Large-scale equipment production and operation and maintenance method and system based on digital twin drive
By constructing and expanding the box graph to identify and remove noise data, the problem of insufficient box graph identification of noise data in traditional methods is solved, and the precise construction of the digital twin model is achieved.
Patent Information
- Application Number
- CN202410507898.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-04-25
- Publication Date
- 2024-12-31
- Estimated Expiration
- 2044-04-25
AI Technical Summary
Traditionally constructing a single box map based on data cannot accurately identify noise data in the operation of the device, resulting in the digital twin model being unable to accurately simulate devices in the real world.
By constructing the 0th box graph, obtain the expansion degree and expansion length based on its left and right adjacent data segments, gradually expand the local range, generate several box graphs, calculate the necessity of each expansion, judge and remove noise data, and finally Build an accurate digital twin.
Accurate identification and removal of noise data is achieved, ensuring the accuracy and completeness of the digital twin model, and avoiding the problem of insufficient identification of a single box graph.
Smart Images
Figure CN118296310B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of data analysis technology, and in particular to a large-scale equipment production and operation maintenance method and system driven by digital twins. Background Art
[0002] As a model that can monitor the status of equipment in real time, the digital twin system can greatly improve the monitoring and management efficiency of equipment. However, in the process of building a digital twin model, a lot of data needs to be collected through many sensors. In the process of collecting data, due to the influence of external environmental factors, a large amount of noise data will be present in the collected data, resulting in the constructed digital twin model being unable to accurately simulate the equipment in the real world. Since the data in the equipment will change due to changes in the operating power of the equipment, the data of the equipment will deviate from the overall performance. Therefore, the traditional method of constructing a single box plot based on the data cannot accurately identify the noise data in the original data. Summary of the invention
[0003] The present invention provides a large-scale equipment production and operation and maintenance method and system driven by digital twins to solve the existing problem that the traditional construction of a single box plot based on data cannot accurately identify noise data in the original data.
[0004] The large-scale equipment production and operation and maintenance method and system based on digital twin drive of the present invention adopts the following technical solutions:
[0005] An embodiment of the present invention provides a large-scale equipment production and operation method based on digital twin drive, the method comprising the following steps:
[0006] Collect raw data;
[0007] Based on the original data, construct the 0th box plot;
[0008] According to the data segments adjacent to the left and right of the 0th box plot, the first left extension degree and the first right extension degree are obtained; according to the first left extension degree and the first right extension degree, the extension length of the first left extension and the extension length of the first right extension are obtained; according to the extension length of the first left extension and the extension length of the first right extension, the first box plot is obtained; the process of obtaining the first box plot according to the 0th box plot is recorded as the first extension, and each extension is obtained;
[0009] Calculate the necessity of each expansion and obtain several box plots based on the necessity of each expansion;
[0010] Determine whether the data is noise data based on several box plots; remove all noise data to obtain construction data, and build a digital twin model of the equipment based on the construction data.
[0011] Preferably, the collecting of raw data includes the following specific methods:
[0012] Through temperature sensors, pressure sensors, flow sensors and other sensors, all data in the device at all times are collected, and the data collected by the sensors are recorded as raw data; the real data collected by the same sensor is recorded as data of the same dimension to obtain raw data of all dimensions.
[0013] Preferably, the method of constructing the 0th box plot according to the original data includes:
[0014] The x-th data in any dimension is recorded as the target data. For the target data, an initial local range α is first preset. The target data and the α data closest to the target data are taken as the 0th local range. According to all the original data in the 0th local range, the 0th box plot is constructed.
[0015] Preferably, the method of obtaining the first left extension degree and the first right extension degree according to the data segments adjacent to the left and right of the 0th box plot includes the following specific methods:
[0016] A maximum expansion range L is preset, and the previous data of the first data in the 0th local range in the same dimension to the previous L data of the first data in the 0th local range in the same dimension are used as the maximum left expansion range of the 0th local range; the next data of the last data in the 0th local range in the same dimension to the next L data of the last data in the 0th local range in the same dimension are used as the maximum right expansion range of the 0th local range; finally, according to the data in the 0th local range, the data in the maximum left expansion range of the 0th local range, and the data in the maximum right expansion range of the 0th local range; the first left expansion degree and the first right expansion degree are obtained; the specific calculation formula is:
[0017]
[0018]
[0019] Where A represents the first left expansion degree; B represents the first right expansion degree; p i+1 represents the i+1th data in the 0th local range; p i Represents the i-th data in the 0-th local range; q v+1 Represents the v+1th data within the maximum left extension range of the 0th local range; q v Indicates the vth data in the maximum left extension range of the 0th local range; Represents the mean of all data in the 0th local range; r u+1Represents the u+1th data within the maximum right extension range of the 0th local range; r u Represents the u-th data within the maximum right extension range of the 0-th local range; L is the preset maximum extension range; α is the preset initial local range; || is the absolute value operation.
[0020] Preferably, the obtaining of the extension length of the first left extension and the extension length of the first right extension according to the first left extension degree and the first right extension degree includes a specific calculation formula as follows:
[0021]
[0022] E=LW
[0023] In the formula, W represents the extension length of the first left extension; E represents the extension length of the first right extension; A represents the first left extension degree; B represents the first right extension degree; L is the preset maximum extension range; round() represents the rounding function.
[0024] Preferably, the method of obtaining the first box plot according to the extension length of the first left extension and the extension length of the first right extension includes the following specific methods:
[0025] From the upper W data in the same dimension of the first data in the 0th local range to the lower E data in the same dimension of the last data in the 0th local range, take the first local range as the first box plot based on all the original data in the first local range.
[0026] Preferably, the calculation of the necessity of each expansion includes the following specific methods:
[0027] For the necessity of the jth expansion, first obtain the standard deviation of all data in each local range, and then obtain the necessity of the jth expansion according to the standard deviation of all data in each local range, the jth left expansion degree, and the jth right expansion degree. The specific calculation formula is:
[0028]
[0029] In the formula, K j Indicates the necessity of the jth expansion; A j Indicates the jth left expansion degree; B j represents the jth right expansion degree; j represents the number of local ranges from the 0th local range to the jth local range; S j-c represents the standard deviation of the data in the jcth local range; S j-c-1 represents the standard deviation of the data in the jc-1th local range; c represents the adjustment variable; norm{} represents the linear normalization function.
[0030] Preferably, the method of obtaining a plurality of box plots according to the necessity of each expansion includes:
[0031] A necessity threshold β is preset, and the necessity of each expansion is calculated. When the necessity of the expansion is greater than or equal to β, the expansion continues. When the necessity of the expansion is less than β, the expansion is stopped, and several local ranges are obtained. According to the several local ranges, several box plots are constructed.
[0032] Preferably, the method of determining whether the data is noise data according to a plurality of box plots; removing all noise data to obtain constructed data includes the following specific methods:
[0033] First, count the data outside the box in each box plot and the number of data in each box plot, and record the median in each box plot; according to the data outside the box in each box plot, the number of data in each box plot, the median in each box plot, and the target data, obtain the noise level of the target data. The specific calculation formula is:
[0034]
[0035] Where F represents the noise level of the target data; N+1 represents the number of box plots of the target data; M j Represents the number of data in the j-th box plot of the target data; m j represents the number of data outside the box in the j-th box plot of the target data; Q represents the value of the target data; H j Represents the median value of the j-th box plot of the target data; norm{} represents the linear normalization function;
[0036] Then, a noise level threshold γ is preset. When the noise level of the target data is less than γ, the target data is normal data. When the noise level of the target data is greater than or equal to γ, the target data is noise data.
[0037] Finally, all noise data are obtained, all noise data are deleted, and then the deleted noise data are filled using linear interpolation to obtain new data, which is recorded as constructed data.
[0038] An embodiment of the present invention provides a large-scale equipment production and operation system driven by digital twins, the system comprising a data acquisition module, an initial selection module, a data expansion module, a termination expansion module and a data cleaning module, wherein:
[0039] Data acquisition module, used for collecting raw data;
[0040] The initial selection module is used to construct the 0th box plot based on the original data;
[0041] The data expansion module is used to obtain the first left expansion degree and the first right expansion degree according to the data segments adjacent to the left and right of the 0th box plot; obtain the extension length of the first left extension and the extension length of the first right extension according to the first left expansion degree and the first right expansion degree; obtain the first box plot according to the extension length of the first left extension and the extension length of the first right extension; record the process of obtaining the first box plot according to the 0th box plot as the first expansion, and obtain each expansion;
[0042] The termination expansion module is used to calculate the necessity of each expansion and obtain several box plots according to the necessity of each expansion;
[0043] The data cleaning module is used to determine whether the data is noise data based on several box plots; remove all noise data to obtain construction data, and build a digital twin model of the equipment based on the construction data.
[0044] The beneficial effects of the technical solution of the present invention are as follows: the present invention first presets a small local range, and then continuously expands the local range to obtain several local ranges; at the same time, based on the stability of the data in the expanded local range, avoids introducing data that is greatly different from the initial local range into the expanded local range; finally, generates several box plots based on the several local ranges, avoiding the problem that a single box plot cannot accurately identify noise data, so that the data can be more comprehensively evaluated to achieve the purpose of accurately identifying noise data. BRIEF DESCRIPTION OF THE DRAWINGS
[0045] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. 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 paying creative work.
[0046] Figure 1 This is a flowchart of the steps of the large-scale equipment production and operation and maintenance method based on digital twin drive of the present invention;
[0047] Figure 2 This is a structural block diagram of the large-scale equipment production and operation and maintenance system driven by digital twins of the present invention. DETAILED DESCRIPTION
[0048] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following, in combination with the accompanying drawings and preferred embodiments, describes in detail the specific implementation methods, structures, features and effects of the large-scale equipment production and operation and maintenance method and system based on digital twin drive proposed by the present invention. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.
[0049] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
[0050] The specific scheme of the large-scale equipment production and operation and maintenance method and system based on digital twin drive provided by the present invention is described in detail below with reference to the accompanying drawings.
[0051] See also Figure 1 , which shows a flowchart of a large-scale equipment production and operation method based on digital twin drive provided by an embodiment of the present invention, the method comprising the following steps:
[0052] Step S001: Collect original data.
[0053] It should be noted that the digital twin model is a virtual model that simulates the real world. It can test and verify various plans and decisions in a digital environment. It has high value and application prospects. This embodiment is a large-scale equipment production and operation method driven by digital twins. Therefore, it is necessary to build a digital twin model first. To build a digital twin model, it is necessary to collect various data first.
[0054] Specifically, all data in the device at all times are collected through sensors such as temperature sensors, pressure sensors, and flow sensors, and the data collected by the sensors are recorded as raw data; the real data collected by the same sensor is recorded as data of the same dimension.
[0055] At this point, the original data of all dimensions are obtained.
[0056] Step S002: construct the 0th box plot based on the original data.
[0057] It should be noted that since the digital twin model is a virtual model that simulates the real world, in order to enable the digital twin model of the device to accurately simulate the device in the real world, no noise data can exist in the data required to build the digital twin model, so as to avoid the digital twin model of the device not being able to accurately simulate the device in the real world due to noise data. Therefore, it is not possible to directly build the digital twin model of the device through the original data collected in step S001. First, it is necessary to denoise the original data obtained in step S001, and then build the digital twin model of the device based on the denoised original data, so that the final digital twin model of the device can accurately simulate the device in the real world.
[0058] It should be further explained that, in the traditional denoising of time series data, a single box plot is first constructed based on the data, and the deviation of the data in the box plot is obtained to determine the degree of abnormality of the data; then the data is denoised according to the degree of abnormality of the data; however, the original data in this embodiment is collected from the data in the device. When the operating power of the device changes, the data of the device will deviate from the overall performance, but it is not noise data; that is, the traditional construction of a single box plot based on the data cannot accurately identify the noise data in the original data. Therefore, this embodiment proposes a denoising method for original data, first constructing the 0th box plot, and then continuously expanding outward, constructing multiple box plots based on the data obtained from each expansion, that is, constructing multiple box plots based on the data, which can accurately identify the noise data in the original data.
[0059] Since the process of determining whether any original data in each dimension is abnormal data is exactly the same, this embodiment records the xth data in any dimension as the target data, and describes it by taking the determination of whether the target data is noise data as an example; therefore, it is necessary to first construct the 0th box plot based on the target data.
[0060] Specifically, for the target data, firstly, an initial local range α is preset. The specific size of α can be set according to the actual situation. This embodiment does not make a rigid requirement. In this embodiment, α=10 is used for description. The target data and the α data closest to the target data are taken as the 0th local range. According to all the original data in the 0th local range, the 0th box plot is constructed. Since the box plot construction method is a well-known prior art, it will not be described in detail in this embodiment.
[0061] Step S003: According to the data segments adjacent to the left and right of the 0th box plot, obtain the first left expansion degree and the first right expansion degree; according to the first left expansion degree and the first right expansion degree, obtain the extension length of the first left extension and the extension length of the first right extension; according to the extension length of the first left extension and the extension length of the first right extension, obtain the first box plot; the process of obtaining the first box plot according to the 0th box plot is recorded as the first expansion, and each expansion is obtained.
[0062] It should be noted that, since this embodiment is a large-scale equipment production and operation method driven by digital twins, a method for denoising the original data of the equipment is proposed. After obtaining the 0th box plot through step S002, it can be expanded according to the 0th box plot to obtain the 1st local range, and then the 1st box plot is constructed with all the data in the 1st local range.
[0063] It should be noted that in the original data of the device, the distribution and change trend of the data on the left and right sides adjacent to the 0th box plot are likely to be different. In order to avoid introducing data with large difference in change trend as data for constructing the box plot when expanding the box plot, which results in the constructed box plot being unable to accurately analyze the data, it is necessary to calculate the first left expansion degree and the first right expansion degree of the 0th box plot.
[0064] Specifically, a maximum expansion range L is preset, and the specific value of L can be set according to actual conditions. This embodiment does not make a rigid requirement. In this embodiment, L=6 is used for description, and the previous data of the first data in the 0th local range in the same dimension to the previous L data of the first data in the 0th local range in the same dimension are used as the maximum left expansion range of the 0th local range; the next data of the last data in the 0th local range in the same dimension to the next L data of the last data in the 0th local range in the same dimension are used as the maximum right expansion range of the 0th local range; finally, according to the data in the 0th local range, the data in the maximum left expansion range of the 0th local range, and the data in the maximum right expansion range of the 0th local range, the first left expansion degree and the first right expansion degree are obtained; the specific calculation formula is:
[0065]
[0066]
[0067] Where A represents the first left expansion degree; B represents the first right expansion degree; p i+1 represents the i+1th data in the 0th local range; p i Represents the i-th data in the 0-th local range; q v+1Represents the v+1th data within the maximum left extension range of the 0th local range; q v Indicates the vth data in the maximum left extension range of the 0th local range; Represents the mean of all data in the 0th local range; r u+1 Represents the u+1th data within the maximum right extension range of the 0th local range; r u Represents the u-th data within the maximum right extension range of the 0-th local range; L is the preset maximum extension range; α is the preset initial local range; || is the absolute value operation.
[0068] It should be noted that It represents the changing trend of data in the 0th local range; It represents the transformation trend of the data within the maximum left extension range of the 0th local range; therefore It represents the difference between the change trend of the data in the 0th local range and the change trend of the data in the maximum left extension range of the 0th local range; The larger the value is, the smaller the difference is; the smaller the difference is, the more similar the change trend of the data in the 0th local range is to the change trend of the data in the maximum left extension range of the 0th local range, that is, the greater the degree of the first left extension is; at the same time It represents the difference between all the data in the maximum left extension range of the 0th local range and the overall data in the 0th local range; The larger the value of , the smaller the difference between all data in the maximum left extension range of the 0th local range and the overall data in the 0th local range, and the greater the first left extension degree. Since the principle of obtaining the first left extension degree is exactly the same as the principle of obtaining the first right extension degree, this embodiment will not repeat the principle of obtaining the first right extension degree.
[0069] It should be noted that after obtaining the first left extension degree and the first right extension degree, the extension length of the first left extension and the extension length of the first right extension can be obtained according to the first left extension degree and the first right extension degree.
[0070] Specifically, the calculation formula for obtaining the extended length of the first left extension and the extended length of the first right extension is:
[0071]
[0072] E=LW
[0073] In the formula, W represents the extension length of the first left extension; E represents the extension length of the first right extension; A represents the first left extension degree; B represents the first right extension degree; L is the preset maximum extension range; round() represents the rounding function.
[0074] Finally, from the upper W data in the same dimension of the first data in the 0th local range to the lower E data in the same dimension of the last data in the 0th local range, as the 1st local range, the 1st box plot is constructed according to all the original data in the 1st local range; since the method for constructing the box plot is a well-known prior art, it will not be described in detail in this embodiment.
[0075] At this point, the first box plot is obtained.
[0076] It should be noted that the process of obtaining the 1st box plot based on the 0th box plot is recorded as the first expansion; the process of obtaining the 2nd box plot based on the 1st box plot is recorded as the second expansion; and so on, each expansion is obtained. Since the process of each expansion is exactly the same, it will not be repeated in this embodiment.
[0077] Step S004: Calculate the necessity of each expansion, and obtain several box plots according to the necessity of each expansion.
[0078] It should be noted that, since the present embodiment constructs multiple box plots for one data to determine whether the data is noise data, the multiple box plots constructed based on the data need to satisfy the requirement that the data has a stable change feature within the extended local range, and this feature will gradually be lost as the local range of the data expands. Therefore, the data cannot be infinitely expanded based on the 0th local range, that is, when expanding based on the 0th local range of the data, there is an expansion limit. In order to obtain all local ranges of the data and generate all box plots of the data, it is necessary to calculate the necessity of each expansion, and obtain several local ranges and generate several box plots based on the necessity of each expansion.
[0079] Specifically, for the necessity of the j-th expansion, first obtain the standard deviation of all data in each local range, and then obtain the necessity of the j-th expansion according to the standard deviation of all data in each local range, the j-th left expansion degree, and the j-th right expansion degree. The specific calculation formula is:
[0080]
[0081] In the formula, K j Indicates the necessity of the jth expansion; A j Indicates the jth left expansion degree; B jrepresents the jth right expansion degree; j represents the number of local ranges from the 0th local range to the jth local range; S j-c represents the standard deviation of the data in the jcth local range; S j-c-1 represents the standard deviation of the data in the jc-1th local range; c represents the adjustment variable; norm{} represents the linear normalization function.
[0082] It should be further explained that the greater the j-th left expansion degree and the j-th right expansion degree, the more similar the maximum left expansion range of the j-th local range and the maximum right expansion range of the j-th local range are to the j-1-th local range, that is, the greater the necessity of the j-th expansion; S j-c-1 is the standard deviation of the data in the jc-1th local range, indicating the stability of the jc-1th local range, S j-c is the standard deviation of the data in the jcth local range, which indicates the stability of the jcth local range. The smaller the standard deviation, the higher the stability. Therefore, S j-c-1 -S j-c The larger the value of , the less the stability from the jc-1th local range to the jcth local range decreases, that is, the greater the necessity of the jth expansion.
[0083] Finally, a necessity threshold β is preset. The specific size of β can be set according to the actual situation. This embodiment does not make a rigid requirement. In this embodiment, β=0.6 is used for description. The necessity of each expansion is calculated. When the necessity of the expansion is greater than or equal to β, the expansion continues. When the necessity of the expansion is less than β, the expansion is stopped to obtain several local ranges. According to the several local ranges, several box plots are constructed. Since the method for constructing the box plot is a well-known prior art, it will not be described in detail in this embodiment.
[0084] At this point, several box plots are obtained.
[0085] Step S005: Determine whether the data is noise data based on several box plots; remove all noise data to obtain construction data, and build a digital twin model of the device based on the construction data.
[0086] It should be noted that, since the present embodiment determines whether the data is noise data by constructing multiple box plots for one data, after obtaining several box plots in step S004, it is possible to determine whether the target data is noise data based on these box plots.
[0087] Specifically, the data outside the box in each box plot and the number of data in each box plot are counted, and the median in each box plot is recorded; according to the data outside the box in each box plot, the number of data in each box plot, the median in each box plot and the target data, the noise degree of the target data is obtained, and the specific calculation formula is:
[0088]
[0089] Where F represents the noise level of the target data; N+1 represents the number of box plots of the target data; M j Represents the number of data in the j-th box plot of the target data; m j represents the number of data outside the box in the j-th box plot of the target data; Q represents the value of the target data; H j Represents the median value of the j-th box plot of the target data; norm{} represents the linear normalization function.
[0090] It needs to be further explained that |QH j The larger the value of |, the more the target data in the j-th box plot of the target data has the degree of noise data, so |QH j The larger the value of |, the greater the possibility that the target data is noise data; The larger the value of , the smaller the number of data outside the box in the box plot; the smaller the number of data outside the box, the higher the credibility of the box plot. Therefore, the larger the value of F, the greater the possibility that the target data is noise data.
[0091] Finally, a noise level threshold γ is preset. The specific value of γ can be set according to the actual situation. This embodiment does not make a rigid requirement. In this embodiment, γ=0.9 is used for description. When the noise level of the target data is less than γ, the target data is normal data. When the noise level of the target data is greater than or equal to γ, the target data is noise data.
[0092] At this point, the determination of whether the target data is noise data is completed; similarly, all original data in all dimensions are determined to be noise data, and all noise data are obtained.
[0093] All noise data is deleted, and then the deleted noise data is filled using linear interpolation to obtain new data, which is recorded as construction data. Since linear interpolation is a well-known prior art, it will not be described in detail in this embodiment. Then, a digital twin model of the device is constructed based on the construction data. Since the construction of the digital twin model is a well-known prior art, it will not be described in detail in this embodiment, so that the digital twin model of the device constructed based on the construction data can accurately simulate the device in the real world.
[0094] At this point, this embodiment is completed.
[0095] See also Figure 2 , which shows a structural block diagram of a large-scale equipment production and operation and maintenance system based on digital twin drive provided by an embodiment of the present invention, and the system includes the following modules:
[0096] Data acquisition module, used for collecting raw data;
[0097] The initial selection module is used to construct the 0th box plot based on the original data;
[0098] The data expansion module is used to obtain the first left expansion degree and the first right expansion degree according to the data segments adjacent to the left and right of the 0th box plot; obtain the extension length of the first left extension and the extension length of the first right extension according to the first left expansion degree and the first right expansion degree; obtain the first box plot according to the extension length of the first left extension and the extension length of the first right extension; record the process of obtaining the first box plot according to the 0th box plot as the first expansion, and obtain each expansion;
[0099] The termination expansion module is used to calculate the necessity of each expansion and obtain several box plots according to the necessity of each expansion;
[0100] The data cleaning module is used to determine whether the data is noise data based on several box plots; remove all noise data to obtain construction data, and build a digital twin model of the equipment based on the construction data.
[0101] The beneficial effects of the technical solution of the present invention are as follows: the present invention first presets a small local range, and then continuously expands the local range to obtain several local ranges; at the same time, based on the stability of the data in the expanded local range, avoids introducing data that is greatly different from the initial local range into the expanded local range; finally, generates several box plots based on the several local ranges, avoiding the problem that a single box plot cannot accurately identify noise data, so that the data can be more comprehensively evaluated to achieve the purpose of accurately identifying noise data.
[0102] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.
Claims
1. A large-scale equipment production and operation method driven by digital twins, characterized in that: The method comprises the following steps: Collect raw data; Based on the original data, construct the 0th box plot; According to the data segments adjacent to the left and right of the 0th box plot, the first left extension degree and the first right extension degree are obtained; according to the first left extension degree and the first right extension degree, the extension length of the first left extension and the extension length of the first right extension are obtained; according to the extension length of the first left extension and the extension length of the first right extension, the first box plot is obtained; the process of obtaining the first box plot according to the 0th box plot is recorded as the first extension, and each extension is obtained; Calculate the necessity of each expansion and obtain several box plots based on the necessity of each expansion; Determine whether the data is noise data based on several box plots; remove all noise data to obtain construction data, and build a digital twin model of the device based on the construction data; The specific calculation formula for obtaining the extension length of the first left extension and the extension length of the first right extension according to the first left extension degree and the first right extension degree is as follows: In the formula, Indicates the extension length of the first left extension; Indicates the extension length of the first right extension; Indicates the first left expansion degree; Indicates the first right expansion degree; The preset maximum expansion range; Represents the rounding function.
2. The large-scale equipment production and operation method based on digital twin drive according to claim 1 is characterized in that: The specific method of collecting raw data includes: Through temperature sensors, pressure sensors, and flow sensors, all data in the device at all times are collected, and the data collected by the sensors are recorded as raw data; the real data collected by the same sensor is recorded as data of the same dimension to obtain raw data of all dimensions.
3. The large-scale equipment production and operation method based on digital twin drive according to claim 1 is characterized in that: The specific method of constructing the 0th box plot according to the original data is as follows: In any dimension, The data is recorded as the target data. For the target data, an initial local range is first preset. , the target data and the closest data as the 0th local range, and construct the 0th box plot based on all the original data in the 0th local range.
4. The large-scale equipment production and operation method based on digital twin drive according to claim 3 is characterized in that: The specific method of obtaining the first left expansion degree and the first right expansion degree according to the data segments adjacent to the left and right of the 0th box plot is as follows: Preset a maximum expansion range , move the previous data in the same dimension of the first data in the 0th local range to the previous data in the same dimension of the first data in the 0th local range data as the maximum left extension range of the 0th local range; extend the next data in the same dimension of the last data in the 0th local range to the next data in the same dimension of the last data in the 0th local range. data as the maximum right extension range of the 0th local range; finally, according to the data in the 0th local range, the data in the maximum left extension range of the 0th local range, and the data in the maximum right extension range of the 0th local range; obtain the first left extension degree and the first right extension degree; The specific calculation formula is: In the formula, Indicates the first left expansion degree; Indicates the first right expansion degree; Indicates the 0th local range. individual data; Indicates the 0th local range. individual data; Indicates the maximum left extension range of the 0th local range. individual data; Indicates the maximum left extension range of the 0th local range. individual data; Represents the mean of all data in the 0th local range; Indicates the maximum right extension range of the 0th local range. individual data; Indicates the maximum right extension range of the 0th local range. individual data; The preset maximum expansion range; is the preset initial local range; It is an absolute value operation.
5. The large-scale equipment production and operation method based on digital twin drive according to claim 1 is characterized in that: The specific method of obtaining the first box plot according to the extension length of the first left extension and the extension length of the first right extension is as follows: From the first data in the 0th local range in the same dimension data, to the next data in the same dimension in the 0th local range data as the first local range, and construct the first box plot based on all the original data in the first local range.
6. The large-scale equipment production and operation method based on digital twin drive according to claim 5 is characterized in that: The calculation of the necessity of each expansion includes the following specific methods: For The necessity of the second expansion is to first obtain the standard deviation of all data in each local range, and then The second left extension and the The second right extension degree; get the The necessity of the secondary expansion is calculated as follows: In the formula, Indicates The necessity of secondary expansion; Indicates Second left extension degree; Indicates Second right extension degree; Indicates the local range from 0th to The number of local ranges in local ranges; Indicates The standard deviation of the data in a local area; Indicates The standard deviation of the data in a local area; represents the adjustment variable; represents the linear normalization function.
7. The large-scale equipment production and operation method based on digital twin drive according to claim 1 is characterized in that: The specific method of obtaining several box plots according to the necessity of each expansion is as follows: Set a threshold of necessity , calculate the necessity of each expansion, when the necessity of expansion is greater than or equal to When the necessity of expansion is less than When , the expansion stops and several local ranges are obtained; according to several local ranges, several box plots are constructed.
8. The large-scale equipment production and operation method based on digital twin drive according to claim 3 is characterized in that: The specific method of determining whether the data is noise data according to a number of box plots and removing all noise data to obtain constructed data includes: First, count the data outside the box in each box plot and the number of data in each box plot, and record the median in each box plot; according to the data outside the box in each box plot, the number of data in each box plot, the median in each box plot, and the target data, obtain the noise level of the target data. The specific calculation formula is: In the formula, Indicates the noise level of the target data; The number of box plots representing the target data; Indicates the target data The number of data in the box plot; Indicates the target data The number of data outside the box in a box plot; Indicates the value of the target data; Indicates the target data The median value in the box plot; represents the linear normalization function; Then, a noise level threshold is preset , when the noise level of the target data is less than , then the target data is normal data. When the noise level of the target data is greater than or equal to , then the target data is noise data; Finally, all noise data are obtained, all noise data are deleted, and then the deleted noise data are filled using linear interpolation to obtain new data, which is recorded as constructed data.
9. A large-scale equipment production and operation system driven by digital twins, characterized by: The system includes the following modules: Data acquisition module, used for collecting raw data; The initial selection module is used to construct the 0th box plot based on the original data; The data expansion module is used to obtain the first left expansion degree and the first right expansion degree according to the data segments adjacent to the left and right of the 0th box plot; obtain the extension length of the first left extension and the extension length of the first right extension according to the first left expansion degree and the first right expansion degree; obtain the first box plot according to the extension length of the first left extension and the extension length of the first right extension; record the process of obtaining the first box plot according to the 0th box plot as the first expansion, and obtain each expansion; The termination expansion module is used to calculate the necessity of each expansion and obtain several box plots according to the necessity of each expansion; The data cleaning module is used to determine whether the data is noise data based on several box plots; remove all noise data to obtain construction data, and build a digital twin model of the device based on the construction data; The specific calculation formula for obtaining the extension length of the first left extension and the extension length of the first right extension according to the first left extension degree and the first right extension degree is as follows: In the formula, Indicates the extension length of the first left extension; Indicates the extension length of the first right extension; Indicates the first left expansion degree; Indicates the first right expansion degree; The preset maximum expansion range; Represents the rounding function.