Reducer manufacturing information optimization analysis method and system

Abstract

The present application provides a method and system for optimizing and analyzing manufacturing information of reducers. The method includes: crawling the real-time Internet information of reducer raw materials, the historical Internet information of reducer raw materials, the real-time Internet information of reducer parts, the historical Internet information of reducer parts, and the customer demand information of reducers; extracting the data features of the information; establishing three single prediction models, namely, artificial neural network model, support vector machine model, and principal component regression analysis prediction model; according to the virtual prediction method, three single models are used to predict the manufacturing capacity of reducers, and compared with the actual manufacturing capacity of reducers; according to the daily prediction error of each model, the prediction decision theory is used to complete the model screening, and the models with poor prediction effect are eliminated. The present application realizes crawling the real-time and historical information data of online raw materials, parts, customer needs, etc. of reducers, realizes intelligent decision-making optimization analysis, and realizes the intelligent manufacturing of reducers.

Description

Technical Field

[0001] The present application relates to the technical field of reducers, and in particular to a method and system for optimizing and analyzing reducer manufacturing information. Background Art

[0002] The reducer is an independent component consisting of a gear drive, worm drive, or gear-worm drive enclosed in a rigid housing. It is often used as a reduction transmission device between the prime mover and the working machine. It plays the role of matching speed and transmitting torque between the prime mover and the working machine or actuator, and is widely used in modern machinery.

[0003] Reducers are generally used for low-speed and high-torque transmission equipment. The power of electric motors, internal combustion engines or other high-speed running engines is decelerated by meshing the gear with fewer teeth on the input shaft of the reducer with the large gear on the output shaft. Ordinary reducers will also have several pairs of gears with the same principle to achieve the ideal deceleration effect. The ratio of the number of teeth of large and small gears is the transmission ratio.

[0004] The reducer is a relatively precise machine that matches the speed and transmits torque between the prime mover and the working machine or actuator. The purpose of using it is to reduce the speed and increase the torque. It has many types and models, and different types have different uses. There are many types of reducers. According to the transmission type, they can be divided into gear reducers, worm reducers and planetary gear reducers; according to the number of transmission stages, they can be divided into single-stage and multi-stage reducers; according to the shape of the gear, they can be divided into cylindrical gear reducers, bevel gear reducers and bevel-cylindrical gear reducers; according to the layout of the transmission, they can be divided into expansion reducers, split reducers and coaxial reducers.

[0005] The main features of the worm gear reducer are reverse self-locking function, large reduction ratio, input shaft and output shaft are not on the same axis, nor on the same plane. However, it is generally large in size, low in transmission efficiency and precision. The harmonic drive of the harmonic reducer uses the controllable elastic deformation of the flexible element to transmit motion and power. It is small in size and very precise, but its disadvantages are that the flexible wheel has a limited life, is not impact-resistant, and has poor rigidity compared to metal parts. The input speed cannot be too high. The advantages of the planetary reducer are compact structure, small return clearance, high precision, long service life, and large rated output torque. But the price is slightly expensive. The gear reducer has the characteristics of small size and large transmission torque. The gear reducer is designed and manufactured on the basis of the modular combination system. There are many motor combinations, installation forms and structural schemes. The transmission ratio is graded finely to meet different working conditions and realize mechatronics. The gear reducer has high transmission efficiency, low energy consumption and superior performance. The cycloid pinwheel reducer is a transmission machine that adopts the cycloid pin gear meshing planetary transmission principle. It is an ideal transmission device with many advantages, wide applications, and can run forward and reverse.

[0006] Reducer products are widely used in various fields of national economy and national defense industry. Food and light industry, electric machinery, construction machinery, metallurgical machinery, cement machinery, environmental protection machinery, electronic appliances, road construction machinery, water conservancy machinery, chemical machinery, mining machinery, transportation machinery, building materials machinery, rubber machinery, petroleum machinery and other industries have a strong demand for reducer products.

[0007] The market with huge potential has given rise to fierce industry competition. In the cruel market competition, reducer industry enterprises must accelerate the elimination of backward production capacity, vigorously develop high-efficiency and energy-saving products, make full use of the policy opportunities of the national energy-saving products benefiting the people project, increase product updates, adjust product structure, pay attention to national industrial policies, in order to cope with the complex and changing economic environment and maintain a good development momentum.

[0008] At present, in response to the different needs of different reducer customers, reducer manufacturers can only design different reducer types, structures, select appropriate parts, working parameters, etc. according to their needs. At the same time, they also need to manually coordinate data on the parts supply side, parts demand side, assembly capacity, and complete machine demand. If the supply or demand on one side decreases or increases, it may affect the production capacity digestion of other aspects. The result of manual coordination is often a long design cycle, complex design schemes, large personnel investment, high time, human, material and financial costs, and no intelligent manufacturing has been achieved.

[0009] At present, the extensive use of network information has made almost all reducer manufacturing information available on the Internet, such as the real-time and historical information data of reducer's online raw materials, parts and components. By analyzing these data, we can make scientific predictions on the manufacturing capacity of reducers in the future. Summary of the invention

[0010] In view of this, the purpose of this application is to propose a method and system for optimizing and analyzing reducer manufacturing information. This application can specifically solve the matching problem between existing reducer manufacturing needs and manufacturing capabilities.

[0011] Based on the above purpose, this application proposes a method for optimizing and analyzing reducer manufacturing information, including:

[0012] Crawling the real-time Internet information of reducer raw materials, the historical Internet information of reducer raw materials, the real-time Internet information of reducer parts, the historical Internet information of reducer parts, and the customer demand information of reducer;

[0013] Extract data features of the reducer raw material real-time Internet information, reducer raw material historical Internet information, reducer parts real-time Internet information, reducer parts historical Internet information, and reducer customer demand information;

[0014] Establish three single prediction models: artificial neural network model, support vector machine model, and principal component regression analysis prediction model;

[0015] According to the virtual prediction method, three single models are used to predict the manufacturing capacity of reducers and compared with the actual manufacturing capacity of reducers;

[0016] According to the daily forecast errors of each model, the forecast decision theory is used to complete the model screening and eliminate the models with poor forecasting effects.

[0017] Furthermore, the method further comprises:

[0018] An optimization model is established based on the error square sum criterion to solve the weight of each single model, which is used as the weight allocated to each model in future predictions;

[0019] The future predicted values ​​of relevant variables are used as input variables, and the selected single model is used to predict the manufacturing capacity of reducers;

[0020] Based on the crawled Internet real-time information on reducer raw materials, Internet historical information on reducer raw materials, Internet real-time information on reducer parts, Internet historical information on reducer parts, and reducer customer demand information for the next month after the current time, the reducer manufacturing capacity forecast results are corrected to obtain the final reducer manufacturing capacity forecast results.

[0021] Furthermore, the crawling of the real-time Internet information of reducer raw materials, the historical Internet information of reducer raw materials, the real-time Internet information of reducer parts, the historical Internet information of reducer parts, and the customer demand information of reducer includes:

[0022] Through three types of crawlers, namely web pages, APIs and URLs, historical text information or real-time text streams under preset conditions can be crawled and stored.

[0023] Furthermore, the data features of extracting the real-time Internet information of reducer raw materials, the historical Internet information of reducer raw materials, the real-time Internet information of reducer parts, the historical Internet information of reducer parts, and the reducer customer demand information include:

[0024] The LSTM model is used to pre-extract features of the reducer raw material real-time Internet information, reducer raw material historical Internet information, reducer parts real-time Internet information, reducer parts historical Internet information, and reducer customer demand information, including:

[0025] Input training samples into the LSTM model for training;

[0026] Add BN layer and fully connected layer to the LSTM model, and use the modified LSTM model as a feature extractor.

[0027] Furthermore, the establishing of the support vector machine model comprises:

[0028] Divide the network data set into k sets, each set is used as a test set in turn, and the remaining sets are used as training sets; the network data includes the real-time Internet information of the reducer raw materials, the historical Internet information of the reducer raw materials, the real-time Internet information of the reducer parts, the historical Internet information of the reducer parts, and the reducer customer demand information;

[0029] Calculate and compare the average cross-validation accuracy of k trainings, select a set of parameters with the highest cross-validation accuracy, and apply the parameters (C, g) as the optimal values ​​of k-fold cross-validation to support vector machine regression prediction, where the penalty parameter C controls the relative weight of the model structure risk and empirical risk;

[0030] The prediction model selects g as the Gaussian kernel function, as shown in the following formula:

[0031] Where x represents the sample value, z represents the sample mean, and σ is a constant;

[0032] Apply the parameters to the model to perform anomaly detection on network traffic.

[0033] Furthermore, the prediction by the artificial neural network model, support vector machine model and principal component regression analysis prediction model is implemented through MATLAB programming.

[0034] Furthermore, the crawled Internet real-time information on reducer raw materials, Internet historical information on reducer raw materials, Internet real-time information on reducer parts, Internet historical information on reducer parts, and reducer customer demand information include reducer raw materials and parts data, text content of reducer customer demand information, text publishing time, text publishing location, text forwarding information, text publisher, text like number and text forwarding number.

[0035] Based on the above purpose, the present application also proposes a reducer manufacturing information optimization analysis system, including:

[0036] An information crawling module is used to crawl the Internet real-time information of reducer raw materials, the Internet historical information of reducer raw materials, the Internet real-time information of reducer parts, the Internet historical information of reducer parts, and the reducer customer demand information;

[0037] A feature extraction module is used to extract data features of the reducer raw material Internet real-time information, reducer raw material Internet historical information, reducer parts Internet real-time information, reducer parts Internet historical information, and reducer customer demand information;

[0038] Model building module, used to build three single prediction models: artificial neural network model, support vector machine model, and principal component regression analysis prediction model;

[0039] A virtual prediction module is used to predict the manufacturing capacity of the reducer using three single models according to the virtual prediction method, and compare it with the actual manufacturing capacity of the reducer;

[0040] The model screening module is used to complete model screening based on the daily forecast error of each model, using forecast decision theory to eliminate models with poor forecasting effects.

[0041] In general, the advantages of this application and the experience it brings to users are:

[0042] This application crawls the real-time and historical information data of online raw materials, parts, customer demand information, etc. of reducers, and realizes intelligent decision-making optimization analysis through feature extraction, modeling and model screening, thereby realizing the intelligent manufacturing of reducers. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In the accompanying drawings, unless otherwise specified, the same reference numerals throughout the multiple drawings represent the same or similar parts or elements. These drawings are not necessarily drawn to scale. It should be understood that these drawings only depict some embodiments disclosed in the present application and should not be regarded as limiting the scope of the present application.

[0044] Figure 1 A schematic diagram showing the system architecture principle of the present application is shown.

[0045] Figure 2 A flow chart of a method for optimizing and analyzing reducer manufacturing information according to an embodiment of the present application is shown.

[0046] Figure 3 A schematic diagram showing the process of inputting reducer parts and raw material information into a neural network to obtain prediction results.

[0047] Figure 4 A configuration diagram of a reducer manufacturing information optimization and analysis system according to an embodiment of the present application is shown.

[0048] Figure 5 A schematic diagram of the structure of an electronic device provided by an embodiment of the present application is shown;

[0049] Figure 6A schematic diagram of a storage medium provided in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0050] The present application will be further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the relevant invention, rather than to limit the invention. It is also necessary to explain that, for ease of description, only the parts related to the relevant invention are shown in the accompanying drawings.

[0051] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0052] Figure 1 The schematic diagram of the system architecture principle of the present application is shown. In the embodiment of the present application, the real-time and historical information data of online raw materials, parts, customer demand information of reducers, etc. are crawled, and intelligent decision-making optimization analysis is realized by extracting features, modeling and model screening, thereby realizing intelligent manufacturing of reducers.

[0053] Figure 2 The flowchart of the method for optimizing and analyzing the manufacturing information of a reducer according to an embodiment of the present application is shown. Figure 2 As shown, the reducer manufacturing information optimization analysis method includes:

[0054] Step 101: crawl the real-time Internet information of reducer raw materials, the historical Internet information of reducer raw materials, the real-time Internet information of reducer parts, the historical Internet information of reducer parts, and the customer demand information of reducer.

[0055] This application uses three types of crawlers, namely web pages, APIs and URLs, to crawl and store historical text information or real-time text streams under preset conditions, including data on raw materials and parts of reducers, text content of reducer customer demand information, time of text publication, location of text publication, forwarding information of text, publisher of text, number of likes of text and number of forwarding of text.

[0056] The components of the reducer disclosed in this embodiment mainly include: a reduction transmission mechanism, a transmission shaft, a transmission bearing, a housing, an automatic fault identification device, and a plurality of first sensors. The first sensor is installed on at least one of the reduction transmission mechanism, the transmission shaft, the transmission bearing, and the housing, and is used to detect and upload the sensor signal of the corresponding part of the reducer. It is understandable that the reducer may also include structures such as a ventilation device, a lubrication device, and an oil seal, and the ventilation device, the lubrication device, and the oil seal may also be equipped with a first sensor.

[0057] The raw materials for reducer manufacturing can include the following types, depending on the type of reducer:

[0058] 1. Forged steel

[0059] Steel has high strength and impact resistance, and its mechanical properties can be significantly improved by heat treatment, so it is the main material for manufacturing gears. Since the fiber direction of the forged blank is conducive to improving the bending strength of the gear teeth, most gears use forged blanks, and only gears with small forces and unimportant gears directly use rolled steel. According to the gear blank processing method and gear cutting process, the steel and heat treatment methods for manufacturing gears are divided into two categories:

[0060] The first category: the tooth surface hardness HB ≤ 350, with medium carbon steel 45 steel, 50 steel or medium carbon alloy steel 40Gr, 40MnB, 35SiMn and other near-tempering or normalizing treatment, the final tooth cutting can be carried out after heat treatment. After tempering, the hardness is not high (HB = 220 ~ 250), the comprehensive performance of the material (mechanical strength and impact toughness, etc.) is relatively good, suitable for low speed, medium speed and medium stable load. This type of gear is mostly used in reducer gears in industrial control equipment machinery. 45 steel is low in price, sufficient in supply, and most widely used. After normalizing, the comprehensive performance is improved, but not as good as tempering, and is mostly used for gears with large diameters that are inconvenient to temper and unimportant. When selecting the first type of material, the hardness of the small gear should be 20 ~ 40HB higher than that of the large gear, so that the life of the two gears is close to equal.

[0061] The second category: tooth surface hardness HB ≥ 350, surface quenching with medium carbon steel and medium carbon alloy steel (tooth surface hardness HRG = 50-55), or surface carbon quenching with low carbon steel and low carbon alloy steel (tooth surface hardness HRG = 58-63). After treatment, the tooth surface hardness is high; the tooth core has good toughness. Therefore, it has strong load-bearing capacity and impact resistance, but it is difficult to process and has a high cost, so it is rarely used in reducers.

[0062] 2. Cast steel

[0063] When the gear diameter is large (D>400-600mm), the gear blank is not easy to forge, so cast steel gear blank is often used and normalized. Commonly used grades include ZG45 and ZH50.

[0064] 3. Cast Iron

[0065] Cast iron is inexpensive and can be cast into complex structural shapes, but gray cast iron has poor bending strength and impact resistance, so it is only used in low-speed and light-load open gear transmissions. Commonly used grades include HT15-33, HT 20-40, HTr 30-54, etc. Ductile iron has higher mechanical properties than gray cast iron and can partially replace high-quality carbon steel. Commonly used grades include QT60-2, etc.

[0066] 4. Non-metallic materials

[0067] For high-speed and light-loaded gear transmission, non-metallic gears are often used in conjunction with another metal gear to reduce the noise of gear transmission. Commonly used non-metallic materials include phenolic laminate (fabric laminate), nylon, etc. This type of gear has low load-bearing capacity and short life, and its allowable load is only 25-30% of that of steel gears.

[0068] When selecting materials for reducers, we should "strive to save money and do more with less money" and should base ourselves on domestic and local materials. In order to facilitate production preparation and supply, the variety, brand and specification of materials should be as few as possible, and appropriate substitute materials should be considered.

[0069] For the crawled results, this application provides data storage functions based on relational and non-relational databases according to the structure of the data. Data storage includes the storage of data crawling results and the storage of analysis-related public data sets (such as public data sets of Yelp and IMDB). At the same time, it supports efficient retrieval of data stored in the database.

[0070] The web crawler module can crawl web pages and parse the text information in the web pages, the API crawler module uses the data stream interface provided by the developer to crawl, and the URL crawler module directly crawls the data set of the given download link, such as providing the data structure and link address of the public data set. The system crawls the file resources of the corresponding link according to the specified data format and parses them and provides them to the data storage module. Each module can start multiple threads to work, and each automated data collection task can set certain crawling conditions, such as setting the task to end after a period of time, and ending after crawling a certain amount of text content. This embodiment can parse the official website of the reducer raw material and parts manufacturers, and the interface crawling unit can set the specific user group that needs to be crawled or the specific keywords contained in the text that needs to be crawled when crawling. The data resources crawled by the crawler are delivered to the data storage module for storage.

[0071] Step 102: extracting data features of the reducer raw material real-time Internet information, reducer raw material historical Internet information, reducer parts real-time Internet information, reducer parts historical Internet information, and reducer customer demand information, including:

[0072] The LSTM model is used to pre-extract features of the reducer raw material real-time Internet information, reducer raw material historical Internet information, reducer parts real-time Internet information, reducer parts historical Internet information, and reducer customer demand information, including:

[0073] Input training samples into the LSTM model for training; Long Short-Term Memory (LSTM) is a time recurrent neural network designed to solve the long-term dependency problem of general RNN (recurrent neural network). All RNNs have a chain form of repeated neural network modules. In the standard RNN, this repeated structural module has only a very simple structure, such as a tanh layer.

[0074] Add BN layer and fully connected layer to the LSTM model, and use the modified LSTM model as a feature extractor.

[0075] Step 103: Establish three single prediction models: an artificial neural network model, a support vector machine model, and a principal component regression analysis prediction model; the predictions of the artificial neural network model, the support vector machine model, and the principal component regression analysis prediction model are implemented through MATLAB programming.

[0076] In the present application, an artificial neural network model is established, specifically: a 3-layer feedforward neural network is established, and the parameters P are defined as the network input value, T is the network target value, S = (S1, S2, S3) is the hidden layer unit input vector, B = (B1, B2, B3) is the hidden layer unit output vector, L is the output layer unit input value, C is the output layer unit output value, W = (W1, W2, W3) is the input layer to hidden layer connection weight, V = (V1, V2, V3) is the hidden layer to output layer connection weight, O = (O1, O2, O3) is the output threshold of each hidden layer unit, R is the output threshold of each output layer unit, E = (E1, E2, E3) is the error of each hidden layer unit, and D is the output layer unit error.

[0077] In this application, a support vector machine model is established, including:

[0078] First, the network data set is divided into k sets, each set is used as a test set in turn, and the remaining sets are used as training sets; the network data includes the real-time Internet information of the reducer raw materials, the historical Internet information of the reducer raw materials, the real-time Internet information of the reducer parts, the historical Internet information of the reducer parts, and the reducer customer demand information;

[0079] Calculate and compare the average cross-validation accuracy of k trainings, select a set of parameters with the highest cross-validation accuracy, and apply the parameters (C, g) as the optimal values ​​of k-fold cross-validation to support vector machine regression prediction, where the penalty parameter C controls the relative weight of the model structure risk and empirical risk;

[0080] The prediction model selects g as the Gaussian kernel function, as shown in the following formula:

[0081] Where x represents the sample value, z represents the sample mean, and σ is a constant;

[0082] Apply the parameters to the model to perform anomaly detection on network traffic.

[0083] In this application, a principal component regression analysis prediction model is established, specifically:

[0084] The daily supply of 5 different kinds of raw materials and parts for reducers was used as the independent variable of the principal component regression model, and the daily production volume of the reducer was used as the dependent variable to establish a prediction model.

[0085] Step 104: According to the virtual prediction method, three single models are used to predict the manufacturing capacity of the reducer, and compared with the actual manufacturing capacity of the reducer;

[0086] For example, Figure 3 As shown in the figure, the data of raw materials and parts of the reducer are used as the input of the artificial neural network model, and the prediction result of the reducer manufacturing capacity is obtained in the output layer.

[0087] By analogy, the data of raw materials and parts of reducers and customer demand information of reducers are used as inputs of the support vector machine model and the principal component regression analysis prediction model respectively to obtain the prediction results of reducer manufacturing capacity.

[0088] Step 105: Based on the daily forecast errors of each model, the forecast decision theory is used to complete model screening and eliminate models with poor forecasting effects.

[0089] The daily forecast errors of each model are shown in the following table:

[0090]

[0091] After analysis, the principal component regression analysis prediction model has a poor prediction effect, so it can be eliminated. Of course, the model can also be retained in this step.

[0092] More preferably, the present application may further include the following steps:

[0093] Step 106: Establish an optimization model based on the error square sum criterion, and solve the weight of each single model as the weight allocated to each model in future prediction;

[0094] The error sum of squares is also called the residual sum of squares, the within-group sum of squares, etc. After fitting an appropriate model based on n observations, the remaining unfitted part (ei=yi-y average) is called the residual, where y average represents the average of n observations, and the sum of all n residual squares is called the error sum of squares. In regression analysis, it is usually represented by SSE, and its size is used to indicate the quality of function fitting. Dividing the residual sum of squares by the degrees of freedom np-1 (where p is the number of independent variables) can be used as an unbiased estimate of the error variance σ2, which is usually used to test whether the fitted model is significant.

[0095] K-Means is a clustering algorithm based on the minimum error square sum criterion. The present application can establish an optimization model based on the K-Means clustering algorithm to solve the weights of each single model. Since K-Means is a relatively well-known algorithm, it will not be described here.

[0096] In the present application, for example, the weights of the three prediction single models, namely, the artificial neural network model, the support vector machine model, and the principal component regression analysis prediction model, which can be solved by the algorithm, are 0.2, 0.7, and 0.1, respectively.

[0097] Step 107: using the future predicted values ​​of the relevant variables as input variables, and using the selected single model and its weight to predict the reducer manufacturing capacity;

[0098] For example, Figure 3 The various reducer raw materials, parts, and reducer customer demand information listed in the table are input into the three models for their future forecast values, and the overall reducer manufacturing capacity forecast results are calculated based on the weights. As shown in the following table:

[0099]

[0100] Step 108: Based on the crawled real-time Internet information on reducer raw materials, historical Internet information on reducer raw materials, real-time Internet information on reducer parts, historical Internet information on reducer parts, and reducer customer demand information for the next month after the current time, the reducer manufacturing capacity forecast result is corrected to obtain the final reducer manufacturing capacity forecast result.

[0101] In this application, some data anomalies may appear in the prediction results of the three single models, causing the reducer manufacturing capacity prediction results to significantly exceed the possible supply of raw materials and parts in the next month, for example. This can be determined as an abnormal prediction result, and the prediction results that exceed the preset threshold will be deleted to obtain the final multiple reducer manufacturing capacity prediction results for user reference.

[0102] This application realizes crawling of real-time and historical information data such as online raw materials, parts, and reducer customer demand information of reducers, and realizes intelligent decision-making optimization analysis through feature extraction, modeling, and model screening, thereby realizing intelligent manufacturing of reducers.

[0103] The embodiment of the application provides a reducer manufacturing information optimization analysis system, which is used to execute the reducer manufacturing information optimization analysis method described in the above embodiment, such as Figure 4 As shown, the system includes:

[0104] The information crawling module 501 is used to crawl the real-time Internet information of reducer raw materials, the historical Internet information of reducer raw materials, the real-time Internet information of reducer parts, the historical Internet information of reducer parts, and the customer demand information of reducer;

[0105] A feature extraction module 502 is used to extract data features of the reducer raw material Internet real-time information, reducer raw material Internet historical information, reducer parts Internet real-time information, reducer parts Internet historical information, and reducer customer demand information;

[0106] Model building module 503, used to build three prediction models: artificial neural network model, support vector machine model, and principal component regression analysis prediction model;

[0107] A virtual prediction module 504 is used to predict the manufacturing capacity of the reducer using three single models according to the virtual prediction method, and compare it with the actual manufacturing capacity of the reducer;

[0108] The model screening module 505 is used to complete model screening based on the daily forecast error of each model using forecast decision theory to eliminate models with poor forecasting effects.

[0109] The reducer manufacturing information optimization and analysis system provided in the above-mentioned embodiments of the present application and the reducer manufacturing information optimization and analysis method provided in the embodiments of the present application are based on the same inventive concept and have the same beneficial effects as the methods adopted, run or implemented by the application programs stored therein.

[0110] The embodiment of the present application also provides an electronic device corresponding to the method for optimizing and analyzing the manufacturing information of a reducer provided in the above embodiment, so as to execute the method for optimizing and analyzing the manufacturing information of a reducer.

[0111] Please refer to Figure 5 , which shows a schematic diagram of an electronic device provided by some embodiments of the present application. Figure 5As shown, the electronic device 2 includes: a processor 200, a memory 201, a bus 202 and a communication interface 203, and the processor 200, the communication interface 203 and the memory 201 are connected via the bus 202; the memory 201 stores a computer program that can be run on the processor 200, and when the processor 200 runs the computer program, it executes the reducer manufacturing information optimization analysis method provided in any of the aforementioned embodiments of the present application.

[0112] The memory 201 may include a high-speed random access memory (RAM), and may also include a non-volatile memory, such as at least one disk memory. The communication connection between the system network element and at least one other network element is realized through at least one communication interface 203 (which may be wired or wireless), and the Internet, wide area network, local area network, metropolitan area network, etc. may be used.

[0113] The bus 202 may be an ISA bus, a PCI bus, or an EISA bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. The memory 201 is used to store programs, and the processor 200 executes the programs after receiving the execution instructions. The reducer manufacturing information optimization analysis method disclosed in any implementation of the embodiment of the present application may be applied to the processor 200, or implemented by the processor 200.

[0114] The processor 200 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the hardware integrated logic circuit or software instructions in the processor 200. The above processor 200 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a readily available programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present application can be implemented or executed. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in the embodiments of the present application can be directly embodied as a hardware decoding processor to be executed, or the hardware and software modules in the decoding processor can be executed. The software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory 201, and the processor 200 reads the information in the memory 201 and completes the steps of the above method in combination with its hardware.

[0115] The electronic device provided in the embodiment of the present application and the reducer manufacturing information optimization and analysis method provided in the embodiment of the present application are based on the same inventive concept and have the same beneficial effects as the methods adopted, operated or implemented therein.

[0116] The present application also provides a computer-readable storage medium corresponding to the reducer manufacturing information optimization analysis method provided in the above embodiment. Figure 6 The computer-readable storage medium shown is a CD 30 on which a computer program (ie, a program product) is stored. When the computer program is run by a processor, it executes the reducer manufacturing information optimization analysis method provided in any of the aforementioned embodiments.

[0117] It should be noted that examples of the computer-readable storage medium may also include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other optical or magnetic storage media, which are not listed here one by one.

[0118] The computer-readable storage medium provided in the above-mentioned embodiments of the present application and the reducer manufacturing information optimization and analysis method provided in the embodiments of the present application are based on the same inventive concept and have the same beneficial effects as the methods adopted, run or implemented by the application programs stored therein.

[0119] It should be noted that:

[0120] The algorithms and displays provided herein are not inherently related to any particular computer, virtual system or other device. Various general purpose systems may also be used together with the teachings based thereon. It is apparent from the above description that the structure required for constructing such systems is not intended for any particular programming language. In addition, the present application is not intended for any particular programming language either. It should be understood that the content of the present application described herein may be implemented using various programming languages, and the description of the particular language above is intended to disclose the best mode of implementation of the present application.

[0121] In the description provided herein, a large number of specific details are described. However, it is understood that the embodiments of the present application can be practiced without these specific details. In some instances, well-known methods, structures and techniques are not shown in detail so as not to obscure the understanding of this description.

[0122] Similarly, it should be understood that in order to streamline the present application and help understand one or more of the various inventive aspects, in the above description of the exemplary embodiments of the present application, the various features of the present application are sometimes grouped together into a single embodiment, figure, or description thereof. However, the disclosed method should not be interpreted as reflecting the following intention: the claimed application requires more features than the features clearly stated in each claim. More specifically, as reflected in the claims below, the inventive aspects are less than all the features of the single embodiment disclosed above. Therefore, the claims following the specific embodiment are hereby expressly incorporated into the specific embodiment, wherein each claim itself serves as a separate embodiment of the present application.

[0123] Those skilled in the art will appreciate that the modules in the devices in the embodiments may be adaptively changed and arranged in one or more devices different from the embodiments. The modules or units or components in the embodiments may be combined into one module or unit or component, and in addition they may be divided into a plurality of submodules or subunits or subcomponents. Except that at least some of such features and / or processes or units are mutually exclusive, all features disclosed in this specification (including the accompanying claims, abstracts and drawings) and all processes or units of any method or device disclosed in this manner may be combined in any combination. Unless otherwise expressly stated, each feature disclosed in this specification (including the accompanying claims, abstracts and drawings) may be replaced by an alternative feature providing the same, equivalent or similar purpose.

[0124] In addition, those skilled in the art will appreciate that, although some embodiments described herein include certain features included in other embodiments but not other features, the combination of features of different embodiments is meant to be within the scope of the present application and form different embodiments. For example, in the claims below, any one of the claimed embodiments may be used in any combination.

[0125] The various component embodiments of the present application can be implemented in hardware, or implemented in software modules running on one or more processors, or implemented in a combination thereof. It should be understood by those skilled in the art that a microprocessor or digital signal processor (DSP) can be used in practice to implement some or all functions of some or all components in the creation system of the virtual machine according to the embodiment of the present application. The application can also be implemented as a device or system program (for example, a computer program and a computer program product) for performing a part or all of the methods described herein. Such a program implementing the present application can be stored on a computer-readable medium, or can have the form of one or more signals. Such a signal can be downloaded from an Internet website, or provided on a carrier signal, or provided in any other form.

[0126] It should be noted that the above embodiments illustrate the present application rather than limit the present application, and that those skilled in the art may design alternative embodiments without departing from the scope of the appended claims. In the claims, any reference symbol between brackets should not be constructed as a limitation to the claims. The word "comprising" does not exclude the presence of elements or steps not listed in the claims. The word "one" or "an" preceding an element does not exclude the presence of multiple such elements. The present application may be implemented by means of hardware including several different elements and by means of appropriately programmed computers. In a unit claim that lists several systems, several of these systems may be embodied by the same hardware item. The use of the words first, second, and third, etc. does not indicate any order. These words may be interpreted as names.

[0127] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of various changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for optimizing and analyzing reducer manufacturing information, characterized in that: include: Crawling the real-time Internet information of reducer raw materials, the historical Internet information of reducer raw materials, the real-time Internet information of reducer parts, the historical Internet information of reducer parts, and the customer demand information of reducer; Extract data features of the reducer raw material real-time Internet information, reducer raw material historical Internet information, reducer parts real-time Internet information, reducer parts historical Internet information, and reducer customer demand information; Establish three single prediction models: artificial neural network model, support vector machine model, and principal component regression analysis prediction model; According to the virtual prediction method, three single models are used to predict the manufacturing capacity of reducers and compared with the actual manufacturing capacity of reducers; According to the daily forecast errors of each model, the forecast decision theory is used to complete the model screening and eliminate the models with poor forecasting effect; The support vector machine model is established, comprising: Divide the network data set into k sets, each set is used as a test set in turn, and the remaining sets are used as training sets; the network data includes the real-time Internet information of the reducer raw materials, the historical Internet information of the reducer raw materials, the real-time Internet information of the reducer parts, the historical Internet information of the reducer parts, and the reducer customer demand information; Calculate and compare the average cross-validation accuracy of k trainings, select a set of parameters with the highest cross-validation accuracy, and apply the parameters (C, g) as the optimal values ​​of k-fold cross-validation to support vector machine regression prediction, where the penalty parameter C controls the relative weight of the model structure risk and empirical risk; The prediction model selects g as the Gaussian kernel function, as shown in the following formula: , where x represents the sample value and z represents the sample mean. is a constant; Apply parameters to the model to detect anomalies in network traffic; The crawling of reducer raw material Internet real-time information, reducer raw material Internet historical information, reducer parts Internet real-time information, reducer parts Internet historical information, reducer customer demand information, including reducer raw material and parts data, reducer customer demand information text content, text publishing time, text publishing location, text forwarding information, text publisher, text likes and text forwarding number; An optimization model is established based on the error square sum criterion to solve the weight of each single model, which is used as the weight allocated to each model in future predictions; The future predicted values ​​of relevant variables are used as input variables, and the selected single model is used to predict the manufacturing capacity of reducers; Based on the crawled Internet real-time information on reducer raw materials, Internet historical information on reducer raw materials, Internet real-time information on reducer parts, Internet historical information on reducer parts, and reducer customer demand information for the next month after the current time, the reducer manufacturing capacity forecast results are corrected to obtain the final reducer manufacturing capacity forecast results.

2. The method according to claim 1, characterized in that The crawled reducer raw material Internet real-time information, reducer raw material Internet historical information, reducer parts Internet real-time information, reducer parts Internet historical information, reducer customer demand information, including: Through three types of crawlers, namely web pages, APIs and URLs, historical text information or real-time text streams under preset conditions can be crawled and stored.

3. The method according to claim 2, characterized in that The data features of the extracted real-time Internet information of reducer raw materials, historical Internet information of reducer raw materials, real-time Internet information of reducer parts, historical Internet information of reducer parts, and reducer customer demand information include: The LSTM model is used to pre-extract features of the reducer raw material real-time Internet information, reducer raw material historical Internet information, reducer parts real-time Internet information, reducer parts historical Internet information, and reducer customer demand information, including: Input training samples into the LSTM model for training; Add BN layer and fully connected layer to the LSTM model, and use the modified LSTM model as a feature extractor.

4. The method according to claim 1, characterized in that: The artificial neural network model, support vector machine model and principal component regression analysis prediction model are used to perform predictions through MATLAB programming.

5. A reducer manufacturing information optimization and analysis system, using the method described in any one of claims 1 to 4, characterized in that: include: An information crawling module is used to crawl the Internet real-time information of reducer raw materials, the Internet historical information of reducer raw materials, the Internet real-time information of reducer parts, the Internet historical information of reducer parts, and the reducer customer demand information; A feature extraction module is used to extract data features of the reducer raw material Internet real-time information, reducer raw material Internet historical information, reducer parts Internet real-time information, reducer parts Internet historical information, and reducer customer demand information; Model building module, used to build three single prediction models: artificial neural network model, support vector machine model, and principal component regression analysis prediction model; A virtual prediction module is used to predict the manufacturing capacity of the reducer using three single models according to the virtual prediction method, and compare it with the actual manufacturing capacity of the reducer; The model screening module is used to complete model screening based on the daily forecast error of each model, using forecast decision theory to eliminate models with poor forecasting effects.

6. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: The processor runs the computer program to implement the method according to any one of claims 1 to 4.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: The program is executed by a processor to implement the method according to any one of claims 1 to 4.

Images