Wind power tower drum manufacturing digital operation cooperation system and method

Abstract

The invention relates to the technical field of digital manufacturing, in particular to a wind power tower manufacturing digital operation cooperation system, which comprises the following modules: a data acquisition layer module, which is used for deploying a sensor, acquiring data in real time and transmitting the data to a data processing layer, and the content of the acquired data comprises one or more of production equipment, a technological process and product quality; the data processing layer module is used for cleaning, converting and storing the collected data, establishing a unified data warehouse, mining and analyzing the data by using a data analysis algorithm, and providing support for production decision making; and the application layer module comprises a production plan management sub-module and a process control sub-module. The technical problems that in traditional wind power tower drum manufacturing, a production plan mode is large in hysteresis quality and cannot adapt to the modern manufacturing industry are solved. On the basis, the invention provides a digital operation cooperation method for manufacturing the wind power tower drum.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of digital manufacturing, and particularly relates to a wind power tower manufacturing digital operation coordination system and method. BACKGROUND

[0002] At present, under the promotion of global energy structure transformation and carbon neutralization target, the wind power industry, as one of the core fields of clean energy, is experiencing the development of scale and intelligence. The wind power tower, as the key supporting structure of the wind turbine generator, its manufacturing quality directly affects the stability and power generation efficiency of the wind turbine generator. However, the traditional wind power tower manufacturing process has the following technical bottlenecks, which restricts the improvement of production efficiency and product quality:

[0003] 1. Wind power tower manufacturing involves multi-link coordination (such as welding, cutting, assembly), but the existing production mode relies on manual experience and offline data recording, which leads to: data fragmentation: production equipment, process and product quality detection data are scattered in different systems, and there is a lack of unified integration platform; decision lag: production plan is made depending on static report, which cannot respond to equipment failure, raw material shortage or order priority change in real time, resulting in unstable delivery cycle; process parameter out of control: key process parameters rely on manual adjustment by workers, and different batches of products have poor consistency.

[0004] 2. Production plan preparation mainly relies on traditional manual experience and static scheduling model, and the traditional planning system cannot integrate dynamic variables such as equipment failure, raw material shortage and personnel skill matching in real time, resulting in frequent plan adjustment. Tower production involves plate rolling, welding, corrosion prevention and other processes, and the traditional planning mode is difficult to coordinate the coupling relationship of equipment capacity, personnel skills and material supply, often resulting in process waiting or resource overload.

[0005] Therefore, the present application proposes a wind power tower manufacturing digital operation coordination system and method to realize technical breakthrough. SUMMARY

[0006] Therefore, in view of the above problems, the present application proposes a wind power tower manufacturing digital operation coordination system to solve the technical problems of large lag of production plan mode in traditional wind power tower manufacturing and inability to adapt to modern manufacturing. Based on this, a wind power tower manufacturing digital operation coordination method is proposed.

[0007] To achieve the above purpose, the present application adopts the following technical scheme: a wind power tower manufacturing digital operation coordination system, comprising the following modules:

[0008] Data acquisition layer module: deploy sensors, including one or more of temperature sensors, pressure sensors, displacement sensors, vibration sensors and current sensors, real-time data acquisition and transmission to the data processing layer, the content of the collected data includes one or more of production equipment, process and product quality;

[0009] Data processing layer module: clean, convert, store and process the collected data, establish a unified data warehouse, use data analysis algorithms to mine and analyze the data, and provide support for production decision-making;

[0010] Application layer module: including production plan management sub-module, process control sub-module;

[0011] Production plan management sub-module: based on order demand, equipment capacity, personnel skill level, raw material supply situation and order delivery priority factors, generate production plan;

[0012] Process control sub-module: manage process parameter database, record welding process parameters and machining process parameters of different specifications of wind power tower drum.

[0013] Further, a hybrid algorithm based on improved genetic algorithm and fuzzy comprehensive evaluation is used to generate production plan, the hybrid algorithm based on improved genetic algorithm and fuzzy comprehensive evaluation,

[0014] The objective function is as follows:

[0015]

[0016] The fitness function is as follows:

[0017]

[0018] Where:

[0019] n: total number of production equipment;

[0020] T i： Actual production time of the ith equipment in the planning period;

[0021] C i : rated production time of the ith equipment in the planning period (maximum available time of the equipment);

[0022] ω1: weight coefficient of equipment capacity factor, reflecting the importance of equipment capacity in production plan optimization;

[0023] m ： Total number of production personnel;

[0024] S jThe skill level score of the jth production personnel, which can be comprehensively evaluated according to the personnel's training record, work experience, operation proficiency and other factors;

[0025] S max： The maximum value of the skill level score among all production personnel;

[0026] ω2: weight coefficient of personnel skill level factor, reflecting the influence degree of personnel skill level on production plan;

[0027] p ： The number of types of required raw materials

[0028] R k The actual supply amount of the kth raw material

[0029] R max The demand amount of the kth raw material in the planning period

[0030] ω3: weight coefficient of raw material supply situation factor, indicating the proportion of raw material supply situation in production plan optimization;

[0031] q ： The total number of orders;

[0032] P l The delivery priority score of the lth order, which is comprehensively determined according to the order delivery deadline urgency, customer importance and other factors;

[0033] P max The maximum value of the delivery priority score among all orders;

[0034] ω4: weight coefficient of order delivery priority factor, reflecting the importance of order delivery priority in production plan;

[0035] Wherein, ω1+ω2+ω3+ω4=1.

[0036] Further, the application layer module further comprises:

[0037] Quality detection submodule: for real-time detection of the weld quality or dimensional accuracy of the wind tower drum, and real-time uploading of the detection data to the digital operation collaborative platform;

[0038] Equipment maintenance submodule: for real-time monitoring of production equipment, collecting operation state data of the equipment,

[0039] Further, video monitoring equipment or image recognition device is also deployed in the data acquisition module, for collecting image data and transmitting to the data processing module;

[0040] A data fusion module is further arranged to receive different types of data from the data acquisition layer module, perform data synchronization, alignment and fusion processing, and form comprehensive data.

[0041] A wind power tower manufacturing digital operation collaboration method, comprising the following steps:

[0042] Production plan intelligent execution step: according to the production plan generated by the production plan management sub-module, execute the production task, monitor the production progress in real time, and ensure that the production plan is carried out according to the plan;

[0043] Process parameter accurate execution step: the process control sub-module establishes a process parameter database, collects process parameters in real time through sensors, compares them with standard parameters in the process parameter database, sets different deviation thresholds, takes different warning and adjustment measures according to the deviation degree, and when the parameters deviate, sends out a warning in time and automatically adjusts the equipment parameters to ensure that the process parameters meet the standard requirements;

[0044] Quality defect traceability step: for the detected welding porosity or crack defects, associate production batch, operator, equipment state and environmental parameters, generate a root cause analysis report and push it to the responsible department.

[0045] Further, it further includes a multi-source data fusion processing step: after the multi-source data is collected by the data acquisition layer module, the multi-source data fusion module performs data synchronization, alignment and fusion processing to form a comprehensive data view.

[0046] Further, the multi-source data fusion processing step:

[0047] Data synchronization and alignment ensure that data from different data sources are synchronized in time stamp and necessary data alignment processing is performed to eliminate time delay and misplacement between data;

[0048] Data fusion algorithm: adopt data fusion algorithm to fuse the multi-source data, improve the accuracy and reliability of the data;

[0049] Abnormality detection and processing: detect and process abnormal data during data fusion to ensure the quality of fused data;

[0050] Comprehensive data view generation: generate a comprehensive data view according to the fused data to provide intuitive and comprehensive data display for subsequent data analysis and application.

[0051] Further, the data fusion algorithm includes:

[0052] State prediction:

[0053] X k|k-1 = F k ·X k-1|k-1+B k ·U k ;

[0054]

[0055] wherein X k|k-1 is a state prediction value; X k-1|k-1 is a state estimation value at a previous time; F k is a state transition matrix; B k is a control input matrix; U k is a control input; P k|k-1 is a prediction covariance matrix; P k-1|k-1 is a covariance matrix at a previous time; Q k is a process noise covariance matrix; is a transpose matrix of the state transition matrix F k ;

[0056] Measurement update:

[0057]

[0058] X k|k = X k|k-1 + K k ·(Z k -H k ·X k|k-1 );

[0059] P k|k =(I-K k ·H k )·P k-1|k-1 ;

[0060] wherein K k is a Kalman gain, H k is a measurement matrix, is a transpose matrix of the measurement matrix H k , Z k is a measurement value, R k is a measurement noise covariance matrix, X k|k is a state estimation value, P k|k is a covariance matrix, and I is an identity matrix.

[0061] Further, in the data fusion process, anomaly detection is achieved by the following steps:

[0062] Statistical features and business rules joint detection: calculate the statistical feature quantity of data distribution, mark the outliers exceeding the preset threshold, and at the same time, construct a business rule engine based on domain knowledge to logically check the data rationality;

[0063] Cross-sensor consistency verification: Kalman filter algorithm is used to estimate the state of multi-source sensor data, calculate the residual and detect the anomaly, if the residual exceeds the dynamically adjusted confidence interval, it is marked as potential anomaly;

[0064] Deep learning time series modeling: LSTM-Autoencoder model is used to encode and reconstruct time series data, and unknown mode anomaly is identified by comparing reconstruction error with adaptive threshold value;

[0065] Among them, the dynamically adjusted confidence interval is updated in real time according to the historical data volatility rate, and the adaptive threshold value is dynamically generated based on the percentile of the reconstruction error of the training set.

[0066] Further, the abnormality processing step comprises:

[0067] Hierarchical repair strategy: the detected abnormal data is classified according to the severity, wherein the fatal abnormality immediately triggers system alarm and terminates the fusion process, and the recoverable abnormality is repaired according to the data type by selecting interpolation method or regression prediction method;

[0068] Dynamic weight fault tolerance mechanism: for the data source containing abnormality, its weight value is reduced in the fusion process.

[0069] By adopting the foregoing technical solutions, the present application has the following beneficial effects:

[0070] 1. By deploying temperature, pressure, displacement, vibration, current and other multi-type sensors, a three-dimensional data sensing network covering production equipment, process and product quality is constructed, solving the problem of "equipment-process-quality" data fragmentation in traditional manufacturing. The modular architecture of data acquisition layer, processing layer and application layer supports flexible expansion, realizing real-time data sensing and closed-loop control of the whole process of wind turbine tower manufacturing. The problem of data island in traditional manufacturing is solved, the production transparency is improved, and data support is provided for process optimization and equipment maintenance, significantly reducing the risk of unplanned downtime.

[0071] 2. The improved genetic algorithm simulates the natural selection mechanism to achieve Pareto optimal solution among equipment capacity, personnel skills, raw material supply and order priority, realizing dynamic balance of multiple objectives. In view of the defect that genetic algorithm is easy to fall into local optimum, fuzzy rules are introduced to dynamically adjust the weight of uncertain factors (such as sudden shortage of raw materials), realizing fuzzy comprehensive evaluation compensation. The algorithm output can include visual tools such as equipment load thermal map and personnel skill matching matrix, helping dispatchers quickly locate bottleneck links and realize visual decision support.

[0072] 3. New sub-modules for quality inspection and equipment maintenance have been added. By monitoring weld quality, dimensional accuracy, and equipment operating status in real time, defect tracing and preventative maintenance are achieved. This design reduces the defect rate, extends equipment life, and promotes process improvement through root cause analysis reports, thereby lowering quality costs.

[0073] 4. Video surveillance equipment and image recognition devices were deployed, and a data fusion module was added to achieve multimodal data synchronization and comprehensive analysis. This technology solves the problem of the one-sidedness of data from a single sensor, improves the accuracy of anomaly detection (such as equipment failure and process deviation), and provides intuitive decision support for management through a comprehensive data view.

[0074] 5. Precise Execution of Process Parameters: By comparing sensor data with a standard parameter library in real time, a three-level early warning mechanism is automatically triggered (yellow warning - parameter drift, orange warning - approaching threshold, red warning - exceeding tolerance and shutdown). Intelligent Execution of Production Plans: Combining equipment status monitoring data, the sequence of processes is dynamically adjusted. This ensures that production proceeds according to plan, while real-time parameter comparison and automatic adjustment guarantee process stability and reduce quality fluctuations caused by human intervention.

[0075] 6. The multi-source data fusion processing step solves the spatiotemporal inconsistency problem of heterogeneous data (such as numerical values, images, and text) through data synchronization, alignment, and fusion. This technology improves data integrity and usability, provides a high-quality data foundation for subsequent analysis, and supports accurate decision-making.

[0076] 7. An anomaly detection and handling mechanism was introduced during the data fusion process. Through statistical features, business rules, cross-sensor consistency verification, and deep learning models, multi-level anomaly identification and isolation were achieved. This design avoids interference from anomalous data with the analysis results, improves system robustness, and reduces the impact of erroneous data through dynamic weight adjustment.

[0077] 8. A Kalman filter algorithm for state prediction and measurement updates was adopted to achieve dynamic estimation and noise filtering of time-series data. This algorithm improves the accuracy of data prediction and is particularly suitable for high-frequency vibration or temperature fluctuation scenarios, providing a reliable basis for equipment health management.

[0078] 9. By jointly detecting statistical features and business rules, verifying cross-sensor consistency, and using LSTM-Autoencoder time-series modeling, intelligent identification of unknown anomaly patterns is achieved. This technology overcomes the limitations of traditional rule engines, can adaptively learn changes in data distribution, and improves the coverage of anomaly detection.

[0079] 10、Hierarchical repair strategy and dynamic weight fault tolerance mechanism balance system real-time and data reliability by differentiating between fatal exceptions and recoverable exceptions. At the same time, dynamic weight adjustment reduces the fusion contribution of abnormal data sources, avoids the "bucket effect", and improves the overall data quality.

[0080] Drawings

[0081] Figure 1 is a process flow chart of the present application.

[0082] Figure 2 is a multi-source data fusion processing step flow chart.

[0083] Figure 3 is a step flow chart of exception detection and processing. DETAILED DESCRIPTION

[0084] The present application will be further described in conjunction with the specific drawings and embodiments of the specification.

[0085] Reference Figure 1 、 Figure 2 、 Figure 3 The embodiment provides a wind tower manufacturing digital operation collaborative system, which comprises the following modules:

[0086] Data acquisition layer module: deploy sensors, including one or more of temperature sensors, pressure sensors, displacement sensors, vibration sensors and current sensors, also deploy video monitoring equipment or image recognition devices, real-time data acquisition and transmission to the data processing layer, the content of the collected data includes one or more of production equipment, process and product quality;

[0087] Data fusion module: used for receiving different types of data from the data acquisition layer module, performing data synchronization, alignment and fusion processing, forming comprehensive data;

[0088] Data processing layer module: cleans, converts, stores and processes the collected data, establishes a unified data warehouse, uses data analysis algorithms to mine and analyze the data, and provides support for production decision-making;

[0089] Application layer module: including production plan management sub-module, process control sub-module, quality detection sub-module, equipment maintenance sub-module;

[0090] Production plan management sub-module: based on order demand, equipment capacity, personnel skill level, raw material supply situation and order delivery priority factors, generates production plan;

[0091] Process control sub-module: manages process parameter database, records welding process parameters and machining process parameters of different specifications of wind towers;

[0092] Quality detection submodule: for real-time detection of the weld quality or dimensional accuracy of the wind tower, and real-time uploading of the detection data to the digital operation collaborative platform;

[0093] Equipment maintenance submodule: for real-time monitoring of production equipment, collecting operation state data of the equipment,

[0094] A hybrid algorithm based on improved genetic algorithm and fuzzy comprehensive evaluation is used to generate production plans. The hybrid algorithm based on improved genetic algorithm and fuzzy comprehensive evaluation is used to generate production plans,

[0095] The objective function is as follows:

[0096]

[0097] The fitness function is as follows:

[0098]

[0099] Wherein:

[0100] n: total number of production equipment;

[0101] T i： Actual production time of the i-th equipment within the planning period;

[0102] C i Rated production time of the i-th equipment within the planning period (maximum available time of the equipment);

[0103] ω1: weight coefficient of equipment capacity factor, reflecting the importance of equipment capacity in production plan optimization;

[0104] m ： Total number of production personnel;

[0105] S j Skill level score of the j-th production personnel, which can be evaluated based on factors such as training records, work experience, and operation proficiency;

[0106] S max： Maximum value of skill level scores among all production personnel;

[0107] ω2: weight coefficient of personnel skill level factor, reflecting the influence of personnel skill level on production plan;

[0108] p ： Number of types of required raw materials

[0109] R k Actual supply amount of the k-th raw material

[0110] R max : demand of the kth raw material in the planning period

[0111] ω3: weight coefficient of raw material supply situation factor, indicating the proportion of raw material supply situation in production plan optimization;

[0112] q ： total number of orders;

[0113] P l : delivery priority score of the lth order, determined according to factors such as order delivery deadline urgency and customer importance;

[0114] P max : maximum value of delivery priority scores of all orders;

[0115] ω4: weight coefficient of order delivery priority factor, reflecting the importance of order delivery priority in production plan;

[0116] wherein ω1+ω2+ω3+ω4=1.

[0117] In the above scheme, no video monitoring device or image recognition device can be deployed in data collection, that is, no image data can be collected, and thus no data fusion module can be set. In the above application layer module, no quality detection submodule or device maintenance submodule can be included. The generation of the production plan can also use a conventional algorithm.

[0118] A wind power tower manufacturing digital operation collaboration method based on the same inventive concept, comprising the following steps:

[0119] S1, production plan intelligent execution step: according to the production plan generated by the production plan management submodule, execute the production task, monitor the production progress in real time, and ensure that the production plan is carried out according to the plan;

[0120] S2, precise execution of process parameters: the process control submodule establishes a process parameter database, collects process parameters in real time through sensors, compares them with the standard parameters in the process parameter database, sets different deviation thresholds, takes different warning and adjustment measures according to the deviation degree, and when the parameters deviate, sends out a warning in time and automatically adjusts the equipment parameters to ensure that the process parameters meet the standard requirements;

[0121] S3, multi-source data fusion processing step: after collecting multi-source data in the data collection layer module, the multi-source data fusion module performs data synchronization, alignment and fusion processing to form a comprehensive data view.

[0122] S4. Quality Defect Tracing Steps: For detected weld porosity or crack defects, associate them with production batch, operators, equipment status, and environmental parameters to generate a root cause analysis report and push it to the responsible department.

[0123] Multi-source data fusion processing steps:

[0124] S31. Data synchronization and alignment: Ensure that data from different data sources are synchronized on timestamps and perform necessary data alignment processing to eliminate delays and misalignments between data.

[0125] S32. Data fusion algorithm: The data fusion algorithm is used to fuse multi-source data to improve the accuracy and reliability of the data.

[0126] S33. Anomaly detection and handling: During the data fusion process, abnormal data is detected and handled to ensure the quality of the fused data.

[0127] S34. Generate a comprehensive data view: Based on the merged data, generate a comprehensive data view to provide an intuitive and comprehensive data display for subsequent data analysis and applications.

[0128] Data fusion algorithms include:

[0129] State prediction:

[0130] X k|k-1 =F k ·X k-1|k-1 +B k ·U k ;

[0131]

[0132] Among them, X k|k-1 It is the predicted state value; X k-1|k-1 It is the state estimate from the previous moment; F k It is the state transition matrix; B k It is the control input matrix; U k It is a control input; P k|k-1 It predicts the covariance matrix; P k-1|k-1 Q is the covariance matrix of the previous time step; k It is the process noise covariance matrix; It is the state transition matrix F k The transpose of the matrix;

[0133] Measurement Update:

[0134]

[0135] X k|k =X k|k-1 +Kk · (Z k k · X k|k-1 );

[0136] P k|k = (I - K k · H k ) · P k-1|k-1 ;

[0137] wherein K k is the Kalman gain, H k is the measurement matrix, is the transpose matrix of the measurement matrix H k , Z k is the measurement value, R k is the measurement noise covariance matrix, X k|k is the state estimate value, P k|k is the covariance matrix, and I is the unit matrix.

[0138] In the data fusion process, the anomaly detection is realized by the following steps:

[0139] S331, statistical feature and business rule joint detection: calculate the statistical feature quantity of data distribution, mark the outliers exceeding the preset threshold, and simultaneously construct a business rule engine based on domain knowledge to logically check the data rationality;

[0140] S332, cross-sensor consistency verification: the Kalman filtering algorithm is used to estimate the state of multi-source sensor data, calculate the residual and detect the anomaly, and if the residual exceeds the dynamically adjusted confidence interval, it is marked as a potential anomaly;

[0141] S333, deep learning time series modeling: the LSTM-Autoencoder model is used to encode and reconstruct the time series data, and the unknown mode anomaly is identified by comparing the reconstruction error with the adaptive threshold;

[0142] wherein the dynamically adjusted confidence interval is updated in real time according to the historical data volatility rate, and the adaptive threshold is dynamically generated based on the percentiles of the reconstruction error of the training set.

[0143] The anomaly processing step includes:

[0144] S334, hierarchical repair strategy: the detected abnormal data is classified according to the severity, wherein the fatal anomaly immediately triggers a system alarm and terminates the fusion process, and the recoverable anomaly is repaired according to the data type by selecting the interpolation method or the regression prediction method;

[0145] S335, dynamic weight fault tolerance mechanism: for the data source containing the anomaly, the weight value thereof is reduced in the fusion process.

[0146] ​The multi-source data fusion processing step described above can also not be set, the data fusion algorithm described above can also use a conventional data fusion algorithm, and the abnormality detection step and the abnormality processing step described above can also use other methods.

[0147] Although the present application has been particularly shown and described with reference to preferred embodiments, it will be understood by those skilled in the art that various changes in form and details can be made therein without departing from the spirit and scope of the application as defined by the appended claims.

Claims

1. A wind tower manufacturing digital operation collaborative system, characterized in that, Comprise the following modules: Data acquisition layer module: deploy sensors, sensors including temperature sensor, pressure sensor, displacement sensor, vibration sensor and current sensor one or more, real-time data acquisition and transmission to the transmission to the data processing layer, the content of the collected data includes one or more of production equipment, process and product quality; Data processing layer module: the collected data is cleaned, converted, stored and processed, a unified data warehouse is established, data analysis algorithms are used to mine and analyze the data, and support is provided for production decision-making; Application layer module: including production plan management sub-module, process control sub-module; Production plan management sub-module: based on order demand, equipment capacity, personnel skill level, raw material supply situation and order delivery priority factors, generate production plan; Process control sub-module: manage process parameter database, record welding process parameters and machining process parameters of different specifications of wind turbine tower drum.

2. The wind turbine tower manufacturing digitalized operation collaborative system according to claim 1, wherein, A hybrid algorithm based on improved genetic algorithm and fuzzy comprehensive evaluation is used to generate production plan, and a hybrid algorithm based on improved genetic algorithm and fuzzy comprehensive evaluation is used, The objective function is as follows: The fitness function is as follows: Wherein: n: total number of production equipment; T i： actual production time of the ith device within the planning period; C i : scheduled production time of the ith equipment in the planning period (maximum available time of the equipment) ω1: weight coefficient of equipment capacity factor, reflecting the importance of equipment capacity in production plan optimization; m ： Total number of production personnel; S j : the skill level score of the jth production personnel, which can be comprehensively evaluated according to the training record, work experience, operation proficiency, and other factors of the personnel; S max： Maximum of skill level score across all production personnel; ω2: weight coefficient of personnel skill level factor, reflecting the influence of personnel skill level on production plan; p ： Number of types of raw materials required R k : actual supply amount of the kth raw material R max : Demand of the kth raw material in the planning period ω3: weight coefficient of raw material supply situation factor, indicating the proportion of raw material supply situation in production plan optimization; q ： Total number of orders; P l : the delivery priority score of the first order, determined according to the urgency of the delivery period of the order, the importance of the customer, and the like P max : maximum value of delivery priority score among all orders; ω4: weight coefficient of order delivery priority factor, reflecting the importance of order delivery priority in production plan; Where, ω1+ω2+ω3+ω4=1.

3. The wind tower manufacturing digitalized operation collaborative system of claim 1, wherein, The application layer module also includes: Quality detection sub-module: for real-time detection of wind turbine tower weld quality or dimensional accuracy, and real-time uploading of detection data to the digital operation collaborative platform; Equipment maintenance sub-module: for real-time monitoring of production equipment and collecting equipment operating state data.

4. The wind tower manufacturing digitalized operation collaborative system of claim 1, wherein, Video monitoring equipment or image recognition device is also deployed in the data acquisition module for collecting image data and transmitting to the data processing module; A data fusion module is also provided for receiving different types of data from the data acquisition layer module, performing data synchronization, alignment and fusion processing, and forming comprehensive data.

5. A wind tower manufacturing digital operation collaboration method applied to the wind tower manufacturing digital operation collaboration system of any one of claims 1 to 4, characterized in that, Comprise the following steps: Production plan intelligent execution step: according to the production plan generated by the production plan management sub-module, execute the production task, real-time monitor the production progress, ensure the production plan according to the plan; Process parameter accurate execution step: the process control sub-module establishes the process parameter database, real-time collects the process parameters through the sensor, compares with the standard parameters in the process parameter database, sets different deviation threshold, takes different early warning and adjustment measures according to the deviation degree, when the parameters deviate, timely sends out early warning and automatically adjusts the equipment parameters, ensures the process parameters meet the standard requirements; Quality defect traceability step: For detected weld porosity or crack defects, correlate production batch, operator, equipment status and environmental parameters, generate root cause analysis report and push to responsible department.

6. The windmill tower manufacturing digitalized operation collaborative method according to claim 5, characterized in that, It also includes multi-source data fusion processing step: After multi-source data is collected by the data acquisition layer module, the multi-source data fusion module performs data synchronization, alignment and fusion processing to form a comprehensive data view.

7. The windmill tower manufacturing digitalized operation collaborative method according to claim 6, characterized in that, Multi-source data fusion processing step: Data synchronization and alignment ensure that data from different data sources are synchronized in time stamp and necessary data alignment processing is performed to eliminate time delay and misplacement between data; Data fusion algorithm, using data fusion algorithm, multi-source data fusion processing, improve the accuracy and reliability of data; Abnormal detection and processing, in the process of data fusion, detect and process abnormal data to ensure the quality of fused data; Comprehensive data view generation, according to the fused data, generate comprehensive data view, provide intuitive, comprehensive data display for subsequent data analysis and application.

8. The windmill tower manufacturing digitalized operation collaborative method according to claim 7, characterized in that, Data fusion algorithm includes: State prediction: X k|k-1 = F k · X k-1|k-1 + B k · U k ; wherein X k|k-1 is the state prediction; X k-1|k-1 is the state estimate at the previous time; F k is the state transition matrix; B k is the control input matrix; U k is the control input; P k|k-1 is the prediction covariance matrix; P k-1|k-1 is the covariance matrix at the previous time; Q k is the process noise covariance matrix; is the transpose of the state transition matrix F k . Measurement update: X k|k = X k|k-1 + K k · (Z k - H k · X k|k-1 ); P k|k = (I - K k · H k ) · P k-1|k-1 ; where K k is the Kalman gain, H k is the measurement matrix, is the transpose of the measurement matrix H k , Z k is the measurement value, R k is the measurement noise covariance matrix, X k|k is the state estimate value, P k|k is the covariance matrix, and I is the identity matrix.

9. The windmill tower manufacturing digitalized operation collaborative method according to claim 7, characterized in that, In the process of data fusion, abnormal detection is realized through the following steps: Statistical characteristics and business rules joint detection: calculate the statistical characteristic quantity of data distribution, mark the outliers beyond the preset threshold, and at the same time, construct business rule engine based on domain knowledge to logically check the data rationality; Cross sensor consistency verification: Kalman filter algorithm is used to estimate the state of multi-source sensor data, calculate the residual and detect the abnormality, if the residual exceeds the dynamically adjusted confidence interval, it is marked as potential abnormality; Deep learning time series modeling: use LSTM-Autoencoder model to encode and reconstruct time series data, compare reconstruction error with adaptive threshold to identify unknown mode anomaly; Among them, the dynamically adjusted confidence interval is updated in real time according to the historical data fluctuation rate, and the adaptive threshold is dynamically generated based on the percentiles of reconstruction error of training set.

10. The wind turbine tower manufacturing digitalized operation collaborative method according to claim 9, characterized in that, The abnormal processing step includes: Hierarchical repair strategy: classify the detected abnormal data according to severity, where fatal abnormality triggers system alarm and terminates fusion process immediately, recoverable abnormality is repaired by interpolation method or regression prediction method according to data type; Dynamic weight fault tolerance mechanism: for data sources containing abnormality, reduce its weight value in the process of fusion.

Images