Industrial edge intelligent decision-making all-in-one machine data analysis system

Abstract

The invention discloses a data analysis system of an industrial edge intelligent decision all-in-one machine, and relates to the technical field of data analysis. The method comprises the steps of analyzing a global data influence rate of data based on real-time data and historical data through a local data analysis module, and performing data screening to divide data of a single node into enhanced data and weakened data; the conversion processing module obtains lightweight data based on conversion processing; and the view angle compensation module performs global optimization compensation on data of a single node by aggregating data of multiple edge nodes. According to the invention, by aggregating multi-edge node data and carrying out global optimization analysis, the real-time performance of edge calculation is kept, local data limitation is broken through, an industrial intelligent decision-making system of local agility and global optimization is realized, and the problem of data islands of an industrial edge intelligent decision-making all-in-one machine is effectively solved.

Description

Technical Field

[0001] The present invention belongs to the field of data analysis technology, and in particular relates to an industrial edge intelligent decision-making all-in-one data analysis system. Background Art

[0002] The Industrial Edge Intelligent Decision-Making All-in-One Machine is an industrial-grade device that integrates edge computing and artificial intelligence technologies, and is designed for real-time data processing and intelligent decision-making in industrial scenarios.

[0003] Key features include: edge computing: processing data near the source, reducing data transmission latency; AI intelligent analysis: built-in machine learning algorithms support real-time data analysis and decision-making; industrial-grade design: rugged and durable, suitable for harsh industrial environments; multi-protocol support: compatible with a variety of industrial communication protocols and interfaces; low-latency response: millisecond-level response speeds, meeting industrial real-time requirements. While industrial edge intelligent decision-making all-in-one machines offer significant advantages in industrial scenarios, their data analysis capabilities still have some inherent flaws and limitations.

[0004] Edge devices typically lack the computing power (CPU / GPU / TPU) of cloud servers, making it difficult to deploy and run large-scale deep learning models (such as large language models and high-precision visual inspection models). This can lead to reduced model accuracy or limited execution of lightweight models (such as TinyML), sacrificing some analytical capabilities. Furthermore, a single device can only capture local data (such as a single production line or piece of equipment), lacking a global perspective, which can affect comprehensive decision-making. For example, optimizing a device's own energy consumption may negatively impact overall production line efficiency.

[0005] For example, Chinese patent CN117014440A provides an edge intelligent decision-making method for the Industrial Internet. This method includes: calculating the cosine distance between each data body received by the edge gateway; constructing a matrix of the cosine distances between each data body, calculating the credibility of each data body; calculating the uncertainty of each data body, fusing the credibility and uncertainty of the data body, calculating the average weight of all data bodies, and using the average weight as the decision result of the edge gateway; using the DS algorithm to fuse the decision results of the edge gateway N-1 times to obtain the final decision result of the edge gateway. For example, Chinese patent CN119603297A discloses a real-time data analysis and decision-making system based on the edge intelligent Internet of Things. The system makes decisions based on analysis results and a decision generation mechanism, generates decision optimization using an adaptive decision optimization algorithm, and exchanges data with a central server layer through a cloud-edge collaboration module. The central server layer is used to update the model based on the actual situation and analysis results corresponding to the analyzed and processed data uploaded by each edge computing node, and returns the updated model parameters to the edge computing node layer.

[0006] However, the industrial edge intelligent decision-making all-in-one machine has the problem of "data islands", that is, a single device can only obtain local data and lacks a global perspective. Now a solution is proposed. Summary of the Invention

[0007] The purpose of the present invention is to provide an industrial edge intelligent decision-making all-in-one data analysis system, which solves the existing problems by analyzing real-time data and historical data, screening the data of a single node into strengthened data and weakened data, and performing lightweight processing on the data.

[0008] To solve the above technical problems, the present invention is achieved through the following technical solutions:

[0009] The present invention is an industrial edge intelligent decision-making all-in-one machine data analysis system, comprising:

[0010] The local data analysis module analyzes the global data impact rate of data based on real-time data and historical data, and performs data screening to divide the data of a single node into strengthened data and weakened data;

[0011] A conversion processing module, which obtains lightweight data based on conversion processing;

[0012] Lightweight data includes data item-number-feature data mapping table, data item-feature value + conversion value mapping table;

[0013] The perspective compensation module is used to perform global optimization compensation on the data of a single node by aggregating the data of multiple edge nodes.

[0014] Furthermore, when the local data analysis module performs data screening, the following steps are performed:

[0015] Step S001: Obtain real-time data and historical data of any node;

[0016] Step S002: Analyze historical data and extract parameters and corresponding data items whose impact on global data exceeds a preset value X1 from the historical data;

[0017] Step S003: extracting parameters corresponding to key data items in real-time data according to key data items preset by the administrator;

[0018] Step S004: Mark the parameters and corresponding data items extracted in steps S002 and S003 as enhanced data, and mark the parameters and corresponding data items in the real-time data other than the enhanced data as weakened data.

[0019] Furthermore, the global data impact rate is:

[0020] Obtain any parameter of any node from historical data, and mark the obtained parameter as a simulation parameter, and mark the data item corresponding to the simulation parameter as a simulation data item;

[0021] Obtain the products and production lines corresponding to the simulation parameters, extract the real-time production capacity of the production line corresponding to the product output, and mark it as simulated capacity;

[0022] Obtain all parameters corresponding to the simulation data items from the historical data and form a parameter set C: {C1, C2, C3, ..., Cn}, where n is a positive integer, indicating that the parameter set contains n parameters;

[0023] Calculate the production line capacity corresponding to all parameter elements in the parameter set, and mark the average value of the obtained capacity as the average capacity;

[0024] The variance corresponding to the data set formed by the simulation parameters and all parameter elements in the parameter set is marked as S1, and the absolute value of the difference between the averaged capacity and the simulated capacity is marked as S2;

[0025] If S1 ≥ X2 and S2 ≥ X3, the global data influence rate corresponding to the simulation parameter is marked as Y1;

[0026] If S1<X2, and S2<X3, the global data influence rate corresponding to the simulation parameter is marked as Y2;

[0027] If S1<X2, and S2≥X3, the global data influence rate corresponding to the simulation parameter is marked as Y3;

[0028] If S1≥X2, and S2<X3, the global data influence rate corresponding to the simulation parameter is marked as Y4;

[0029] Among them, X2, X3, Y1, Y2, Y3, and Y4 are preset values; and Y1>Y2>X1>Y3>Y4.

[0030] Furthermore, the global data impact rate is:

[0031] Step A01: Obtain any data item of any node from historical data and mark it as a simulation data item;

[0032] Step A02: Obtain all parameters corresponding to the simulation data item, and extract the maximum value from the obtained parameters and mark it as the simulation parameter;

[0033] Step A03: Obtain the product and production line corresponding to the simulation parameters, extract the real-time production capacity of the production line corresponding to the product output, and mark it as the simulated capacity;

[0034] Step A04: Calculate the production capacity and the average of the production capacity corresponding to all the parameters obtained in step A02, and mark the average of the production capacity as the comparison capacity;

[0035] Step A05: Among them, the benchmark difference is a preset value.

[0036] Furthermore, when the conversion processing module performs the conversion processing, the following steps are performed:

[0037] Obtain the real-time strengthening and weakening data corresponding to a single node, as well as the number of strengthening and weakening data items;

[0038] Generate a dynamic numbering algorithm based on the number of strengthening data and weakening data, perform mapping processing on the strengthening data and weakening data, and obtain the data item number and the corresponding data item-number mapping table;

[0039] Retrieve the parameter corresponding to the data item number, convert the parameter into a characteristic data of a rectangle, and form a data item-number-characteristic data mapping table;

[0040] For parameters that do not have data item numbers, the parameters are converted into characteristic values + conversion values to form a data item-characteristic value + conversion value mapping table;

[0041] When data is transmitted, lightweight data represented by the mapping table is transmitted.

[0042] Furthermore, the dynamic numbering algorithm is:

[0043] If the number of enhanced data items is greater than the number of weakened data items, the enhanced data items are numbered in sequence according to the transmission time of the corresponding parameters;

[0044] If the number of strengthened data items is less than or equal to the number of weakened data items, the weakened data items are numbered in sequence according to the transmission time of the corresponding parameters;

[0045] When numbering, the starting number is: number of strengthened data items - number of weakened data items.

[0046] Furthermore, the data item-number-feature data mapping table is:

[0047] Retrieve the specific value of the parameter corresponding to the data item number and mark it as the value to be mapped;

[0048] If the data item number is an odd number, the number in the tens place of the value to be mapped + 1 is used as the width K of the rectangle, and the value obtained by dividing the value to be mapped by the width is used as the length C of the rectangle;

[0049] If the data item number is an even number, the number in the tens place of the value to be mapped + 1 is used as the length C of the rectangle, and the value obtained by dividing the value to be mapped by the width is used as the width K of the rectangle;

[0050] Form a rectangle with a width of K, a length of C, and an area equal to the value to be mapped;

[0051] Convert the value to be mapped into the area value of the rectangle to form a data item-number-feature data mapping table.

[0052] Furthermore, the data item-feature value+conversion value mapping table is:

[0053] Get the specific value corresponding to the parameter with no data item number, marked as initial value CZi, i = 1, 2, 3, ..., m, m is a positive integer;

[0054] Retrieve the initial value CZ1 and the time when it was generated, obtain a parameter with a data item number that is closest to the time when it was generated, and mark the corresponding data item number as the characteristic value;

[0055] If the eigenvalue is an odd number, a new parameter XCZ1 is formed with the initial value CZ1 in front and the eigenvalue in the back;

[0056] If the eigenvalue is an even number, a new parameter XCZ1 is formed with the initial value CZ1 at the back and the eigenvalue at the front;

[0057] In addition, i is accumulated by 1 in sequence until all new parameters are obtained;

[0058] All the new parameters obtained are marked as new parameters XCZi to form a data item-feature value+conversion value mapping table.

[0059] Furthermore, when the perspective compensation module performs global optimization compensation by aggregating data of multiple edge nodes, the following steps are performed:

[0060] Select any parameter in a node and the corresponding production line, and mark the parameter and production line as: parameter to be compensated, production line to be compensated;

[0061] Obtain the parameters upstream and downstream of the parameters to be compensated on the production line to be compensated;

[0062] If the upstream and downstream parameters and the parameters to be compensated are all enhanced data, the upstream and downstream parameters are marked as optimized compensation parameters of the parameters to be compensated;

[0063] If the upstream and downstream parameters are all enhanced data and the parameters to be compensated are all weakened data, then the upstream and downstream parameters are marked as optimized compensation parameters of the parameters to be compensated;

[0064] If both upstream and downstream parameters are weakened data and the parameters to be compensated are strengthened data, the upstream parameters are marked as optimized compensation parameters of the parameters to be compensated;

[0065] Otherwise, the mean capacity corresponding to the adjacent upstream nodes in the process is used as the capacity corresponding to the optimized compensation parameter;

[0066] When performing capacity analysis, the capacity corresponding to the parameter to be compensated = 0.62*the capacity corresponding to the parameter to be compensated + 0.38*the capacity corresponding to the optimized compensation parameter.

[0067] Furthermore, the nodes are divided by an administrator, and the nodes include multiple production lines that are adjacent in terms of process.

[0068] The present invention has the following beneficial effects:

[0069] The present invention uses a local data analysis module to analyze the global data impact rate of data based on real-time data and historical data, and performs data screening to divide the data of a single node into strengthened data and weakened data; the conversion processing module obtains lightweight data based on conversion processing; the perspective compensation module performs global optimization compensation on the data of a single node by aggregating data from multiple edge nodes; aggregates data from multiple edge nodes and performs global optimization analysis, while retaining the real-time performance of edge computing, breaking through local data limitations, and realizing an industrial intelligent decision-making system of "local agility + global optimization".

[0070] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing 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.

[0072] Figure 1 This is a schematic diagram of an industrial edge intelligent decision-making integrated machine data analysis system of the present invention;

[0073] Figure 2 Flowchart for data screening for the local data analysis module. DETAILED DESCRIPTION

[0074] In the following description, specific details such as specific system structures and techniques are provided for purposes of illustration rather than limitation to facilitate a thorough understanding of the embodiments of the present application. However, it will be apparent to those skilled in the art that the present application may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to avoid obscuring the description of the present application with unnecessary detail.

[0075] It should be understood that when used in the present specification and the appended claims, the term "comprising" indicates the presence of described features, integers, steps, operations, elements and / or components, but does not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or collections thereof.

[0076] It will also be understood that the term "and / or" used in this specification and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.

[0077] As used in this specification and the appended claims, the term "if" can be interpreted as "when" or "upon" or "in response to determining" or "in response to detecting," depending on the context. Similarly, the phrase "if it is determined" or "if [described condition or event] is detected" can be interpreted as meaning "upon determination" or "in response to determining" or "upon detection of [described condition or event]" or "in response to detecting [described condition or event]," depending on the context.

[0078] In addition, in the description of the present application specification and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish the descriptions and cannot be understood as indicating or implying relative importance.

[0079] References to "one embodiment" or "some embodiments" in this specification mean that a particular feature, structure, or characteristic described in conjunction with that embodiment is included in one or more embodiments of the present application. Thus, phrases such as "in one embodiment," "in some embodiments," "in other embodiments," and "in other embodiments" appearing in various places in this specification do not necessarily refer to the same embodiment, but rather mean "one or more but not all embodiments," unless otherwise specifically emphasized. The terms "including," "comprising," "having," and variations thereof all mean "including but not limited to," unless otherwise specifically emphasized.

[0080] Example 1:

[0081] See also Figure 1-2 As shown, the present invention is an industrial edge intelligent decision-making integrated machine data analysis system, including:

[0082] The local data analysis module analyzes the global data impact rate of the data based on real-time data and historical data, and performs data screening to divide the data of a single node into strengthening data and weakening data; as an embodiment provided by the present invention, preferably, the node is divided by the administrator, and the node includes multiple production lines adjacent to each other in terms of process

[0083] A conversion processing module, which obtains lightweight data based on conversion processing;

[0084] Lightweight data includes data item-number-feature data mapping table, data item-feature value + conversion value mapping table;

[0085] The perspective compensation module is used to perform global optimization compensation on the data of a single node by aggregating the data of multiple edge nodes.

[0086] As an embodiment provided by the present invention, preferably, when the local data analysis module performs data screening, the following steps are performed:

[0087] Step S001: Obtain real-time data and historical data of any node;

[0088] Step S002: Analyze historical data and extract parameters and corresponding data items whose impact on global data exceeds a preset value X1 from the historical data;

[0089] Step S003: extracting parameters corresponding to key data items in real-time data according to key data items preset by the administrator;

[0090] Step S004: Mark the parameters and corresponding data items extracted in steps S002 and S003 as enhanced data, and mark the parameters and corresponding data items in the real-time data other than the enhanced data as weakened data.

[0091] As an embodiment provided by the present invention, preferably, the global data impact rate is:

[0092] Obtain any parameter of any node from the historical data, and mark the obtained parameter as a simulation parameter, and mark the data item corresponding to the simulation parameter as a simulation data item; specifically, select any parameter of a node from the historical data, such as the temperature value of a temperature sensor of a device, and record it as a simulation parameter, and its corresponding data record is called a simulation data item;

[0093] Get the product and production line corresponding to the simulation parameters, extract the real-time production capacity of the production line when the product is produced, and mark it as the simulated capacity; find the product batch and production line corresponding to the parameters, record the real-time capacity (such as output per unit time) when the batch is completed, and record it as the simulated capacity;

[0094] All parameters corresponding to the simulation data items are obtained from the historical data, and a parameter set C is formed: {C1, C2, C3, ..., Cn}, where n is a positive integer, indicating that the parameter set contains n parameters; all parameters associated with the simulation data items (such as pressure and speed in the same time period) are extracted from the historical data to form a parameter set;

[0095] Calculate the production line capacity corresponding to all parameter elements in the parameter set and mark the average of the obtained capacity as the average capacity; for each parameter in parameter set C, calculate the corresponding production line capacity and take the average value to obtain the average capacity (reflecting the historical average performance);

[0096] The variance corresponding to the data set formed by the simulation parameters and all parameter elements in the parameter set is marked as S1, and the absolute value of the difference between the averaged capacity and the simulated capacity is marked as S2;

[0097] If S1 ≥ X2 and S2 ≥ X3, the global data influence rate corresponding to the simulation parameter is marked as Y1;

[0098] If S1<X2, and S2<X3, the global data influence rate corresponding to the simulation parameter is marked as Y2;

[0099] If S1<X2, and S2≥X3, the global data influence rate corresponding to the simulation parameter is marked as Y3;

[0100] If S1≥X2, and S2<X3, the global data influence rate corresponding to the simulation parameter is marked as Y4;

[0101] Among them, X2, X3, Y1, Y2, Y3, and Y4 are preset values; and Y1>Y2>X1>Y3>Y4.

[0102] Specific scenarios can be:

[0103] Simulation parameters: robot welding gun current value;

[0104] Analog data items: current records during a batch of welding;

[0105] Parameter set C: voltage, pressure, wire feeding speed, etc. recorded during the same period;

[0106] Simulated capacity: This batch welds 100 workpieces per hour;

[0107] Average production capacity: historical average of 105 workpieces per hour;

[0108] S1 (variance): current and related parameter variance = 8.2 (preset X2 = 5);

[0109] S2 (capacity deviation): |105-100| = 5 (preset X3 = 3);

[0110] S1=8.2>X2=5, S2=5>X3=3→Influence rate=Y1;

[0111] Therefore, the welding gun current is a key parameter, which needs to be uploaded to the cloud in real time and trigger global optimization.

[0112] As an embodiment provided by the present invention, preferably, when the conversion processing module performs the conversion processing, the following steps are performed:

[0113] Obtain the real-time strengthening and weakening data corresponding to a single node, as well as the number of strengthening and weakening data items;

[0114] Generate a dynamic numbering algorithm based on the number of strengthening data and weakening data, perform mapping processing on the strengthening data and weakening data, and obtain the data item number and the corresponding data item-number mapping table;

[0115] Retrieve the parameter corresponding to the data item number, convert the parameter into a characteristic data of a rectangle, and form a data item-number-characteristic data mapping table;

[0116] For parameters that do not have data item numbers, the parameters are converted into characteristic values + conversion values to form a data item-characteristic value + conversion value mapping table;

[0117] When data is transmitted, lightweight data represented by the mapping table is transmitted.

[0118] For example, more than 100 robots work together in the welding workshop of a certain automobile company, and the welding quality (such as weld strength and position accuracy) needs to be guaranteed.

[0119] A single edge machine can monitor the status of a single robot, but cannot optimize the beat of the entire line. As an embodiment provided by the present invention, preferably, when the perspective compensation module aggregates data from multiple edge nodes for global optimization compensation, the following steps are performed:

[0120] Select any parameter in a node and the corresponding production line, and mark the parameter and production line as: parameter to be compensated, production line to be compensated;

[0121] Obtain the parameters upstream and downstream of the parameters to be compensated on the production line to be compensated;

[0122] If the upstream and downstream parameters and the parameters to be compensated are all enhanced data, the upstream and downstream parameters are marked as optimized compensation parameters of the parameters to be compensated;

[0123] If the upstream and downstream parameters are all enhanced data and the parameters to be compensated are all weakened data, then the upstream and downstream parameters are marked as optimized compensation parameters of the parameters to be compensated;

[0124] If the parameters of the upstream and downstream are all weakened data and the parameters to be compensated are all strengthened data, the upstream parameters will be marked as the optimized compensation parameters of the parameters to be compensated; upstream and downstream refer to the upper and lower coherence of the process;

[0125] otherwise:

[0126] The average production capacity corresponding to the adjacent upstream nodes in the process is used as the production capacity corresponding to the optimized compensation parameters;

[0127] When conducting a capacity analysis:

[0128] The production capacity corresponding to the parameter to be compensated = 0.62*the production capacity corresponding to the parameter to be compensated + 0.38*the production capacity corresponding to the optimized compensation parameter.

[0129] Example 2:

[0130] As a second embodiment provided by the present invention, preferably, the global data impact rate is:

[0131] Step A01: Obtain any data item of any node from historical data and mark it as a simulation data item;

[0132] Step A02: Obtain all parameters corresponding to the simulation data item, and extract the maximum value from the obtained parameters and mark it as the simulation parameter;

[0133] Step A03: Obtain the product and production line corresponding to the simulation parameters, extract the real-time production capacity of the production line corresponding to the product output, and mark it as the simulated capacity;

[0134] Step A04: Calculate the production capacity and the average of the production capacity corresponding to all the parameters obtained in step A02, and mark the average of the production capacity as the comparison capacity;

[0135] Step A05: Among them, the benchmark difference is a preset value.

[0136] Example 3:

[0137] As a third embodiment provided by the present invention, preferably, the dynamic numbering algorithm is:

[0138] If the number of enhanced data items is greater than the number of weakened data items, the enhanced data items are numbered in sequence according to the transmission time of the corresponding parameters;

[0139] If the number of strengthened data items is less than or equal to the number of weakened data items, the weakened data items are numbered in sequence according to the transmission time of the corresponding parameters;

[0140] When numbering, the starting number is: number of strengthened data items - number of weakened data items.

[0141] As an embodiment provided by the present invention, preferably, the data item-number-feature data mapping table is:

[0142] Retrieve the specific value of the parameter corresponding to the data item number and mark it as the value to be mapped;

[0143] If the data item number is an odd number, the number in the tens place of the value to be mapped + 1 is used as the width K of the rectangle, and the value obtained by dividing the value to be mapped by the width is used as the length C of the rectangle;

[0144] If the data item number is an even number, the number in the tens place of the value to be mapped + 1 is used as the length C of the rectangle, and the value obtained by dividing the value to be mapped by the width is used as the width K of the rectangle;

[0145] Form a rectangle with a width of K, a length of C, and an area equal to the value to be mapped;

[0146] Convert the value to be mapped into the area value of the rectangle to form a data item-number-feature data mapping table.

[0147] As an embodiment provided by the present invention, preferably, the data item-feature value+conversion value mapping table is:

[0148] Get the specific value corresponding to the parameter with no data item number, marked as initial value CZi, i = 1, 2, 3, ..., m, m is a positive integer;

[0149] Retrieve the initial value CZ1 and the time when it was generated, obtain a parameter with a data item number that is closest to the time when it was generated, and mark the corresponding data item number as the characteristic value;

[0150] If the eigenvalue is an odd number, a new parameter XCZ1 is formed with the initial value CZ1 in front and the eigenvalue in the back;

[0151] If the eigenvalue is an even number, a new parameter XCZ1 is formed with the initial value CZ1 at the back and the eigenvalue at the front;

[0152] In addition, i is accumulated by 1 in sequence until all new parameters are obtained;

[0153] All the new parameters obtained are marked as new parameters XCZi to form a data item-feature value+conversion value mapping table.

[0154] A data analysis system for an industrial edge intelligent decision-making all-in-one machine uses a local data analysis module to analyze the global data impact rate of data based on real-time data and historical data, and performs data screening to divide the data of a single node into strengthened data and weakened data; a conversion processing module obtains lightweight data based on conversion processing; a perspective compensation module performs global optimization compensation on the data of a single node by aggregating data from multiple edge nodes; aggregates data from multiple edge nodes and performs global optimization analysis, retaining the real-time performance of edge computing while breaking through local data limitations to achieve an industrial intelligent decision-making system of "local agility + global optimization"; and effectively solves the data island problem of the industrial edge intelligent decision-making all-in-one machine.

[0155] Throughout this specification, references to terms such as "one embodiment," "example," or "specific example" indicate that the specific features, structures, materials, or characteristics described in conjunction with that embodiment or example are included in at least one embodiment or example of the present invention. In this specification, schematic representations of these terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0156] The preferred embodiments of the present invention disclosed above are intended only to help illustrate the present invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the present invention to the specific embodiments described. Obviously, many modifications and variations are possible based on the content of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the present invention, thereby enabling those skilled in the art to better understand and utilize the present invention. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. An industrial edge intelligent decision-making integrated machine data analysis system, characterized in that: include: The local data analysis module analyzes the global data impact rate of data based on real-time data and historical data, and performs data screening to divide the data of a single node into strengthened data and weakened data; A conversion processing module, which obtains lightweight data based on conversion processing; Lightweight data includes data item-number-feature data mapping table, data item-feature value + conversion value mapping table; The perspective compensation module is used to perform global optimization compensation on the data of a single node by aggregating the data of multiple edge nodes.

2. The industrial edge intelligent decision-making integrated machine data analysis system according to claim 1 is characterized in that: When the local data analysis module performs data screening, the following steps are performed: Step S001: Obtain real-time data and historical data of any node; Step S002: Analyze historical data and extract parameters and corresponding data items whose impact on global data exceeds a preset value X1 from the historical data; Step S003: extracting parameters corresponding to key data items in real-time data according to key data items preset by the administrator; Step S004: Mark the parameters and corresponding data items extracted in steps S002 and S003 as enhanced data, and mark the parameters and corresponding data items in the real-time data other than the enhanced data as weakened data.

3. The industrial edge intelligent decision-making integrated machine data analysis system according to claim 2 is characterized in that: The global data impact rate is: Obtain any parameter of any node from historical data, and mark the obtained parameter as a simulation parameter, and mark the data item corresponding to the simulation parameter as a simulation data item; Obtain the products and production lines corresponding to the simulation parameters, extract the real-time production capacity of the production line corresponding to the product output, and mark it as simulated capacity; Obtain all parameters corresponding to the simulation data items from the historical data and form a parameter set C: {C1, C2, C3, ..., Cn}, where n is a positive integer, indicating that the parameter set contains n parameters; Calculate the production line capacity corresponding to all parameter elements in the parameter set, and mark the average value of the obtained capacity as the average capacity; The variance corresponding to the data set formed by the simulation parameters and all parameter elements in the parameter set is marked as S1, and the absolute value of the difference between the averaged capacity and the simulated capacity is marked as S2; If S1 ≥ X2 and S2 ≥ X3, the global data influence rate corresponding to the simulation parameter is marked as Y1; If S1<X2, and S2<X3, the global data influence rate corresponding to the simulation parameter is marked as Y2; If S1<X2, and S2≥X3, the global data influence rate corresponding to the simulation parameter is marked as Y3; If S1≥X2, and S2<X3, the global data influence rate corresponding to the simulation parameter is marked as Y4; Among them, X2, X3, Y1, Y2, Y3, and Y4 are preset values; and Y1>Y2>X1>Y3>Y4.

4. The industrial edge intelligent decision-making integrated machine data analysis system according to claim 2 is characterized in that: The global data impact rate is: Step A01: Obtain any data item of any node from historical data and mark it as a simulation data item; Step A02: Obtain all parameters corresponding to the simulation data item, and extract the maximum value from the obtained parameters and mark it as the simulation parameter; Step A03: Obtain the product and production line corresponding to the simulation parameters, extract the real-time production capacity of the production line corresponding to the product output, and mark it as the simulated capacity; Step A04: Calculate the production capacity and the average of the production capacity corresponding to all the parameters obtained in step A02, and mark the average of the production capacity as the comparison capacity; Step A05: Among them, the benchmark difference is a preset value.

5. The industrial edge intelligent decision-making integrated machine data analysis system according to claim 1 is characterized in that: When the conversion processing module performs the conversion processing, the following steps are performed: Obtain the real-time strengthening and weakening data corresponding to a single node, as well as the number of strengthening and weakening data items; Generate a dynamic numbering algorithm based on the number of strengthening data and weakening data, perform mapping processing on the strengthening data and weakening data, and obtain the data item number and the corresponding data item-number mapping table; Retrieve the parameter corresponding to the data item number, convert the parameter into a characteristic data of a rectangle, and form a data item-number-characteristic data mapping table; For parameters that do not have data item numbers, the parameters are converted into characteristic values + conversion values to form a data item-characteristic value + conversion value mapping table; When data is transmitted, lightweight data represented by the mapping table is transmitted.

6. The industrial edge intelligent decision-making integrated machine data analysis system according to claim 5 is characterized in that: The dynamic numbering algorithm is: If the number of enhanced data items is greater than the number of weakened data items, the enhanced data items are numbered in sequence according to the transmission time of the corresponding parameters; If the number of strengthened data items is less than or equal to the number of weakened data items, the weakened data items are numbered in sequence according to the transmission time of the corresponding parameters; When numbering, the starting number is: number of strengthened data items - number of weakened data items.

7. The industrial edge intelligent decision-making integrated machine data analysis system according to claim 5 is characterized in that: The data item-number-feature data mapping table is: Retrieve the specific value of the parameter corresponding to the data item number and mark it as the value to be mapped; If the data item number is an odd number, the number in the tens place of the value to be mapped + 1 is used as the width K of the rectangle, and the value obtained by dividing the value to be mapped by the width is used as the length C of the rectangle; If the data item number is an even number, the number in the tens place of the value to be mapped + 1 is used as the length C of the rectangle, and the value obtained by dividing the value to be mapped by the width is used as the width K of the rectangle; Form a rectangle with a width of K, a length of C, and an area equal to the value to be mapped; Convert the value to be mapped into the area value of the rectangle to form a data item-number-feature data mapping table.

8. The industrial edge intelligent decision-making integrated machine data analysis system according to claim 5 is characterized in that: The data item-feature value+conversion value mapping table is: Get the specific value corresponding to the parameter with no data item number, marked as initial value CZi, i = 1, 2, 3, ..., m, m is a positive integer; Retrieve the initial value CZ1 and the time when it was generated, obtain a parameter with a data item number that is closest to the time when it was generated, and mark the corresponding data item number as the characteristic value; If the eigenvalue is an odd number, a new parameter XCZ1 is formed with the initial value CZ1 in front and the eigenvalue in the back; If the eigenvalue is an even number, a new parameter XCZ1 is formed with the initial value CZ1 at the back and the eigenvalue at the front; In addition, i is accumulated by 1 in sequence until all new parameters are obtained; All the new parameters obtained are marked as new parameters XCZi to form a data item-feature value+conversion value mapping table.

9. The industrial edge intelligent decision-making integrated machine data analysis system according to claim 1 is characterized in that: When the perspective compensation module performs global optimization compensation by aggregating data from multiple edge nodes, the following steps are performed: Select any parameter in a node and the corresponding production line, and mark the parameter and production line as: parameter to be compensated, production line to be compensated; Obtain the parameters upstream and downstream of the parameters to be compensated on the production line to be compensated; If the upstream and downstream parameters and the parameters to be compensated are all enhanced data, the upstream and downstream parameters are marked as optimized compensation parameters of the parameters to be compensated; If the upstream and downstream parameters are all enhanced data and the parameters to be compensated are all weakened data, then the upstream and downstream parameters are marked as optimized compensation parameters of the parameters to be compensated; If both upstream and downstream parameters are weakened data and the parameters to be compensated are strengthened data, the upstream parameters are marked as optimized compensation parameters of the parameters to be compensated; Otherwise, the mean capacity corresponding to the adjacent upstream nodes in the process is used as the capacity corresponding to the optimized compensation parameter; When performing capacity analysis, the capacity corresponding to the parameter to be compensated = 0.62*the capacity corresponding to the parameter to be compensated + 0.38*the capacity corresponding to the optimized compensation parameter.

10. The industrial edge intelligent decision-making integrated machine data analysis system according to claim 1, characterized in that: The nodes are divided by an administrator and include multiple production lines that are adjacent in terms of process.

Images