i-th object data point. From this set, a quality indicator difference quantity is selected each time, denoted as the j-th quality indicator difference quantity. Then, the j-th quality indicator difference quantity is used to traverse the data source-scenario type matrix to obtain the j-th deviation confidence level for each of the W data source entities; i = 1, 2, ..., P; j = 1, 2, ..., N. In this way, a total of N data source entity groups can be obtained, each data source entity group having N deviation confidence levels.

[0104] (12) Using the j-th quality index difference, schedule the j-th deviation confidence of each of the W data source entities to obtain the j-th scheduling confidence of each of the W data source entities;

[0105] In a specific implementation, the j-th quality indicator difference can be used to schedule the j-th deviation confidence level of each of the W data source entities, resulting in the j-th scheduling confidence level for each of the W data source entities. In other words, the deviation confidence level of each of the W data source entities is scheduled based on the quality indicator difference. The specific scheduling process can be as follows: Scheduling confidence level = Deviation confidence level × (1 + Indicator difference). This amplifies the influence of data sources corresponding to high-difference indicators, allowing the confidence level to reflect the actual quality fluctuations of the current object.

[0106] (13) The scheduling confidence scores of each of the W data source entities are weighted and summed to obtain the i-th heterogeneous confidence score of each of the W data source entities;

[0107] In a specific implementation, the weighted sum of all scheduling confidence scores for each of the W data source entities can be used to obtain the i-th heterogeneous confidence score for each of the W data source entities. In other words, in a group of N data source entities, each data source entity has a heterogeneous confidence score. By weighted summing the N heterogeneous confidence scores, the fused heterogeneous confidence score of each data entity under the indicator difference set of each object data can be obtained.

[0108] (14) The weighted sum of all heterogeneous confidence scores of each of the W data source entities is obtained to obtain the total heterogeneous confidence score of each of the W data source entities;

[0109] Specifically, in the set of indicator differences for P object data, each data entity has a heterogeneous confidence level. By weighted summing of all heterogeneous confidence levels of each data entity under the set of indicator differences for P object data, we can obtain the total heterogeneous confidence level of each of the W data source entities.

[0110] (15) Select H data source entities whose total heterogeneity confidence is greater than the preset heterogeneity confidence as the H heterogeneous data source entities.

[0111] In a specific implementation, H is less than W, and the H heterogeneous data source entities have H heterogeneous confidence levels. Thus, through detailed heterogeneous analysis and evaluation, it is possible to accurately identify which data source entities have the greatest impact on product quality, thereby enabling targeted scheduling, constructing heterogeneous synchronous records and fused heterogeneous confidence levels, quantifying quality heterogeneity, identifying and locating highly heterogeneous entities in the industrial IoT data sources, and taking corresponding measures to reduce heterogeneity and improve product quality.

[0112] 105. Based on the H heterogeneous confidence levels and the standard procedures for obtaining data sources from the Industrial Internet of Things, the H heterogeneous data source entities are hierarchically classified for scheduling to obtain a multi-level set of entities to be scheduled.

[0113] In a specific implementation process, this step can be achieved in the following way:

[0114] (21) Obtain the H project stable cycles, H data source scheduling resources and H data source scheduling complexity of the H heterogeneous data source entities through interaction;

[0115] In a specific implementation, it can interact with other objects to obtain H project stability periods, H data source scheduling resources, and H data source scheduling complexities for H heterogeneous data source entities. For example, by connecting to the monitoring system's API, it can query the project stability period, data source scheduling resources, and data source scheduling complexity of each data source entity over a past period.

[0116] (22) Obtain the H data source execution positions of the H heterogeneous data source entities from the standard procedure call;

[0117] In a specific implementation process, the standard procedure refers to the standard execution sequence in the Industrial Internet of Things (IIoT) process. It refers to the order in which various data source entities are executed within the fusion process, ensuring collaboration and smooth connection between the steps of each data source. The standard procedure retrieves the execution positions of H heterogeneous data source entities, representing the position of each heterogeneous data source entity within the entire fusion process.

[0118] (23) Pre-construct a scheduling heterogeneous quantization function, and obtain H scheduling heterogeneous quantization values ​​by loading the H project stable cycles, H data source scheduling resources, H data source scheduling complexity, H data source execution position and H heterogeneous confidence into the scheduling heterogeneous quantization function;

[0119] In a specific implementation, the scheduling priority of each data source entity can be calculated by combining its heterogeneity confidence level with its position in the standard process. Simultaneously, considering the project stability period, data source scheduling resources, and data source scheduling complexity of each data source entity, a scheduling heterogeneity quantification value can be calculated for each entity, reflecting the scheduling heterogeneity urgency duration of each entity. A larger scheduling heterogeneity quantification value indicates a higher scheduling priority for that entity.

[0120] Specifically, a scheduling heterogeneity quantification function can be pre-constructed to quantify the degree of heterogeneity among different data source entities during the scheduling process. The purpose of the scheduling heterogeneity quantification function is to comprehensively consider multiple factors and measure the scheduling urgency of each data source entity, thereby providing a basis for data source scheduling decisions.

[0121] H project stability periods, H data source scheduling resources, H data source scheduling complexities, H data source execution positions, and H heterogeneous confidence levels can be loaded into the scheduling heterogeneous quantization function to obtain H scheduling heterogeneous quantization values. For each heterogeneous data source entity, its corresponding project stability period, scheduling resources, scheduling complexity, execution position, and heterogeneous confidence level are substituted into the scheduling heterogeneous quantization function to calculate the scheduling heterogeneous quantization value for each heterogeneous data source entity.

[0122] The scheduling heterogeneous quantization function is as follows:

[0123]

[0124] Where F is the scheduling heterogeneity quantization value, T is the engineering steady period, C is the data source scheduling resource, O is the data source scheduling complexity, S is the data source execution position, and R is the heterogeneity confidence.

[0125] In a specific implementation, F represents the scheduling heterogeneity quantification value, measuring the degree of heterogeneity of a data source entity during scheduling. A higher F value indicates greater scheduling heterogeneity and higher scheduling urgency for the data source entity. Typically, an F value close to 0 indicates a lower scheduling priority for the entity, while a value close to 1 indicates a higher scheduling priority. T represents the engineering stability period, referring to the period during which the data source entity can maintain stable quality, usually measured in hours or minutes. A longer engineering stability period means the entity is better able to maintain consistent and stable quality.

[0126] C represents the data source scheduling resources, which are the costs or resources required to schedule a data source entity, including equipment upgrades, raw materials, technical support, and human resources.

[0127] O represents the data source scheduling complexity, which measures the complexity and duration of the scheduling process. It is related to factors such as the technical difficulty of the data source entity, equipment requirements, and the operational experience required.

[0128] S represents the execution position of the data source, which is the execution order of the data source entities during the fusion process. Different data source entities will affect the operation of subsequent entities, so their execution position will affect the urgency of scheduling.

[0129] R represents the heterogeneity confidence level, which is the potential heterogeneity duration of the data source entity. The larger the value, the greater the impact of the entity on product quality, and the higher its scheduling priority.

[0130] T·O·e -R This represents the relationship between the stability, complexity, and heterogeneity of data source entity scheduling. A larger product of the project's stable period T and the data source scheduling complexity O indicates more stable and difficult scheduling, as expressed by the exponential function e. -R To adjust the heterogeneity confidence level R, representing the impact of heterogeneity on scheduling urgency duration, if R is large, e -R If R is smaller, the urgency of scheduling increases; if R is smaller, the urgency of scheduling decreases. C·S·(In(S)) 2 This represents the suppressive effect of data source scheduling resources, execution position, and the nonlinear influence of execution position on scheduling. The larger the scheduling resource C, the greater the scheduling pressure; the execution position S affects the criticality of the data source entity, with earlier entities typically having a greater impact; and the square of the natural logarithm (In(S))... 2 This indicates the non-linear impact of the execution position S on scheduling—the earlier the entity is executed, the greater the impact on scheduling. By calculating the scheduling heterogeneity quantization value for each data source entity, data source entities with higher scheduling heterogeneity quantization values ​​are prioritized, thereby achieving higher fusion efficiency and product quality.

[0131] (24) Preset scheduling grading scale, and perform scheduling grading on the H heterogeneous data source entities according to the scheduling grading scale and the H scheduling heterogeneous quantification values ​​to obtain a multi-level set of entities to be scheduled.

[0132] In a specific implementation process, the scheduling hierarchy scale is a preset standard used to classify and sort different data source entities based on scheduling heterogeneity quantization values. The scheduling hierarchy scale determines the urgency and priority of scheduling different data source entities, typically divided into several levels, such as high heterogeneity, medium heterogeneity, and low heterogeneity. The range of scheduling heterogeneity quantization values ​​is set according to specific fusion requirements. The calculated scheduling heterogeneity quantization value for each data source entity is compared with the preset hierarchy scale to determine the scheduling level of each entity. Based on the scheduling heterogeneity quantization value and its corresponding scheduling level for each data source entity, all entities are divided into different groups of entities to be scheduled. By combining the scheduling heterogeneity quantization value with the preset scheduling hierarchy scale, the scheduling level to which each data source entity belongs is determined. Typically, the hierarchy scale includes different heterogeneity thresholds, dividing data source entities into multiple scheduling levels.

[0133] By tiering scheduling priorities, each data source entity is assigned a scheduling priority, ultimately determining which entities require scheduling. Highly heterogeneous entities should be scheduled first to reduce potential heterogeneity during the fusion process; moderately heterogeneous entities can have their scheduling appropriately delayed; and low-heterogeneous entities can have their scheduling postponed if insufficient resources are available. By considering the engineering stability cycle, scheduling resources, complexity, and heterogeneity confidence of each data source entity, the system accurately assesses and prioritizes scheduling those entities that have the greatest impact on product quality. Based on the scheduling tier scale and the scheduling heterogeneity quantification value, unnecessary resources and complexity in the data source scheduling process are reduced, which not only improves product quality but also reduces uncertainty during the fusion process while ensuring quality.

[0134] 106. Using the multi-level set of entities to be scheduled as the scheduling direction, perform multi-level progressive scheduling on the H heterogeneous data source entities until the heterogeneous data source entities in the current level set of entities to be scheduled are scheduled, and the difference value of the industrial IoT full quality data in the key quality indicators is less than the preset difference value, then stop scheduling.

[0135] In a specific implementation process, this step can be implemented in the following way:

[0136] (31) Initialize scheduling hierarchy: Obtain a multi-level set of entities to be scheduled, starting from the highest priority first-level set of entities to be scheduled;

[0137] (33) Execute the current level scheduling: Based on the execution position of H data sources, optimize the heterogeneous data source entities to be scheduled in the current level to be scheduled entity set one by one, and use the preset scheduling algorithm to determine the optimal data source index of each heterogeneous data source entity to be scheduled, aggregate all the optimal data source indexes, and obtain the control index of the current level to be scheduled entity set.

[0138] (33) Scheduling effect verification: Based on the control indicators of all scheduled entity sets, and according to the control indicators of the current level entity set to be scheduled, after scheduling the data source indicators of the industrial IoT data source system, check whether the difference between the industrial IoT full quality data after scheduling and the key quality indicators is less than the preset difference value.

[0139] (34) Decision on the next scheduling step: If the difference between the scheduled industrial IoT full quality data and the key quality indicators is less than the preset difference value, the scheduling is stopped; if the difference between the scheduled industrial IoT full quality data and the key quality indicators is not less than the preset difference value, the scheduling of the next level of the set of entities to be scheduled is entered.

[0140] The specific processes of (31)-(34) above are as follows:

[0141] Obtain a first-level set of entities to be scheduled; perform data source scheduling on each of the first-level set of entities to be scheduled according to the execution position of the H data sources to obtain control indicators for the first-level set of entities to be scheduled; if there are discrepancies in the key quality indicators of the industrial IoT data obtained after scheduling the data source indicators of the industrial IoT data source line based on the control indicators of the first-level set of entities to be scheduled, then obtain a second-level set of entities to be scheduled from the multi-level set of entities to be scheduled; use the control indicators of the first-level set of entities to be scheduled as the scheduling benchmark (i.e., use the control indicators of all scheduled entity sets as the scheduling benchmark) to perform data source scheduling on each of the second-level set of entities to be scheduled to obtain control indicators for the second-level set of entities to be scheduled; if there are discrepancies in the key quality indicators of the industrial IoT data obtained after scheduling the data source indicators of the industrial IoT data source line based on the control indicators of the first-level set of entities to be scheduled, then obtain a second-level set of entities to be scheduled from the multi-level set of entities to be scheduled; if there are discrepancies in the key quality indicators of the industrial IoT data obtained after scheduling the data source indicators of the industrial IoT data source line based on the control indicators of the first-level set of entities to be scheduled, then perform data source scheduling on each of the second-level set of entities to be scheduled according to the control indicators of the first-level set of entities to be scheduled, then perform data source scheduling on each of the second-level set of entities to be scheduled; if there are discrepancies in the key quality indicators of the industrial IoT data obtained after scheduling the data source indicators of the industrial IoT data source line based on the control indicators of the first-level set of entities to be scheduled, then perform data source scheduling on each of the second-level set of entities to be scheduled, ... If the overall quality data of the Industrial Internet of Things (IIoT) after scheduling the data source indicators of the IIoT data source line based on the control indicators of the first-level set of entities to be scheduled has differences in the key quality indicators, then a third-level scheduling entity is obtained from the multi-level set of entities to be scheduled. Conversely, if the differences in the key quality indicators of the overall quality data of the IIoT data source line after scheduling the data source indicators based on the control indicators of the first-level set of entities to be scheduled and the control indicators of the second-level set of entities to be scheduled are zero, then the progressive scheduling stops. This process continues, using the multi-level set of entities to be scheduled as the scheduling direction, to perform multi-level progressive scheduling of the IIoT data source for the H heterogeneous data source entities until the differences in the key quality indicators of the overall quality data of the IIoT after scheduling the data source indicators are zero.

[0142] Specifically, the multi-level set of entities to be scheduled is a collection of data source entities obtained based on heterogeneous evaluation and scheduling hierarchy results. These entities have different scheduling priorities and are divided into multiple levels (e.g., Level 1, Level 2, Level 3, etc.), which are scheduled sequentially. The first-level set of entities to be scheduled is obtained from the multi-level set. Based on the execution order of H data sources, the first-level set is scheduled one by one, including adjusting data source indicators, improving operational processes, and replacing equipment or materials. Data source scheduling typically requires the use of appropriate scheduling algorithms to select the most suitable data source indicators, including particle swarm optimization (PSO), genetic algorithms, and simulated annealing algorithms. Taking PSO as an example, the particle swarm is initialized, and the positions and velocities of the particles are set, usually randomly generated. A scheduling objective is defined, and each particle is scheduled based on its own optimal experience and the optimal experience of the swarm. This scheduling continues until a stopping criterion is met (e.g., reaching the maximum number of iterations or convergence of the objective function).

[0143] After data source scheduling, control indicators for the first-level set of entities to be scheduled are determined. These indicators are used to schedule the data source lines from the industrial IoT to improve product quality. Based on the control indicators of the first-level set of entities to be scheduled, the data source lines from the industrial IoT to be scheduled are then used for quality inspection of the scheduled industrial IoT data, checking for differences in key quality indicators.

[0144] If discrepancies persist, it indicates that the first-level scheduling failed to completely eliminate all quality differences. The next step is to select a second-level set of entities to be scheduled and continue scheduling. If discrepancies still exist after the first-level scheduling, the control indicators of the first-level set of entities to be scheduled are used as the scheduling benchmark. Based on this, data source scheduling is performed on each entity in the second-level scheduling set to obtain the control indicators of the second set of entities to be scheduled. The data source indicators of the industrial IoT data source lines are then scheduled again, and the quality of the scheduled industrial IoT full-quality data is checked again to examine the differences in key quality indicators. If discrepancies still exist, a third-level set of entities to be scheduled is obtained from the multi-level set of entities to be scheduled and scheduling continues.

[0145] Conversely, if all indicators meet the standards, and the differences in key quality indicators of the industrial IoT full-quality data after data source indicator scheduling reach zero, the progressive scheduling process stops. This process continues, using a multi-level set of entities to be scheduled as the scheduling direction, continuously scheduling the data source indicators of the industrial IoT data source lines. Through layer-by-layer progressive scheduling, it ensures that the quality differences of each data source entity eventually reach zero. In essence, it involves the gradual adjustment and scheduling of each data source link to ultimately achieve the ideal quality standard. Multi-level progressive scheduling refers to adjusting and scheduling each link in a data source in a step-by-step, progressive manner until the expected goal or standard is achieved. In the industrial IoT data source, the aim is to eliminate quality differences by gradually adjusting key data source entities, ensuring that the final product meets quality standards. Each level of scheduling is based on the results of the previous level, therefore the scheduling process is progressive. Through layer-by-layer scheduling, the quality differences of the industrial IoT full-quality data are gradually reduced, ensuring the optimal operating environment for each data source entity. By scheduling each data source entity, it ultimately ensures that the industrial IoT full-quality data has no differences in key quality indicators, improving the product pass rate.

[0146] This embodiment presents a multi-source heterogeneous data fusion and scheduling method based on an industrial internet data platform. It performs full-quality sampling of industrial IoT data to obtain multiple object data for key quality indicator detection, resulting in an object indicator difference set. Data source entity correspondence analysis yields a data source-scenario type matrix, which is then traversed for fusion quality evaluation to locate heterogeneous data source entities. Standard processes for the data source lines are obtained to hierarchically classify and schedule heterogeneous data source entities at multiple levels. Multi-level progressive scheduling is performed on the heterogeneous data source entities, and data source indicator scheduling is conducted for the industrial IoT data source lines. This achieves efficient multi-source heterogeneous data fusion, solves the problem of large data differences in the industrial internet environment, fully explores the potential value of data, and allows for reasonable task allocation, improving data processing timeliness and resource utilization, and reducing industrial production costs.

[0147] Based on the same general inventive concept, this invention also protects a multi-source heterogeneous data fusion and scheduling system based on an industrial internet data platform. The multi-source heterogeneous data fusion and scheduling system based on an industrial internet data platform provided by this invention will be described below. The multi-source heterogeneous data fusion and scheduling system based on an industrial internet data platform described below can be referred to in correspondence with the multi-source heterogeneous data fusion and scheduling method based on an industrial internet data platform described above.

[0148] Figure 2 This is a schematic diagram of the structure of a multi-source heterogeneous data fusion and scheduling system based on an industrial internet data platform provided in an embodiment of the present invention, as shown below. Figure 3 As shown, the multi-source heterogeneous data fusion and scheduling system based on the industrial internet data platform in this embodiment includes a sampling module 21, a detection module 22, an analysis module 23, a positioning module 24, a hierarchical module 25, and a scheduling module 26.

[0149] The sampling module 21 is used to perform full-quality data sampling of the Industrial Internet of Things (IIoT) during the finished product stage of the IIoT data source line to obtain P object data; wherein the IIoT data source line includes W data source entities.

[0150] The detection module 22 is used to detect key quality indicators for each object data and obtain the indicator difference set for each object data.

[0151] Analysis module 23 is used to analyze the W data source entities based on industrial IoT fusion logs to obtain a data source-scenario type matrix;

[0152] The positioning module 24 is used to traverse the data source-scene type matrix based on the indicator difference set of each object data to perform fusion quality evaluation and locate H heterogeneous data source entities; wherein, H is less than W, and the H heterogeneous data source entities have H heterogeneous confidence levels.

[0153] The hierarchical module 25 is used to hierarchically classify the H heterogeneous data source entities according to the standard procedures of the H heterogeneous confidence levels and the data source lines obtained by the Industrial Internet of Things, and obtain a multi-level set of entities to be scheduled.

[0154] The scheduling module 26 is used to take the multi-level set of entities to be scheduled as the scheduling direction and perform multi-level progressive scheduling on the H heterogeneous data source entities until the heterogeneous data source entities in the current level set of entities to be scheduled are scheduled, and the difference value of the industrial IoT full quality data in the key quality indicators is less than the preset difference value, and then the scheduling stops.

[0155] Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. The multi-source heterogeneous data fusion and scheduling system based on an industrial internet data platform may include: a processor 310, a communication interface 320, a memory 330, and a communication bus 340. The processor 310, communication interface 320, and memory 330 communicate with each other via the communication bus 340. The processor 310 can call logical instructions in the memory 330 to execute the multi-source heterogeneous data fusion and scheduling method based on the industrial internet data platform.

[0156] Furthermore, the logical instructions in the aforementioned memory 330 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0157] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the multi-source heterogeneous data fusion and scheduling method based on the industrial internet data platform provided by the above methods.

[0158] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, is implemented to perform the multi-source heterogeneous data fusion and scheduling method based on the industrial internet data platform provided by the above methods.

[0159] It should be noted that all relevant information that may be involved in the various embodiments of this application is processed in strict accordance with the requirements of laws and regulations, following the principles of legality, legitimacy, and necessity, based on the reasonable purpose of the business scenario, and is information that users actively provide or that is generated as a result of using the product / service, as well as information that is obtained with the user's authorization.

[0160] The information processed in this application may vary depending on the specific product / service scenario and should be based on the specific scenario in which the user uses the product / service. It may involve the user's account information, device information, or other related information. This application will treat the relevant information and its processing with the utmost diligence.

[0161] This application attaches great importance to the security of relevant information and has taken reasonable and feasible security protection measures that comply with industry standards to protect the relevant information and prevent it from being accessed, disclosed, used, modified, damaged or lost without authorization.

[0162] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0163] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0164] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention 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; and these 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 the present invention.

Claims

1. A method for multi-source heterogeneous data fusion and scheduling based on an industrial internet data platform, characterized in that, include: In the finished product stage of obtaining data sources for the Industrial Internet of Things (IIoT), full-quality data sampling of the IIoT is performed to obtain P object data; wherein, the IIoT data source line includes W data source entities; Perform key quality indicator detection on each object's data to obtain a set of indicator differences for each object's data; Based on the industrial IoT fusion logs, the W data source entities are analyzed to obtain a data source-scenario type matrix; Based on the index difference set of each object's data, the data source-scene type matrix is ​​traversed to perform fusion quality evaluation, and H heterogeneous data source entities are located; where H is less than W, and the H heterogeneous data source entities have H heterogeneous confidence levels; Based on the H heterogeneous confidence levels and the standard procedures for obtaining data sources from the Industrial Internet of Things, the H heterogeneous data source entities are hierarchically classified for scheduling to obtain a multi-level set of entities to be scheduled. Using the multi-level set of entities to be scheduled as the scheduling direction, multi-level progressive scheduling is performed on the H heterogeneous data source entities until the heterogeneous data source entities in the current level set of entities to be scheduled are scheduled, and the difference value of the industrial IoT full quality data in the key quality indicators is less than the preset difference value, then the scheduling stops.

2. The method for multi-source heterogeneous data fusion and scheduling based on an industrial internet data platform according to claim 1, characterized in that, Each object's data set includes N quality indicator differences; Based on the index difference set of each object's data, the data source-scene type matrix is ​​traversed to perform a fusion quality evaluation, locating H heterogeneous data source entities, including: For the index difference set of the i-th object data, the j-th quality index difference is used to traverse the data source-scenario type matrix to obtain the j-th deviation confidence score of each of the W data source entities; i = 1, 2, ..., P; j = 1, 2, ..., N; The j-th quality index difference is used to schedule the j-th deviation confidence of each of the W data source entities, thereby obtaining the j-th scheduling confidence of each of the W data source entities; The weighted sum of all scheduling confidence scores of each of the W data source entities is used to obtain the i-th heterogeneous confidence score of each of the W data source entities; The total heterogeneous confidence scores of each of the W data source entities are obtained by weighted summation of all their heterogeneous confidence scores. H data source entities with a total heterogeneity confidence level greater than a preset heterogeneity confidence level are selected as the H heterogeneous data source entities.

3. The method for multi-source heterogeneous data fusion and scheduling based on an industrial internet data platform according to claim 1, characterized in that, Based on the industrial IoT fusion logs, the W data source entities are analyzed to obtain a data source-scenario type matrix, including: A comprehensive historical data collection is performed on the data sources obtained from the Industrial Internet of Things (IIoT) fusion logs, recording the differences in key quality indicators of the W data source entities during historical fusion. Analyze each set of discrepancies to obtain the frequency of discrepancies for each key quality indicator, the magnitude of discrepancies for each key quality indicator, and the duration of a single discrepancy for each key quality indicator in each set of discrepancies. The weighted total duration of the key quality indicator is calculated by multiplying the weight of the difference magnitude of the key quality indicator with the duration of a single difference of the key quality indicator. The ratio of the weighted total duration to the frequency of differences in the key quality indicators is taken as the confidence duration of the key quality indicators. Construct a data source-scenario type matrix with behavioral data source entities, columns as key quality indicators, and elements as confidence durations of key quality indicators.

4. The method for multi-source heterogeneous data fusion and scheduling based on an industrial internet data platform according to claim 1, characterized in that, Based on the H heterogeneous confidence levels and the standard procedures for obtaining data sources from the Industrial Internet of Things (IIoT), the H heterogeneous data source entities are hierarchically scheduled to obtain a multi-level set of entities to be scheduled, including: The H project steady-state cycles, H data source scheduling resources, and H data source scheduling complexity of the H heterogeneous data source entities are obtained through interaction. The execution positions of the H heterogeneous data source entities are obtained from the standard procedure call; A scheduling heterogeneous quantization function is pre-constructed, and H scheduling heterogeneous quantization values ​​are obtained by loading the H project steady-state cycles, H data source scheduling resources, H data source scheduling complexity, H data source execution order and H heterogeneous confidence into the scheduling heterogeneous quantization function. A preset scheduling hierarchy scale is established, and the H heterogeneous data source entities are classified into scheduling levels according to the scheduling hierarchy scale and the H scheduling heterogeneous quantification values ​​to obtain a multi-level set of entities to be scheduled.

5. The method for multi-source heterogeneous data fusion and scheduling based on an industrial internet data platform according to claim 4, characterized in that, The scheduling heterogeneous quantization function is: Where F is the scheduling heterogeneity quantization value, T is the engineering steady period, C is the data source scheduling resource, O is the data source scheduling complexity, S is the data source execution position, and R is the heterogeneity confidence.

6. The method for multi-source heterogeneous data fusion and scheduling based on an industrial internet data platform according to claim 1, characterized in that, Using the multi-level set of entities to be scheduled as the scheduling direction, multi-level progressive scheduling is performed on the H heterogeneous data source entities until the heterogeneous data source entities in the current level set of entities to be scheduled are scheduled. Then, if the difference value of the key quality indicators in the industrial IoT full-quality data is less than a preset difference value, scheduling stops, including: Initialize scheduling hierarchy: Obtain a multi-level set of entities to be scheduled, starting with the highest priority first-level set of entities to be scheduled; Execute the current level scheduling: Based on the execution position of H data sources, optimize the heterogeneous data source entities to be scheduled in the current level's entity set one by one, and use a preset scheduling algorithm to determine the optimal data source index for each heterogeneous data source entity to be scheduled. Aggregate all the optimal data source indexes to obtain the control index of the current level's entity set to be scheduled. Scheduling effect verification: Based on the control indicators of all scheduled entity sets, and according to the control indicators of the entity set to be scheduled at the current level, after scheduling the data source indicators of the industrial IoT data source system, check whether the difference between the industrial IoT full quality data and the key quality indicators after scheduling is less than the preset difference value. Decision on the next scheduling step: If the difference between the key quality indicators of the scheduled industrial IoT full quality data is less than the preset difference value, the scheduling stops; if the difference between the key quality indicators of the scheduled industrial IoT full quality data is not less than the preset difference value, the scheduling proceeds to the next level of the set of entities to be scheduled.

7. The method for multi-source heterogeneous data fusion and scheduling based on an industrial internet data platform according to any one of claims 1-6, characterized in that, Perform full-quality data sampling for the Industrial Internet of Things (IIoT) to obtain data from multiple objects, including: Preset sampling period window, sampling confidence interval, and sampling confidence threshold; Within the sampling period window, full-quality data of the Industrial Internet of Things (IIoT) is sampled from the finished product stage of the data source line obtained by the IIoT to obtain K backup data, wherein the K backup data have K sampling period labels. The K backup data sources are evaluated by fusion to obtain a sequence of K data source indicators; Using the sampling confidence interval as a grouping constraint, the K spare data are classified into M spare data groups according to the K sampling period labels, where M is a positive integer less than K; After evaluating the M backup data groups based on the K data source indicator sequences, the sampling confidence threshold is used as the filtering environment to filter and obtain the P object data from the M backup data groups.

8. The method for multi-source heterogeneous data fusion and scheduling based on an industrial internet data platform according to claim 7, characterized in that, After evaluating the M backup data groups based on the K data source indicator sequences, the sampling confidence threshold is used as the filtering environment to filter and obtain the P object data from the M backup data groups, including: Based on the mapping relationship between the M backup data groups and the K backup data, the K data source indicator sequences are classified into M data source indicator sequence groups; The data source key is assigned to the data source line of the Industrial Internet of Things to obtain the W entity keyness of the W data source entities, where W is a positive integer greater than or equal to H; Obtain the first data source indicator sequence and the second data source indicator sequence from the m-th data source indicator sequence group; m = 1, 2, ..., M; After calculating the correlation degree between the first data source indicator sequence and the second data source indicator sequence in the W entities of the W data source entities based on the cosine distance, the correlation degree between the W entities is calculated by weighting the W entities based on their criticality to obtain the correlation degree of the first data source. The m-th data source indicator sequence group is combined and enumerated, and the data source correlation degree is calculated to obtain the m-th data source correlation degree set of the m-th data source indicator sequence group; Based on the set of data source correlation in the mth group, filter the mth backup data group to obtain the mth group of object data in which the correlation of all data sources satisfies the sampling confidence threshold. The object data of all groups are merged to obtain the P object data.

9. A multi-source heterogeneous data fusion and scheduling system based on an industrial internet data platform, characterized in that, include: The sampling module is used to perform full-quality data sampling of the Industrial Internet of Things (IIoT) during the finished product stage of the IIoT data source line to obtain P object data; wherein the IIoT data source line includes W data source entities; The detection module is used to detect key quality indicators for each object's data and obtain a set of indicator differences for each object's data. The analysis module is used to analyze the W data source entities based on the industrial IoT fusion logs to obtain a data source-scenario type matrix; The positioning module is used to traverse the data source-scene type matrix based on the indicator difference set of each object data to perform fusion quality evaluation and locate H heterogeneous data source entities; wherein, H is less than W, and the H heterogeneous data source entities have H heterogeneous confidence levels. The hierarchical module is used to hierarchically classify the H heterogeneous data source entities according to the standard procedures of the H heterogeneous confidence levels and the data source lines obtained by the Industrial Internet of Things, and obtain a multi-level set of entities to be scheduled. The scheduling module is used to take the multi-level set of entities to be scheduled as the scheduling direction and perform multi-level progressive scheduling on the H heterogeneous data source entities until the heterogeneous data source entities in the current level set of entities to be scheduled are scheduled, and the difference value of the industrial IoT full quality data in the key quality indicators is less than the preset difference value, and then the scheduling stops.

10. An electronic device, characterized in that, It includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the multi-source heterogeneous data fusion and scheduling method based on an industrial internet data platform as described in any one of claims 1-8.

Images