CN114781760B - Fault prediction method based on big data - Google Patents
Fault prediction method based on big data Download PDFInfo
- Publication number
- CN114781760B CN114781760B CN202210683170.9A CN202210683170A CN114781760B CN 114781760 B CN114781760 B CN 114781760B CN 202210683170 A CN202210683170 A CN 202210683170A CN 114781760 B CN114781760 B CN 114781760B
- Authority
- CN
- China
- Prior art keywords
- data
- input end
- extraction unit
- feature extraction
- output end
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- General Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Physics & Mathematics (AREA)
- Economics (AREA)
- Development Economics (AREA)
- Medical Informatics (AREA)
- Mathematical Physics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Game Theory and Decision Science (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention belongs to the technical field of weapon equipment fault prediction, and discloses a fault prediction method based on big data, which determines an abnormal data set by comparing a data set in a normal state with a data set in an abnormal state, but the abnormal data set has huge data volume and can not obviously reflect the data abnormity, so that the abnormal data set is converted into an abnormal data characteristic, a corresponding relation between the abnormal data characteristic and a fault event is found through a fault prediction model, the complex relation among sensing data does not need to be analyzed, and the fault prediction model is specific to the whole complex system, so that the fault prediction method has more accurate prediction precision compared with the fault prediction through a single model.
Description
Technical Field
The invention relates to the technical field of weapon equipment fault prediction, in particular to a fault prediction method based on big data.
Background
With the development of science and technology and industry, mechanical equipment is becoming large-sized, high-speed and complicated. Therefore, the existing weaponry is generally composed of a plurality of components, and the weaponry is multi-level in structure, complex in relation among different components and high in coupling performance. Weaponry systems, once malfunctioning, often endanger personnel's lives.
Therefore, in order to improve the reliability and safety of the weaponry system, it is necessary to perform equipment failure prediction, monitor the status data of weaponry at any time, and predict the future failure of the weaponry.
The existing fault estimation method based on principal component analysis is successfully used for fault prediction, but for data in a multi-working-condition process, the fault estimation method based on principal component analysis cannot accurately predict faults, when a complex system runs under multiple working conditions, data of a sensor changes according to the current working mode, the relation among the data of the whole system is very complex, and a single model is difficult to accurately predict the conditions of all the parties.
Disclosure of Invention
Aiming at the defects in the prior art, the fault prediction method based on big data provided by the invention solves the problem that the existing method for predicting equipment faults has inaccurate prediction on a complex system.
In order to achieve the purpose of the invention, the invention adopts the technical scheme that: a big data-based fault prediction method comprises the following steps:
s1, extracting a data set based on the time sequence under the normal working state and a data set based on the time sequence under the abnormal state of the equipment from the big data;
s2, constructing an abnormal data set according to the data set based on the time sequence under the normal working state and the data set based on the time sequence under the abnormal state;
s3, extracting abnormal data characteristics from the abnormal data set, and constructing an event set;
s4, training the fault prediction model by adopting the event set to obtain a trained fault prediction model;
s5, acquiring equipment data through a sensor, and filtering the equipment data;
and S6, calculating and inputting the abnormal data characteristics of the filtered equipment data into the trained fault prediction model to obtain the fault type of the equipment to be generated.
Further, in S2, the time-series-based data set in the normal operating state is compared with the data points at the same time point in the time-series-based data set in the abnormal state, and abnormal data is extracted to construct an abnormal data set.
Further, the step S3 includes the following sub-steps:
s31, extracting abnormal data features of the abnormal data set;
s32, finding out corresponding fault events according to the abnormal data characteristics;
and S33, constructing the abnormal data characteristics and the fault events into event sets.
Further, the formula for extracting the abnormal data features in S31 is as follows:
wherein the content of the first and second substances,in order to be a characteristic of the anomaly data,the sample data size of the same type of fault event data,is as followsThe number of the abnormal data is counted,is the sample data sizeThe mean value of the data in (1),is a first weight parameter that is a function of,is a second weight parameter.
The beneficial effects of the above further scheme are: the abnormal condition of the data is characterized through the abnormal data characteristics, and compared with the original data, the abnormal condition of the data can be reflected, so that the prediction precision is higher.
Further, the fault prediction model in S4 includes: the system comprises a first feature extraction unit, a second feature extraction unit, a third feature extraction unit, a fourth feature extraction unit, a first attention fusion module, a second attention fusion module, a first splicing module and a full connection layer;
the input end of the first feature extraction unit is respectively connected with the input end of the second feature extraction unit, the input end of the third feature extraction unit and the input end of the fourth feature extraction unit and is used as the input end of the fault prediction model; the first input end of the first attention fusion module is connected with the output end of the first feature extraction unit, the second input end of the first attention fusion module is connected with the output end of the second feature extraction unit, and the output end of the first attention fusion module is connected with the first input end of the first splicing module; the first input end of the second attention fusion module is connected with the output end of the third feature extraction unit, the second input end of the second attention fusion module is connected with the output end of the fourth feature extraction unit, and the output end of the second attention fusion module is connected with the second input end of the first splicing module; and the input end of the full connection layer is connected with the output end of the first splicing module, and the output end of the full connection layer is used as the output end of the fault prediction model.
Furthermore, the convolution kernel size of the convolution layer of the first feature extraction unit is 1 × 1, the convolution kernel size of the convolution layer of the second feature extraction unit is 3 × 3, the convolution kernel size of the convolution layer of the third feature extraction unit is 5 × 5, and the convolution kernel size of the convolution layer of the fourth feature extraction unit is 7 × 7.
The beneficial effects of the above further scheme are: 4 kinds of convolution kernels with different sizes extract features from four dimensions, and an attention fusion module is introduced to pay attention to the feature of the focus.
Further, the first and second attention fusion modules include: the first convolution layer, the second convolution layer, the third convolution layer, the global average pooling layer, the weighted global average pooling layer, the global maximum pooling layer and the second splicing module;
the input end of the first convolution layer is respectively connected with the input end of the second convolution layer and the input end of the third convolution layer and is used as the input end of the first attention fusion module or the second attention fusion module; the input end of the global average pooling layer is connected with the output end of the first convolution layer, and the output end of the global average pooling layer is connected with the first input end of the second splicing module; the input end of the weighted global average pooling layer is connected with the output end of the second convolution layer, and the output end of the weighted global average pooling layer is connected with the second input end of the second splicing module; the input end of the global maximum pooling layer is connected with the output end of the third convolution layer, and the output end of the global maximum pooling layer is connected with the third input end of the second splicing module; and the output end of the second splicing module is used as the output end of the first attention fusion module or the second attention fusion module.
The beneficial effects of the above further scheme are: and features concerned by the global average pooling layer, the weighted global average pooling layer and the global maximum pooling layer are fused, so that the data features are abundant, and important features are prevented from being lost.
Further, the filtering formula in step S5 is as follows:
wherein the content of the first and second substances,is as followsThe number of the filtered data is reduced to one,is as followsThe number of the filtered data is reduced to one,in order to determine the number of data,is composed ofThe mean value of the individual filtered data,is a firstThe data of the individual equipment is stored in a memory,is composed ofThe mean value of the data of the individual pieces of equipment,is as followsThe data of the individual equipment is stored in a memory,is as followsAnd filtering the data.
In conclusion, the beneficial effects of the invention are as follows: according to the method, the abnormal data set is determined by comparing the data set in the normal state with the data set in the abnormal state, but the data volume of the abnormal data set is huge, and the abnormal data set cannot obviously show the data abnormality, so that the abnormal data set is converted into the abnormal data characteristic, the corresponding relation between the abnormal data characteristic and a fault event is found through the fault prediction model, the complex relation among all sensing data does not need to be analyzed, and the fault prediction model is specific to the whole complex system, so that compared with the fault prediction through a single model, the method has more accurate prediction precision.
Drawings
FIG. 1 is a flow chart of a big data based failure prediction method;
FIG. 2 is a schematic diagram of a fault prediction model;
fig. 3 is a schematic structural diagram of a first attention fusion module and a second attention fusion module.
Detailed Description
The following description of the embodiments of the present invention is provided to facilitate the understanding of the present invention by those skilled in the art, but it should be understood that the present invention is not limited to the scope of the embodiments, and it will be apparent to those skilled in the art that various changes may be made without departing from the spirit and scope of the invention as defined and defined in the appended claims, and all matters produced by the invention using the inventive concept are protected.
As shown in fig. 1, a big data based failure prediction method includes the following steps:
s1, extracting a data set based on the time sequence under the normal working state and a data set based on the time sequence under the abnormal state of the equipment from the big data;
in the big data, data of temperature, pressure, current, voltage, load, heat and the like in a normal working state and an abnormal state of the equipment can be extracted.
S2, constructing an abnormal data set according to the data set based on the time sequence under the normal working state and the data set based on the time sequence under the abnormal state;
the time series are used for alignment, so the time series start point can be the weapon starting point, and can also be any time point of the two data sets which is the same.
In S2, the data set based on the time series in the normal operating state is compared with the data points at the same time point in the data set based on the time series in the abnormal state, and abnormal data is extracted to construct an abnormal data set.
S3, extracting abnormal data characteristics from the abnormal data set, and constructing an event set;
the step S3 includes the following sub-steps:
s31, extracting abnormal data features of the abnormal data set;
the formula for extracting the abnormal data features in S31 is as follows:
wherein the content of the first and second substances,in order to be a characteristic of the anomaly data,the sample data size of the same type of fault event data,is as followsThe number of the abnormal data is counted,is the sample data sizeThe mean value of the data in (1),is a first weight parameter that is a function of,is a second weight parameter.
S32, finding out corresponding fault events according to the abnormal data characteristics;
and S33, constructing the abnormal data characteristics and the fault events into event sets.
The abnormal condition of the data is characterized by the abnormal data characteristics, and compared with the original data, the abnormal condition of the data can be embodied, so that the prediction precision is higher.
S4, training the fault prediction model by adopting the event set to obtain a trained fault prediction model;
as shown in fig. 2, the fault prediction model in S4 includes: the system comprises a first feature extraction unit, a second feature extraction unit, a third feature extraction unit, a fourth feature extraction unit, a first attention fusion module, a second attention fusion module, a first splicing module and a full connection layer;
the input end of the first feature extraction unit is respectively connected with the input end of the second feature extraction unit, the input end of the third feature extraction unit and the input end of the fourth feature extraction unit and is used as the input end of the fault prediction model; the first input end of the first attention fusion module is connected with the output end of the first feature extraction unit, the second input end of the first attention fusion module is connected with the output end of the second feature extraction unit, and the output end of the first attention fusion module is connected with the first input end of the first splicing module; the first input end of the second attention fusion module is connected with the output end of the third feature extraction unit, the second input end of the second attention fusion module is connected with the output end of the fourth feature extraction unit, and the output end of the second attention fusion module is connected with the second input end of the first splicing module; and the input end of the full connection layer is connected with the output end of the first splicing module, and the output end of the full connection layer is used as the output end of the fault prediction model.
The convolution kernel size of the convolution layer of the first feature extraction unit is 1 x 1, the convolution kernel size of the convolution layer of the second feature extraction unit is 3 x 3, the convolution kernel size of the convolution layer of the third feature extraction unit is 5 x 5, and the convolution kernel size of the convolution layer of the fourth feature extraction unit is 7.
4 kinds of convolution kernels with different sizes extract features from four dimensions, and an attention fusion module is introduced to pay attention to the feature of the focus.
As shown in fig. 3, the first and second attention fusion modules include: the first convolution layer, the second convolution layer, the third convolution layer, the global average pooling layer, the weighted global average pooling layer, the global maximum pooling layer and the second splicing module;
the input end of the first convolution layer is respectively connected with the input end of the second convolution layer and the input end of the third convolution layer and serves as the input end of the first attention fusion module or the second attention fusion module; the input end of the global average pooling layer is connected with the output end of the first convolution layer, and the output end of the global average pooling layer is connected with the first input end of the second splicing module; the input end of the weighted global average pooling layer is connected with the output end of the second convolution layer, and the output end of the weighted global average pooling layer is connected with the second input end of the second splicing module; the input end of the global maximum pooling layer is connected with the output end of the third convolution layer, and the output end of the global maximum pooling layer is connected with the third input end of the second splicing module; and the output end of the second splicing module is used as the output end of the first attention fusion module or the second attention fusion module.
The characteristics concerned by the global average pooling layer, the weighted global average pooling layer and the global maximum pooling layer are fused, so that the data characteristics are abundant, the loss of important characteristics is prevented, and the data are spliced and then input to a follow-up module by the splicing module.
S5, acquiring equipment data through a sensor, and filtering the equipment data;
the filtering formula in step S5 is:
wherein the content of the first and second substances,is as followsThe number of the filtered data is reduced to one,is as followsThe number of the filtered data is set to be,the number of the data is the number of the data,is composed ofThe mean value of the individual filtered data,is as followsThe data of the individual equipment is stored in a memory,is composed ofThe mean value of the data of each piece of equipment,is as followsThe data of the individual equipment is stored in a memory,is as followsAnd filtering the data.
And S6, calculating and inputting the abnormal data characteristics of the filtered equipment data into the trained fault prediction model to obtain the fault type of the equipment to be generated.
The formula of calculating the abnormal data characteristic of the filter-processed equipment data in step S6 is the same as that in step S31.
According to the method, the abnormal data set is determined by comparing the data set in the normal state with the data set in the abnormal state, but the data volume of the abnormal data set is huge, and the abnormal data set cannot obviously show the data abnormality, so that the abnormal data set is converted into the abnormal data characteristic, the corresponding relation between the abnormal data characteristic and a fault event is found through the fault prediction model, the complex relation among all sensing data does not need to be analyzed, and the fault prediction model is specific to the whole complex system, so that compared with the fault prediction through a single model, the method has more accurate prediction precision.
Claims (1)
1. A big data-based fault prediction method is characterized by comprising the following steps:
s1, extracting a data set based on the time sequence under the normal working state and a data set based on the time sequence under the abnormal state of the equipment from the big data;
s2, constructing an abnormal data set according to the data set based on the time sequence under the normal working state and the data set based on the time sequence under the abnormal state;
s3, extracting abnormal data characteristics from the abnormal data set, and constructing an event set;
s4, training the fault prediction model by adopting the event set to obtain a trained fault prediction model;
s5, acquiring equipment data through a sensor, and filtering the equipment data;
s6, calculating and inputting the abnormal data characteristics of the filtered equipment data into the trained fault prediction model to obtain the type of the fault of the equipment;
in S2, comparing the data set based on the time sequence in the normal operating state with the data points at the same time point in the data set based on the time sequence in the abnormal state, extracting abnormal data, and constructing an abnormal data set;
the S3 comprises the following substeps:
s31, extracting abnormal data features of the abnormal data set;
s32, finding out corresponding fault events according to the abnormal data characteristics;
s33, constructing the abnormal data characteristics and the fault events into event sets;
the formula for extracting the abnormal data features in S31 is as follows:
wherein R is abnormal data characteristic, I is sample data size of similar fault event data, x (I) is ith abnormal data,the data mean value in the sample data volume I is shown, alpha is a first weight parameter, and beta is a second weight parameter;
the fault prediction model in S4 includes: the system comprises a first feature extraction unit, a second feature extraction unit, a third feature extraction unit, a fourth feature extraction unit, a first attention fusion module, a second attention fusion module, a first splicing module and a full connection layer;
the input end of the first feature extraction unit is respectively connected with the input end of the second feature extraction unit, the input end of the third feature extraction unit and the input end of the fourth feature extraction unit and is used as the input end of the fault prediction model; the first input end of the first attention fusion module is connected with the output end of the first feature extraction unit, the second input end of the first attention fusion module is connected with the output end of the second feature extraction unit, and the output end of the first attention fusion module is connected with the first input end of the first splicing module; the first input end of the second attention fusion module is connected with the output end of the third feature extraction unit, the second input end of the second attention fusion module is connected with the output end of the fourth feature extraction unit, and the output end of the second attention fusion module is connected with the second input end of the first splicing module; the input end of the full connection layer is connected with the output end of the first splicing module, and the output end of the full connection layer is used as the output end of the fault prediction model;
the convolution kernel size of the convolution layer of the first feature extraction unit is 1 x 1, the convolution kernel size of the convolution layer of the second feature extraction unit is 3 x 3, the convolution kernel size of the convolution layer of the third feature extraction unit is 5 x 5, and the convolution kernel size of the convolution layer of the fourth feature extraction unit is 7;
the first and second attention fusion modules comprise: the first convolution layer, the second convolution layer, the third convolution layer, the global average pooling layer, the weighted global average pooling layer, the global maximum pooling layer and the second splicing module;
the input end of the first convolution layer is respectively connected with the input end of the second convolution layer and the input end of the third convolution layer and is used as the input end of the first attention fusion module or the second attention fusion module; the input end of the global average pooling layer is connected with the output end of the first convolution layer, and the output end of the global average pooling layer is connected with the first input end of the second splicing module; the input end of the weighted global average pooling layer is connected with the output end of the second convolution layer, and the output end of the weighted global average pooling layer is connected with the second input end of the second splicing module; the input end of the global maximum pooling layer is connected with the output end of the third convolution layer, and the output end of the global maximum pooling layer is connected with the third input end of the second splicing module; the output end of the second splicing module is used as the output end of the first attention fusion module or the second attention fusion module;
the filtering formula in S5 is:
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210683170.9A CN114781760B (en) | 2022-06-17 | 2022-06-17 | Fault prediction method based on big data |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210683170.9A CN114781760B (en) | 2022-06-17 | 2022-06-17 | Fault prediction method based on big data |
Publications (2)
Publication Number | Publication Date |
---|---|
CN114781760A CN114781760A (en) | 2022-07-22 |
CN114781760B true CN114781760B (en) | 2022-09-20 |
Family
ID=82421510
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210683170.9A Active CN114781760B (en) | 2022-06-17 | 2022-06-17 | Fault prediction method based on big data |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN114781760B (en) |
Family Cites Families (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102355230B (en) * | 2011-07-05 | 2016-06-22 | 中兴通讯股份有限公司 | A kind of digital filter apparatus and method thereof |
CN102279343B (en) * | 2011-07-23 | 2013-08-07 | 华北电力大学(保定) | Online detection method of broken rotor bar fault of cage-type asynchronous motor based on ESPRIT-SAA |
CN102507230B (en) * | 2011-10-08 | 2013-12-25 | 中北大学 | Method for diagnosing fault of automatic ammunition supply and transportation device |
CN107480731A (en) * | 2017-09-06 | 2017-12-15 | 西安西热电站信息技术有限公司 | A kind of EARLY RECOGNITION method of thermal power plant's automobile assembly welding Iine fault signature |
CN108414923A (en) * | 2018-02-05 | 2018-08-17 | 武汉大学 | A kind of analog-circuit fault diagnosis method based on the extraction of depth confidence network characterization |
CA3123025A1 (en) * | 2018-12-17 | 2020-06-25 | Saudi Arabian Oil Company | Image based inspection of well equipment |
CN111723925B (en) * | 2020-06-17 | 2023-05-02 | 中南大学 | Fault diagnosis method, device, equipment and medium for on-road intelligent train air conditioning unit |
CN112528472A (en) * | 2020-11-27 | 2021-03-19 | 南京邮电大学 | Multi-innovation hybrid Kalman filtering and H-infinity filtering algorithm |
CN113484738A (en) * | 2021-05-25 | 2021-10-08 | 北京航空航天大学 | Circuit fault diagnosis method based on multi-feature information fusion |
CN114118166A (en) * | 2021-12-03 | 2022-03-01 | 甘肃电通电力工程设计咨询有限公司 | Transformer fault diagnosis system and method based on intelligent frequency spectrum monitoring |
CN114626433A (en) * | 2022-01-27 | 2022-06-14 | 宁夏隆基宁光仪表股份有限公司 | Fault prediction and classification method, device and system for intelligent electric energy meter |
-
2022
- 2022-06-17 CN CN202210683170.9A patent/CN114781760B/en active Active
Also Published As
Publication number | Publication date |
---|---|
CN114781760A (en) | 2022-07-22 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111563606A (en) | Equipment predictive maintenance method and device | |
CN113343581B (en) | Transformer fault diagnosis method based on graph Markov neural network | |
CN115828466A (en) | Fan main shaft component fault prediction method based on wide kernel convolution | |
CN113011248A (en) | Nuclear power plant fault diagnosis method based on LSTM | |
CN115980560A (en) | CNN-GRU-based high-voltage circuit breaker mechanical fault diagnosis system, method and equipment | |
CN114781760B (en) | Fault prediction method based on big data | |
CN114625009A (en) | Fault detection method based on system identification and optimal filtering | |
CN114519382A (en) | Nuclear power plant key operation parameter extraction and abnormity monitoring method | |
CN113487086A (en) | Method and device for predicting remaining service life of equipment, computer equipment and medium | |
WO2024087404A1 (en) | Nuclear reactor fault determination method, apparatus, device, storage medium, and product | |
CN115660198B (en) | Method for predicting residual service life of rolling bearing | |
CN116578922A (en) | Valve cooling system fault diagnosis method and device based on multichannel convolutional neural network | |
CN115375039A (en) | Industrial equipment fault prediction method and device, electronic equipment and storage medium | |
CN111061254B (en) | PHM system performance evaluation method and system | |
CN113586554A (en) | Hydraulic motor detection system and method | |
CN117250942B (en) | Fault prediction method, device, equipment and storage medium for determining model | |
CN118052255A (en) | First prediction time determining method, device, equipment and storage medium | |
CN116125936A (en) | Fault prediction method, device, equipment and readable storage medium | |
CN116882287A (en) | Lead-cooled reactor state monitoring and fault diagnosis method and system | |
CN113761103A (en) | Batch data processing method and device and electronic equipment | |
CN116464693A (en) | Method, device, equipment and storage medium for diagnosing faults of hydro-pneumatic system | |
CN111459126A (en) | Method for determining fault event occurrence probability of electrical system under condition of few data | |
CN117668717A (en) | Power supply protection method and device for fault-tolerant server | |
CN105630642B (en) | Abnormality diagnostic method and device based on PCA overdose space | |
Li et al. | Impact of AI-Assisted System on Vehicle Maintenance: A System Dynamics Approach |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |