Intelligent online detection and defect early warning system for wind power ring forging production
By using multimodal data acquisition and cloud platform modeling, the challenge of online detection across all processes, modes, and time periods in the production of wind power ring forgings has been solved, enabling high-precision identification of defects and risk prediction, thereby improving the safety and reliability of the production process.
Patent Information
- Application Number
- CN202511013621.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-23
- Publication Date
- 2025-11-18
- Estimated Expiration
- 2045-07-23
AI Technical Summary
The existing production system for wind power ring forgings cannot achieve online detection of all processes, all modes, and all time periods. It also lacks spatiotemporal statistical modeling of defect evolution patterns, making it difficult to accurately predict defect development trends. The serious data silo phenomenon hinders the refinement and dynamism of risk warning.
A multimodal data acquisition module is used to simultaneously trigger the probe to acquire longitudinal wave and EMA data. A fused feature vector is generated through correction and feature extraction. Combined with the cloud platform, a spatiotemporal probability model of defect growth is constructed to calculate the real-time risk level index and trigger an early warning.
It enables high-precision identification and risk prediction of minute and early-stage defects, improves the safety and reliability of the production process, realizes closed-loop detection and early warning throughout the entire chain, and significantly enhances the predictability of the production process.
Smart Images
Figure CN120522284B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of intelligent manufacturing technology, and in particular to an intelligent online detection and defect early warning system for the production of wind power generation ring forgings. Background Technology
[0002] With the large-scale production of wind power equipment and the advancement of the "carbon neutrality" strategy, the key component of wind turbine generators—ring forgings—is facing higher requirements for reliability and service life. Traditional processing quality control methods mainly rely on offline non-destructive testing (NDT), such as magnetic particle testing, radiographic testing, and single ultrasonic testing. These methods often only perform inspections after forging, limiting the detection rate and accuracy of detection locations. Furthermore, with the rise of smart manufacturing and the Industrial Internet of Things (IIoT), online monitoring technology has gradually become a research hotspot. Domestic and international scholars have explored combining various sensing technologies such as ultrasonic, eddy current, and acoustic emission with edge computing and cloud platforms to achieve real-time data acquisition and preliminary analysis during the production process. However, existing systems mostly support only a single sensing mode or deploy simple data analysis functions only at local workstations, making it difficult to meet the full-process, full-modality, and all-time online inspection needs of wind turbine ring forgings under high-temperature, high-speed, and closed-loop production conditions.
[0003] The current technology still has the following main shortcomings: First, data acquisition is mostly single-mode, and neither longitudinal wave nor EMA (electromagnetic acoustic excitation) signals can fully reflect the multidimensional characteristics of internal defects in materials; second, online detection systems mostly remain at the threshold alarm level at the edge level, lacking spatiotemporal statistical modeling of defect evolution laws, making it difficult to accurately predict defect development trends; third, data silos between different processes and detection stations have not been broken down, hindering the refinement and dynamism of risk warning. Summary of the Invention
[0004] This invention addresses the problems existing in current intelligent online inspection and defect early warning systems for wind power generation ring forgings. Therefore, the problem this invention aims to solve is how to provide an intelligent online inspection and defect early warning system for wind power generation ring forgings.
[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution:
[0006] In a first aspect, the present invention provides an intelligent online detection and defect early warning system for the production of wind power generation ring forgings, comprising a data acquisition module for synchronously triggering probes to acquire raw longitudinal wave data and raw EMA data of the produced wind power generation ring forgings;
[0007] The calibration and fusion module is used to calibrate the acquired raw P-wave data, extract P-wave features from the calibrated P-wave data, extract features from the raw EMA data to generate EMA spectral features, generate a fused feature vector from the P-wave features and EMA spectral features, and upload it to the cloud platform via the network.
[0008] The cloud platform is used to construct a spatiotemporal probability model of defect growth by uploading fused feature vectors and historical fused feature vectors, calculate and solve model parameters, generate a real-time risk level index of defects based on model parameters and risk thresholds, and trigger defect warnings.
[0009] As a preferred embodiment of the intelligent online detection and defect early warning system for wind power generation ring forging production described in this invention, the step of extracting longitudinal wave features from the corrected longitudinal wave data includes:
[0010] The acquired raw P-wave data is inversely attenuated, as shown below:
[0011] ;
[0012] in: For the corrected P-wave data, For raw P-wave data, The material attenuation coefficient; Sampling time point; The rotation angle of the forging;
[0013] The envelope is calculated from the corrected P-wave data and expressed as:
[0014] ;
[0015] in: For envelope data, For Hilbert transform operators;
[0016] The extracted P-wave features include the maximum envelope amplitude and time delay features, expressed as follows:
[0017] ;
[0018] ;
[0019] in: For the maximum envelope amplitude, This is a time delay characteristic.
[0020] As a preferred embodiment of the intelligent online detection and defect early warning system for wind power generation ring forging production described in this invention, the feature extraction of the original EMA data includes:
[0021] The original EMA data is represented by a Fast Fourier Transform as follows:
[0022] ;
[0023] in: The transformed EMA data, For raw EMA data, For Fast Fourier Transform;
[0024] The resonant peak value extracted from the transformed EMA data within the EMA probe bandwidth is taken as the EMA spectral characteristic and expressed as:
[0025] ;
[0026] in: EMA spectral characteristics; For the first The center frequency of each resonant component; This represents the number of resonant components.
[0027] As a preferred embodiment of the intelligent online detection and defect early warning system for wind power generation ring forgings described in this invention, the step of generating a fused feature vector from the longitudinal wave features and EMA spectrum features includes: combining the longitudinal wave features and EMA spectrum features into a fused feature vector, expressed as:
[0028] ;
[0029] in: To fuse feature vectors, This is the transpose symbol.
[0030] As a preferred embodiment of the intelligent online detection and defect early warning system for wind power generation ring forging production described in this invention, the step of constructing a spatiotemporal probability model for defect growth by uploading fused feature vectors and historical fused feature vectors includes:
[0031] Obtain the current fused feature vector, forging rotation angle, and forging identifier ID. Based on the forging identifier ID, query the fused feature vector sequence with the same rotation angle as the current forging in the historical detection records of the same forging.
[0032] The evolution of defect features at the same location is modeled, and a spatiotemporal probability model of defect growth is constructed, which is expressed as:
[0033] ;
[0034] in: For state The conditional probability density function, For Mean, matrix Let be the probability density function of the multivariate normal distribution of covariance; for The fused feature vector at each time step; The detection time interval; This is the growth rate vector; Let be the diffusion covariance matrix.
[0035] As a preferred embodiment of the intelligent online detection and defect early warning system for wind power generation ring forging production described in this invention, the calculation and solution model parameters include:
[0036] The model parameters are solved using maximum likelihood estimation (MLE) or Bayesian inference, and are expressed as follows:
[0037] ;
[0038] ;
[0039] ;
[0040] ;
[0041] in: For the estimated growth rate vector, Number of tests The detection time interval This refers to the historical number of tests. This represents the increment of the fused feature vector between two adjacent detections. The estimated diffusion covariance matrix is... Historical number of tests The fused feature vector, Historical number of tests The fused feature vector, and Sampling time point;
[0042] The instantaneous growth rate parameter is calculated using the current features and the latest historical features, and is expressed as follows:
[0043] ;
[0044] ;
[0045] in: For instantaneous growth rate, To fuse feature vector differences, To detect the time difference, This is the current fused feature vector. Historical number of tests fused feature vectors;
[0046] The component of the dimension of the maximum rate of change in the instantaneous growth rate is used as the real-time risk level index, expressed as:
[0047] ;
[0048] in: This is a real-time risk level index. The first of the instantaneous growth rate Each feature component.
[0049] As a preferred embodiment of the intelligent online detection and defect early warning system for wind power generation ring forgings described in this invention, the real-time risk level index of defects generated based on model parameters and risk thresholds includes:
[0050] By pre-setting risk thresholds through model estimation parameters, if the real-time risk level index is less than the second risk threshold but greater than the first risk threshold, a first-level warning is triggered, and a visual warning map is sent to the HMI terminal. A cylindrical coordinate cross-section flattening map is constructed with the forging rotation angle as the horizontal axis and the forging length direction as the vertical axis. The forging parts are color-mapped according to the real-time risk level index. The forging parts with a real-time risk level index less than the first risk threshold are mapped to green, and the forging parts with a real-time risk level index less than the second risk threshold but greater than the first risk threshold are mapped to yellow, and the map is pushed to the HMI terminal.
[0051] If the real-time risk level index is greater than the second risk threshold, a level two warning is triggered, a pressure correction command is sent to the forging press control system, a forging quality traceability report is generated simultaneously, and the parameters of the historical defect spatiotemporal probability model are updated.
[0052] As a preferred embodiment of the intelligent online detection and defect early warning system for wind power generation ring forging production described in this invention, the parameters of the updated historical defect spatiotemporal probability model include:
[0053] ;
[0054] ;
[0055] in, This is the corrected growth rate vector. Let the growth rate vector be... To accumulate testing time, This is the corrected diffusion covariance matrix. Let be the diffusion covariance matrix.
[0056] The beneficial effects of this invention are: it effectively overcomes the blind spots of traditional single detection modes, improves the identification rate and risk prediction accuracy of minute and early defects, realizes a closed loop of the entire chain from data acquisition and feature analysis to model prediction and decision execution, and significantly enhances the safety, reliability and predictability of the production process. Attached Figure Description
[0057] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0058] Figure 1 This is a flowchart illustrating the implementation of an intelligent online inspection and defect early warning system for the production of ring forgings for wind power generation. Detailed Implementation
[0059] To make the above-mentioned objects, features, and advantages of the present invention more readily understood, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.
[0060] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.
[0061] Secondly, the term "an embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places throughout this specification does not necessarily refer to the same embodiment, nor is it a single embodiment or an embodiment selectively excluded from other embodiments.
[0062] Reference Figure 1 This is the first embodiment of the present invention, which provides an intelligent online detection and defect early warning system for the production of wind power generation ring forgings, including:
[0063] The acquisition module is used to synchronously trigger the probe to acquire the raw longitudinal wave data and raw EMA data of the wind power generation ring forging produced;
[0064] Specifically, a floating coupling mechanism is deployed circumferentially on the rotating tooling of the ring forging, and its interior is filled with high-temperature resistant silicone grease (≥350℃). The acoustic coupling between the probe and the surface of the high-temperature forging is dynamically maintained by the air pressure regulation module.
[0065] Simultaneously trigger the longitudinal wave probe (center frequency 5MHz) and the EMA probe (bandwidth 13MHz) to acquire raw longitudinal wave data and raw EMA data of the same detection area;
[0066] A forging position code is generated using a speed encoder to acquire two raw ultrasonic data streams—longitudinal wave (LW) and EMA—from the surface of the high-temperature forging in the same detection area, providing spatiotemporal identification for subsequent positioning and analysis. During forging rotation, acoustic coupling between the probe and the workpiece surface is maintained. The LW probe (5MHz) and EMA probe (1–3MHz) are simultaneously triggered to continuously acquire raw LW and EMA data for the corresponding areas. The forging rotation angle is recorded for each frame of data using the speed encoder, achieving spatial coordinate binding of the data. The acquisition results include raw LW data, raw EMA data, and the corresponding forging rotation angle.
[0067] The calibration and fusion module is used to calibrate the acquired raw P-wave data, extract P-wave features from the calibrated P-wave data, extract features from the raw EMA data to generate EMA spectral features, generate a fused feature vector from the P-wave features and EMA spectral features, and upload it to the cloud platform via the network.
[0068] Specifically, at the edge nodes, the synchronously acquired longitudinal wave and EMA raw data are preprocessed and feature extracted to form a fusion vector that can simultaneously reflect acoustic attenuation characteristics and resonance spectrum characteristics.
[0069] The acquired raw longitudinal wave data is subjected to longitudinal wave time-domain attenuation correction. The amplitude of the sound wave decays exponentially as it propagates in the high-temperature forging. The following formula is used for inverse attenuation:
[0070] ;
[0071] in: For the corrected P-wave data, For raw P-wave data, The material attenuation coefficient; Sampling time point; The rotation angle of the forging;
[0072] The envelope is calculated from the corrected P-wave data and expressed as:
[0073] ;
[0074] in: This is the envelope data, used for subsequent amplitude and time delay feature extraction. For Hilbert transform operators;
[0075] The extracted P-wave features include the maximum envelope amplitude and time delay features, expressed as follows:
[0076] ;
[0077] ;
[0078] in: For the maximum envelope amplitude, It is a time delay characteristic;
[0079] EMA frequency domain resonance features are extracted from the original EMA data, and a Fast Fourier Transform is performed on the original EMA data, which is expressed as:
[0080] ;
[0081] in: The transformed EMA data, For raw EMA data, For Fast Fourier Transform;
[0082] In EMA probe bandwidth (Typically, 1–3 MHz) of the transformed EMA data is used to extract several resonant peaks, which are represented as follows:
[0083] ;
[0084] in: For the first The center frequency of each resonant component; The number of resonant components; This is a characteristic of the EMA spectrum.
[0085] The longitudinal wave characteristics and EMA spectral characteristics are combined into a unified vector to generate a fused feature vector, represented as follows:
[0086]
[0087] in: To fuse feature vectors, This is the transpose symbol.
[0088] The edge nodes encapsulate the fused feature vectors corresponding to the rotation angle of each forging and upload them to the cloud platform through a secure communication link to prepare for subsequent spatiotemporal probability model input.
[0089] The cloud platform is used to construct a spatiotemporal probability model of defect growth by uploading fused feature vectors and historical fused feature vectors, calculate and solve model parameters, generate a real-time risk level index of defects based on model parameters and risk thresholds, and trigger defect warnings.
[0090] Specifically, the cloud platform receives the fused feature vector and the forging position code; retrieves historical detection data for the same forging ID, and matches historical fused feature vector sequences with the same spatial coordinates; inputs the current fused feature vector and the historical fused feature vector sequences into the spatiotemporal probability model, and outputs the defect growth rate parameter; and generates a real-time risk level index based on the defect growth rate parameter and a preset safety threshold.
[0091] Based on a spatiotemporal probability model, this method integrates current and historical multimodal features to quantify defect growth rates and dynamically assess the risk level of surface defects in forgings, providing a basis for decision-making in subsequent early warning and process adjustments.
[0092] Data retrieval and sequence matching require inputting the current fused feature vector, forging rotation angle, and forging identifier ID.
[0093] By performing historical data retrieval, the cloud platform queries the fused feature vector sequence from previous inspection records of the same forging, based on the forging ID, for the same forging with the same (or similar) rotation angle.
[0094] A spatiotemporal probability model for defect growth is constructed by modeling the evolution of defect features at the same location, as follows:
[0095] ;
[0096] in: For state The conditional probability density function, For Mean, matrix Let be the probability density function of the multivariate normal distribution of covariance; for The fused feature vector at each time step; The detection time interval; This is the growth rate vector; This is the diffusion covariance matrix, used to describe random fluctuations.
[0097] Using maximum likelihood estimation (MLE) or Bayesian inference, combining historical sequences and time labels, the model parameters are solved, as follows:
[0098] ;
[0099] ;
[0100] ;
[0101] ;
[0102] in: For the estimated growth rate vector, Number of tests The detection time interval This refers to the historical number of tests. This represents the increment of the fused feature vector between two adjacent detections. The estimated diffusion covariance matrix is... Historical number of tests The fused feature vector, Historical number of tests The fused feature vector, and Sampling time point;
[0103] To calculate the instantaneous growth rate parameter, substitute the current features and the latest historical features into the model to estimate the instantaneous growth rate, which is expressed as:
[0104] ;
[0105] ;
[0106] in: For instantaneous growth rate, To fuse feature vector differences, To detect the time difference, This is the current fused feature vector. Historical number of tests fused feature vectors;
[0107] The component of the maximum rate of change dimension (usually the envelope amplitude or a certain resonance peak) in the instantaneous growth rate is used as the real-time risk level index, expressed as:
[0108] ;
[0109] in: This is a real-time risk level index. The first of the instantaneous growth rate Each feature component;
[0110] By pre-setting risk thresholds through model estimation parameters, if the real-time risk level index is less than the second risk threshold but greater than the first risk threshold, a first-level warning is triggered, and a visual warning map is sent to the HMI terminal. A cylindrical coordinate cross-section flattening diagram is constructed with the forging rotation angle as the horizontal axis and the forging length direction as the vertical axis. The forging parts are color-mapped according to the real-time risk level index. Forging parts with a real-time risk level index less than the first risk threshold are mapped to green, and forging parts with a real-time risk level index less than the second risk threshold but greater than the first risk threshold are mapped to yellow. The map is packaged into binary image data through a protocol and pushed to the HMI terminal.
[0111] If the real-time risk level index exceeds the second risk threshold, a level two warning is triggered, sending a pressure correction command to the forging press control system. Simultaneously, a forging quality traceability report containing defect coordinates, risk level, and process correction amount is generated; and the parameters of the historical defect spatiotemporal probability model are updated.
[0112] ;
[0113] ;
[0114] in, This is the corrected growth rate vector. Let the growth rate vector be... To accumulate testing time, This is the corrected diffusion covariance matrix. Let be the diffusion covariance matrix.
[0115] This embodiment also provides a computer device applicable to the intelligent online detection and defect early warning system for wind power generation ring forging production, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement all or part of the steps of the method described in the above embodiments of the present invention.
[0116] This embodiment also provides a storage medium on which a computer program is stored. When the computer program is executed by a processor, it performs the method in any optional implementation of the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.
[0117] The storage medium proposed in this embodiment and the data storage method proposed in the above embodiments belong to the same inventive concept. Technical details not described in detail in this embodiment can be found in the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.
[0118] In summary, this system is the first to achieve multimodal, synchronized online data acquisition in the production process of wind power generation ring forgings. It accurately corrects and integrates longitudinal wave and EMA characteristics at the edge, constructing a spatiotemporal probability model capable of quantifying the average growth rate and random fluctuations. Based on model parameters and confidence thresholds, it generates interpretable real-time risk levels, driving visualized early warning and closed-loop process control. This solution effectively overcomes the blind spots of traditional single-detection modes, improves the identification rate of minute and early-stage defects and the accuracy of risk prediction, and achieves a closed-loop process from data acquisition and feature analysis to model prediction and decision execution, significantly enhancing the safety, reliability, and predictability of the production process.
[0119] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.
Claims
1. An intelligent online inspection and defect early warning system for the production of wind power generation ring forgings, characterized in that: include, The acquisition module is used to synchronously trigger the probe to acquire the raw longitudinal wave data and raw EMA data of the wind power generation ring forging produced; The calibration and fusion module is used to calibrate the acquired raw P-wave data, extract P-wave features from the calibrated P-wave data, extract features from the raw EMA data to generate EMA spectral features, generate a fused feature vector from the P-wave features and EMA spectral features, and upload it to the cloud platform via the network. The cloud platform is used to construct a spatiotemporal probability model of defect growth by uploading fused feature vectors and historical fused feature vectors, calculate and solve model parameters, generate a real-time risk level index of defects based on model parameters and risk thresholds, and trigger defect warnings. The construction of the spatiotemporal probability model for defect growth using the uploaded fused feature vector and historical fused feature vectors includes: Obtain the current fused feature vector, forging rotation angle, and forging identifier ID. Based on the forging identifier ID, query the fused feature vector sequence with the same rotation angle as the current forging in the historical detection records of the same forging. The evolution of defect features at the same location is modeled, and a spatiotemporal probability model of defect growth is constructed, which is expressed as: ； in: For state The conditional probability density function, For Mean, matrix Let be the probability density function of the multivariate normal distribution of covariance; for The fused feature vector at each time step; The detection time interval; This is the growth rate vector; The diffusion covariance matrix; The calculation and solution model parameters include: The model parameters are solved using maximum likelihood estimation (MLE) or Bayesian inference, and are expressed as follows: ； ； ； ； in: For the estimated growth rate vector, Number of tests The detection time interval This refers to the historical number of tests. This represents the increment of the fused feature vector between two adjacent detections. The estimated diffusion covariance matrix is... Historical number of tests The fused feature vector, Historical number of tests The fused feature vector, and Sampling time point; The instantaneous growth rate parameter is calculated using the current features and the latest historical features, and is expressed as follows: ； ； in: For instantaneous growth rate, To fuse feature vector differences, To detect the time difference, This is the current fused feature vector. Historical number of tests fused feature vectors; The component of the dimension of the maximum rate of change in the instantaneous growth rate is used as the real-time risk level index, expressed as: ； in: This is a real-time risk level index. The first of the instantaneous growth rate Each feature component.
2. The intelligent online detection and defect early warning system for wind power generation ring forging production as described in claim 1, characterized in that: The extraction of P-wave features from the corrected P-wave data includes: The acquired raw P-wave data is inversely attenuated, as shown below: ； in: For the corrected P-wave data, For raw P-wave data, The material attenuation coefficient; Sampling time point; The rotation angle; The envelope is calculated from the corrected P-wave data and expressed as: ； in: For envelope data, For Hilbert transform operators; The extracted P-wave features include the maximum envelope amplitude and time delay features, expressed as follows: ； ； in: For the maximum envelope amplitude, This is a time delay characteristic.
3. The intelligent online detection and defect early warning system for wind power generation ring forging production as described in claim 2, characterized in that: The feature extraction of the raw EMA data includes: The original EMA data is represented by a Fast Fourier Transform as follows: ； in: The transformed EMA data, For raw EMA data, For Fast Fourier Transform; The resonant peak value extracted from the transformed EMA data within the EMA probe bandwidth is taken as the EMA spectral characteristic and expressed as: ； in: EMA spectral characteristics; For the first The center frequency of each resonant component; This represents the number of resonant components.
4. The intelligent online detection and defect early warning system for wind power generation ring forging production as described in claim 3, characterized in that: The step of generating a fused feature vector from the P-wave features and EMA spectral features includes: combining the P-wave features and EMA spectral features into a fused feature vector, represented as: ； in: To fuse feature vectors, This is the transpose symbol.
5. The intelligent online detection and defect early warning system for wind power generation ring forging production as described in claim 4, characterized in that: The real-time risk level index for defects generated based on model parameters and risk thresholds includes: By pre-setting risk thresholds through model estimation parameters, if the real-time risk level index is less than the second risk threshold but greater than the first risk threshold, a first-level warning is triggered, and a visual warning map is sent to the HMI terminal. A cylindrical coordinate cross-section flattening map is constructed with the forging rotation angle as the horizontal axis and the forging length direction as the vertical axis. The forging parts are color-mapped according to the real-time risk level index. The forging parts with a real-time risk level index less than the first risk threshold are mapped to green, and the forging parts with a real-time risk level index less than the second risk threshold but greater than the first risk threshold are mapped to yellow, and the map is pushed to the HMI terminal. If the real-time risk level index is greater than the second risk threshold, a level two warning is triggered, a pressure correction command is sent to the forging press control system, a forging quality traceability report is generated simultaneously, and the parameters of the historical defect spatiotemporal probability model are updated.
6. The intelligent online detection and defect early warning system for wind power generation ring forging production as described in claim 5, characterized in that: The parameters of the updated historical defect spatiotemporal probability model include: ； ； in, This is the corrected growth rate vector. Let the growth rate vector be... To accumulate testing time, This is the corrected diffusion covariance matrix. Let be the diffusion covariance matrix.
Citation Information
Patent Citations
ULTRASONIC METHOD FOR MEASURING PRODUCT THICKNESS
RU2009130291A
Method for spatio-temporal prediction of growth state of mushrooms
WO2024045749A1