Safety braking system for external rotor fans based on automatic sensing

Abstract

The present invention discloses an external rotor fan safety braking system based on automatic sensing, which relates to the field of safety control technology. The system includes: dynamically acquiring first operating state information through multi-source sensing equipment; acquiring target characteristic parameters, and judging whether they meet the parameter threshold constraint; if they meet, predicting and analyzing the generated first operating parameter time series to obtain first target predicted operating parameters; judging whether the first target predicted operating parameters meet the parameter threshold constraint; if they do not meet, performing predictive braking analysis to obtain predictive braking analysis results; and performing safe braking according to the predictive braking analysis results. The present invention solves the technical problem in the prior art that the monitoring of the fan operating state is not accurate and timely enough, resulting in a shortened fan service life and a threat to personnel and equipment safety, and achieves the technical effect of realizing accurate monitoring and effective prediction of the fan operating state, extending the service life, and reducing the risk of accidents.

Description

Technical Field

[0001] The invention relates to the technical field of safety control, and in particular to an outer rotor fan safety braking system based on automatic induction. Background Art

[0002] In modern industry and daily life, external rotor fans are widely used in ventilation, cooling and air circulation. However, with the continuous increase in fan speed and load, safety issues have become increasingly prominent. Traditional external rotor fan braking systems often rely on manual monitoring and operation, which is inefficient. The monitoring of the fan operating status is not accurate and timely enough, and it is unable to accurately capture key information such as sudden speed changes, excessive temperature, abnormal vibration, etc., thus failing to effectively avoid accidents.

[0003] The existing technology has the technical problem that the monitoring of the fan operating status is not accurate and timely enough, which shortens the service life of the fan and threatens the safety of personnel and equipment. Summary of the invention

[0004] The present application provides an automatic sensing-based outer rotor fan safety braking system, which is used to solve the technical problem that the existing technology is not accurate and timely enough in monitoring the operating status of the fan, resulting in a shortened service life of the fan and a threat to the safety of personnel and equipment.

[0005] In view of the above problems, the present application provides an outer rotor fan safety braking system based on automatic sensing.

[0006] The present application provides an outer rotor fan safety braking system based on automatic induction, the system comprising:

[0007] A first operating state information acquisition module, the first operating state information acquisition module is used to dynamically acquire target operating state information of the target external rotor fan through a multi-source sensing device, and extract first operating state information from the target operating state information, wherein the first operating state information includes a plurality of operating characteristic parameters with time marks; a target characteristic parameter acquisition module, the target characteristic parameter acquisition module is used to acquire target characteristic parameters from the plurality of operating characteristic parameters with time marks, and determine whether the target characteristic parameters meet the parameter threshold constraints; a first target prediction operating parameter acquisition module, the first target prediction operating parameter acquisition module is used to, if met, predict the target characteristic parameters based on the plurality of operating characteristic parameters with time marks. The first operating parameter time series generated by the operating characteristic parameters identified by the time is predicted and analyzed to obtain the first target predicted operating parameters under the target time; a first target predicted operating parameter judgment module, the first target predicted operating parameter judgment module is used to judge whether the first target predicted operating parameter meets the parameter threshold constraint; a predictive braking analysis result acquisition module, the predictive braking analysis result acquisition module is used to call the predictive braking decision rule to perform predictive braking analysis on the target outer rotor fan if it does not meet the requirements, and obtain a predictive braking analysis result; a safety braking module, the safety braking module is used to perform safety braking on the target outer rotor fan according to the predictive braking analysis result.

[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0009] The first operation status information acquisition module dynamically extracts the first operation status information through multi-source sensing equipment; the target characteristic parameter acquisition module acquires the target characteristic parameters in the operation characteristic parameters; the first target prediction operation parameter acquisition module performs prediction analysis to obtain the first target prediction operation parameters; the first target prediction operation parameter judgment module is used to judge whether the first target prediction operation parameters meet the parameter threshold constraint; the predictive braking analysis result acquisition module performs predictive braking analysis to obtain the predictive braking analysis result; the safety braking module performs safety braking on the target outer rotor fan. The technical effect of realizing accurate monitoring and effective prediction of the fan operation status, extending the service life, and reducing the risk of accidents has been achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0011] Figure 1A schematic structural diagram of an outer rotor fan safety braking system based on automatic sensing provided in an embodiment of the present application.

[0012] Explanation of the reference numerals: first operating status information acquisition module 10 , target characteristic parameter acquisition module 20 , first target predicted operating parameter acquisition module 30 , first target predicted operating parameter judgment module 40 , predictive braking analysis result acquisition module 50 , safety braking module 60 . DETAILED DESCRIPTION

[0013] The present application provides an automatic sensing-based outer rotor fan safety braking system to solve the technical problem in the prior art that the fan operating status monitoring is not accurate and timely enough, resulting in a shortened fan service life and a threat to the safety of personnel and equipment.

[0014] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0015] Embodiment 1, as Figure 1 As shown, the present application provides an outer rotor fan safety braking system based on automatic sensing, the system comprising:

[0016] The first operating status information acquisition module 10 is used to dynamically acquire target operating status information of the target external rotor fan through a multi-source sensing device, and extract the first operating status information from the target operating status information, wherein the first operating status information includes a plurality of operating characteristic parameters with time marks.

[0017] Specifically, the first operation status information acquisition module 10 realizes dynamic acquisition of the target operation status information of the target outer rotor fan by closely cooperating with multi-source sensing devices such as torque sensors and speed sensors. The torque sensor can monitor the torque changes of the fan in real time during operation, thereby reflecting the load condition of the fan; the speed sensor is responsible for accurately measuring the speed of the fan. These multi-source sensing devices continuously transmit the collected information about the operation status of the fan to the first operation status information acquisition module 10. After receiving this information, the target operation status information is deeply analyzed and screened to extract the first operation status information from it. The first operation status information here includes multiple operation characteristic parameters with time marks, which are the speed or load of the fan. As long as any of the two indicators of speed or load is abnormal, the safety braking program will be activated to ensure the safe operation of the fan. For example, when the speed sensor detects that the speed of the fan suddenly exceeds the normal range, or the torque sensor finds that the load of the fan increases or decreases abnormally, the first operation status information acquisition module 10 will capture these abnormal conditions in time and record the corresponding operation characteristic parameters and the corresponding time marks. These time-stamped operating characteristic parameters can accurately reflect the time point and specific manifestations of abnormal situations, providing an important basis for subsequent analysis and processing. Through multi-source sensing equipment (torque sensor and speed sensor), the operating status information of the target external rotor fan can be obtained in real time and dynamically, ensuring timely understanding of the fan operation status, making the judgment of abnormal situations more accurate.

[0018] The target characteristic parameter acquisition module 20 is used to acquire the target characteristic parameter from the plurality of operation characteristic parameters with time marks, and determine whether the target characteristic parameter meets the parameter threshold constraint.

[0019] Specifically, the target characteristic parameter acquisition module interacts with the first operation status information acquisition module, receives multiple operation characteristic parameters with time marks, selects the operation characteristic parameters corresponding to the real time from these parameters, and determines them as the target characteristic parameters. For example, if the current time is 10:00, the module will obtain the operation characteristic parameters such as the fan speed and load at this time point. Compare the target characteristic parameters with the preset parameter threshold constraints. These parameter threshold constraints are set according to the normal operating range and safety requirements of the fan. For example, an upper limit and a lower limit may be set for the fan speed. If the speed exceeds or falls below this range, it is considered that the parameter threshold constraint is not met. When comparing, each dimension of the target characteristic parameter is analyzed one by one. If it is found that the target characteristic parameter does not meet the parameter threshold constraint in any dimension, it will be judged as not meeting the requirements. Through such specific and precise operations, the target characteristic parameter acquisition module can timely and accurately determine whether the operation status of the fan is normal at the real time, providing an important basis for subsequent control and decision-making, thereby ensuring the safe and stable operation of the fan and extending its service life.

[0020] The first target prediction operation parameter acquisition module 30 is used to predict and analyze the first operation parameter time series generated based on the multiple operation characteristic parameters with time marks if it meets the requirements, so as to obtain the first target prediction operation parameters under the target time.

[0021] Specifically, when the target characteristic parameter acquisition module determines that the target characteristic parameter meets the parameter threshold constraint, the module begins to work. The first operating parameter time series generated based on multiple operating characteristic parameters with time marks will be deeply predicted and analyzed. These operating characteristic parameters were previously acquired and recorded in real time through multi-source sensing equipment. They reflect the changes in the operating state of the target external rotor fan in the past period of time. The first operating parameter time series is a data sequence formed by sorting and arranging these operating characteristic parameters in chronological order. By analyzing this time series data, a prediction model based on a long short-term memory network (LSTM) can be used. LSTM is a neural network model suitable for processing time series data. The first operating parameter time series generated by multiple operating characteristic parameters with time marks is input into the LSTM model, and the LSTM model automatically learns the patterns and trends in these time series data. In the training stage, the LSTM model is trained using historical data, and the parameters of the model are adjusted so that the model can accurately predict future operating parameters. In the prediction stage, when the target characteristic parameters meet the parameter threshold constraints, the current operating parameter time series is input into the trained LSTM model, and the model predicts the first target predicted operating parameters at the target time based on the learned patterns and trends. For example, the operating characteristic parameters include speed and load, and the time mark is one data point per second. The LSTM model will learn the changing rules of speed and load over the past period of time, and predict the speed and load values ​​at the future target time based on the current trend. This helps to detect possible abnormal conditions of the fan in advance, take timely measures for safe braking or adjustment, and enhance the safety of fan operation.

[0022] The first target prediction operation parameter judgment module 40 is used to judge whether the first target prediction operation parameter meets the parameter threshold constraint.

[0023] Specifically, after obtaining the first target predicted operating parameters under the target time, these parameters will be input into the first target predicted operating parameter judgment module 40. First, the preset parameter threshold constraints are obtained. These constraints are usually formulated by the system based on factors such as the design specifications, safety standards and actual operating experience of the target external rotor fan. For example, for the speed, a normal range may be set, such as 1000-2000 revolutions per minute; for the load, there is a maximum allowable load value; for the temperature, there is also a safe temperature range. Each parameter in the first target predicted operating parameter is judged one by one. Taking the speed as an example, if the predicted speed is 1500 revolutions per minute, and the preset speed threshold range is 1000-2000 revolutions, then this speed parameter meets the constraint. If the predicted speed exceeds this range, such as reaching 2500 revolutions per minute, then it does not meet the constraint. Similar judgments will be made for other parameters. If all parameters are within their respective threshold ranges, the module will determine that the first target predicted operating parameter meets the parameter threshold constraint; otherwise, it will be determined that it does not meet. Through such a specific judgment process, the first target prediction operation parameter judgment module 40 can accurately determine whether the first target prediction operation parameter meets the parameter threshold constraint, and by judging whether the first target prediction operation parameter meets the parameter threshold constraint, potential safety hazards can be discovered in time. If the prediction parameter exceeds the threshold, it means that the wind turbine will have an abnormal operation, so that measures can be taken in advance to avoid accidents and ensure the safe operation of the wind turbine.

[0024] The predictive braking analysis result acquisition module 50 is used to retrieve the predictive braking decision rule to perform predictive braking analysis on the target outer rotor fan if it does not meet the requirements, so as to obtain the predictive braking analysis result.

[0025] Specifically, when it is determined that the first target predicted operating parameter does not meet the parameter threshold constraint, the module will respond quickly and immediately call the pre-set predictive braking decision rules. These rules are designed to predict the target time, that is, to determine the possible unsafe situation of the wind turbine at a future moment, and take corresponding countermeasures and measures in advance, such as conducting certain maintenance inspections. When performing predictive braking analysis, the module will comprehensively consider multiple relevant factors. First, the current operating status of the wind turbine will be analyzed in detail, including but not limited to the specific conditions of key parameters such as speed, load, temperature, and the degree to which they deviate from the normal range. Secondly, the module will refer to historical operating data to understand the performance and possible problems of the wind turbine under similar conditions. In addition, factors such as the overall performance and working environment of the wind turbine will be considered. Through a comprehensive analysis of these factors, the module can accurately assess the possible failure risks and their potential impacts that the wind turbine may face in the future. For example, if it is predicted that the speed of the wind turbine may rise sharply and exceed the safe range at a certain moment in the future, resulting in damage to wind turbine components or instability of the entire system, corresponding countermeasures will be formulated according to the decision rules. Finally, a detailed predictive braking analysis result will be obtained. The result will clearly indicate whether the fan needs to be braked, as well as the specific degree and timing of braking. At the same time, it will also include specific maintenance suggestions and measures for the fan to eliminate potential safety hazards in advance. Timely and effective prediction and analysis of abnormal conditions of the fan, and provide accurate and reliable basis for subsequent safe braking and maintenance operations, thereby maximizing the safety and stability of the fan and the entire system and extending its service life.

[0026] The safety braking module 60 is used to safely brake the target outer rotor fan according to the predictive braking analysis result.

[0027] Specifically, the safety braking module 60 is a key link in the entire system to ensure the safety of the target outer rotor fan. The safety braking module 60 will receive the predictive braking analysis results and interpret them. If the results indicate that the target outer rotor fan needs to be safely braked, the module will immediately start the braking procedure. During the braking process, the safety braking module 60 will interact with the control system of the fan and send a braking command to gradually slow down the operation of the fan until it stops. This process will ensure the smoothness and safety of the braking and avoid sudden braking from causing damage to the fan and related equipment. The status of the fan will also be continuously monitored to ensure the smooth progress of the braking process. If any abnormality is found during the braking process, the module will take corresponding measures in a timely manner, such as adjusting the braking force or issuing an alarm signal. In this way, the target outer rotor fan can be effectively safely braked according to the predictive braking analysis results, ensuring the safety of the fan during operation, preventing accidents caused by faults or abnormal conditions, and reducing the risk of accidents.

[0028] In a possible implementation, the first operation status information acquisition module 10 further includes:

[0029] The multi-source sensing device includes a torque sensor and a speed sensor.

[0030] Specifically, multi-source sensing equipment plays a vital role. It can fully and accurately obtain the operating status information of the target external rotor fan. Among them, the torque sensor is mainly used to measure the torque condition of the fan during operation. By monitoring the change of torque, the load condition of the fan can be understood and whether the fan is operating within the normal range. The speed sensor focuses on measuring the speed of the fan. The speed is a key parameter of the fan operation, which directly affects the performance and efficiency of the fan. The two sensors work together to provide multi-dimensional information about the operating status of the fan, providing strong support for subsequent data analysis and decision-making.

[0031] In a possible implementation, the first target prediction operation parameter acquisition module 30 further includes:

[0032] When the target characteristic parameter does not meet the parameter threshold constraint, a predetermined safety braking mechanism is triggered to perform safety braking on the target outer rotor fan.

[0033] Specifically, when the target characteristic parameter does not meet the parameter threshold constraint, this means that the operating state of the target external rotor fan has an abnormal situation, which will pose a threat to the safe operation of the fan. In order to avoid potential dangers, the system will immediately trigger the predetermined safety braking mechanism. In this case, the predetermined safety braking mechanism will start quickly and send a braking command to the target external rotor fan. This command will prompt the fan's braking system to start working, causing the fan to gradually slow down until it stops completely. By triggering the safety braking mechanism in a timely manner, the fan can be effectively prevented from continuing to operate in an abnormal state, thereby reducing the risk of accidents caused by faults or abnormal conditions and ensuring the safety of the fan and the entire system. At the same time, this also helps to reduce damage to the equipment and extend the service life of the fan.

[0034] In a possible implementation, the first target prediction operation parameter acquisition module 30 further includes:

[0035] The target outer rotor fan is dynamically monitored for operation by using multi-source monitoring equipment to obtain target operation monitoring information.

[0036] A first monitoring indicator among the predetermined monitoring indicators is extracted, and a first indicator monitoring data set of the first monitoring indicator is matched in the target operation monitoring information.

[0037] The mean value of the first indicator monitoring data set is taken as the first monitoring parameter of the first monitoring indicator.

[0038] The target monitoring parameter set formed based on the first monitoring parameter is input into the security classifier to obtain a target security classification result. The security classifier is an intelligent model trained based on the support vector machine principle.

[0039] The triggering verification of the predetermined safety braking mechanism is performed according to the target safety classification result.

[0040] Specifically, when it is initially determined that the target characteristic parameter value is not within the threshold, it is initially determined to be unsafe and braking is required. At this time, the large model intelligent judgment classification is combined with the monitoring data of more sources. If it is indeed unsafe, the safety braking is initiated. Multi-source monitoring equipment plays a key role. These devices include various sensors, such as temperature sensors, pressure sensors, vibration sensors, etc., which monitor the operating status of the target external rotor fan in real time from different angles. For example, the temperature sensor can monitor the temperature changes of various parts of the fan, the pressure sensor can measure the pressure inside the fan, and the vibration sensor can detect the vibration amplitude and frequency of the fan. These sensors transmit the monitored data to the monitoring system in real time, and the monitoring system integrates and analyzes these data to obtain the target operation monitoring information. This information includes parameters of the fan speed, temperature, pressure, vibration and other aspects, which can fully reflect the operating status of the fan. By dynamically monitoring the operation of the target external rotor fan, abnormal conditions in the operation of the fan can be discovered in time.

[0041] First, the predetermined monitoring indicators will be clarified. These indicators are determined in advance based on the operating characteristics, safety requirements, and previous experience of the target external rotor fan. For example, the predetermined monitoring indicators include speed, temperature, vibration amplitude, current, etc. Then the first monitoring indicator is selected, assuming that the first monitoring indicator here is the speed. Next, in the target operation monitoring information previously obtained through multi-source monitoring equipment, find data related to the first monitoring indicator of speed. These related data constitute the first indicator monitoring data set. For example, the target operation monitoring information contains the speed values ​​of the fan at different time points. Through screening and matching, these speed values ​​are extracted to form a first indicator monitoring data set specifically for the speed. Such operations can classify and refine complex target operation monitoring information, focus on specific monitoring indicators, and provide a targeted data basis for subsequent more in-depth analysis and processing.

[0042] The target monitoring parameter set is input into the safety classifier, which is a powerful intelligent model trained based on the principle of support vector machine (SVM). Support vector machine is a very effective classification algorithm in the field of machine learning. When training this safety classifier, a large amount of data covering various normal and abnormal wind turbine operation conditions is used. This data contains various different monitoring parameter combinations and corresponding safety classification labels, such as safe, mild risk, moderate risk, severe risk, etc. By learning from these massive data, the safety classifier gradually grasps the complex relationship and pattern between different monitoring parameter combinations and wind turbine safety status. When a new target monitoring parameter set is input into this trained safety classifier, it will quickly perform complex mathematical calculations and pattern matching on the input parameters. It will analyze the value of each parameter and the relationship between them, and compare them with the previously learned patterns. Finally, the safety classifier will output a clear target safety classification result. The result is a specific classification label, for example, "safe" means that the wind turbine is currently in good operating condition and no special measures are needed; "mild risk" means that some minor adjustments or increased monitoring frequency are needed; "moderate risk" indicates the need for more in-depth inspections and possible preventive maintenance; and "serious risk" will directly trigger emergency safety measures, such as immediate shutdown and repair. Through this sophisticated and accurate classification process, the powerful capabilities of the intelligent model can be used to efficiently and accurately evaluate and classify the operational safety of the wind turbine, providing key decision-making support for ensuring the stable operation of the wind turbine and avoiding potential safety accidents.

[0043] If the target safety classification result shows that the fan is in a high-risk or dangerous state, such as severe overspeed, excessive temperature, abnormally strong vibration, etc., this meets the conditions for triggering the predetermined safety braking mechanism. The system will further verify the accuracy of the relevant monitoring data and classification results to ensure that it is not a misjudgment. If it is verified that the fan is indeed in a dangerous state that requires emergency braking, the system will activate the predetermined safety braking mechanism. This includes a series of operations such as quickly cutting off the power supply, starting the mechanical braking device, and adjusting the working mode of the fan to ensure that the fan can stop running as soon as possible to avoid possible accidents and equipment damage. On the contrary, if the target safety classification result shows that the fan is in a lower risk or safe state, then the predetermined safety braking mechanism will not be triggered, and the fan will continue to operate normally or only take some milder monitoring and adjustment measures. In short, the triggering verification of the predetermined safety braking mechanism based on the target safety classification result is a rigorous and precise process aimed at ensuring the safe operation of the target external rotor fan to the greatest extent.

[0044] In a possible implementation, the first target prediction operation parameter acquisition module 30 further includes:

[0045] The predetermined monitoring indicators include temperature, vibration, current, pressure, flow, sound and acceleration.

[0046] Specifically, in the monitoring of the target external rotor fan, the predetermined monitoring indicators cover several key aspects. The temperature indicator reflects the heating of each component of the fan, which can be used to determine whether there is an overheating fault and avoid component damage. The vibration indicator can reflect the smoothness of the fan operation. Abnormal vibration means loose components, imbalance or other mechanical failures. The current indicator helps to understand the energy consumption and load of the fan. Abnormal current indicates motor failure or overload. The pressure indicator reflects the pressure state inside the fan, which is crucial to evaluating the working efficiency of the fan and the stability of the system. The flow indicator can indicate the ability of the fan to transport gas or liquid. Abnormal flow indicates pipeline blockage or deterioration of fan performance. The sound indicator can determine whether the fan has abnormal noise through audio characteristics, thereby discovering potential faults. The acceleration indicator can more keenly capture the changes in the fan's motion state and detect rapid abnormal movements in time. Combining these predetermined monitoring indicators, we can fully and accurately understand the operating status of the target external rotor fan, discover and solve problems in time, and ensure the safe, stable and efficient operation of the fan.

[0047] In a possible implementation, the first target prediction operation parameter acquisition module 30 further includes:

[0048] The historical safety braking records of the same type of fans as the target outer rotor fan are obtained, and a first historical record in the historical safety braking records is extracted, wherein the first historical record has an identification of whether the first historical record is safe or not.

[0049] A first historical monitoring parameter set is obtained by matching and analyzing the first historical operation monitoring information in the first historical record based on the predetermined monitoring indicator.

[0050] Based on the support vector machine principle, supervised learning and verification are performed on a first training data set formed by the first historical monitoring parameter set and the first historical safety identification to obtain the safety classifier.

[0051] Specifically, all safety braking records of fans of the same type as the target external rotor fan in the past operation process are widely collected. These records record in detail the braking conditions and related data of fans of the same type under various working conditions. The first historical record is selected from the numerous historical safety braking records. This first historical record has a clear mark about its safety status at the time, that is, whether it is a "safe" braking situation or an "unsafe" braking situation. If this mark shows "safe", it may mean that during the braking process at that time, fans of the same type did not have any parts damage, performance degradation, abnormal noise or vibration and other adverse conditions. The entire braking operation smoothly and effectively ensured the safe stop of the fan, and did not have a negative impact on the subsequent operation and service life of the fan. On the contrary, if the mark shows "unsafe", it means that during that braking, there were problems such as excessive wear of the brake pads, breakage of parts of the braking system, current shock caused by the sudden stop of the fan damaging related electrical components, or loosening of connecting parts due to excessive braking. These problems not only affected the braking effect at the time, but also may bury safety hazards for the subsequent operation of the fan. By carefully extracting and deeply studying such first historical records with clear safety or non-safety identification, the safety braking performance of the target external rotor fan can be accurately evaluated and possible problems can be predicted.

[0052] According to the previously determined predetermined monitoring indicators, such as temperature, vibration, current, etc., the first historical operation monitoring information contained in the first historical record is compared and analyzed in detail. The first historical operation monitoring information contains a large amount of various data about the operation of the fan. According to the type of predetermined monitoring indicators, the data corresponding to the indicator is screened out from these complex data. For example, if the predetermined monitoring indicator includes temperature, the system will search and extract specific values ​​related to temperature in the first historical operation monitoring information; for vibration indicators, data related to vibration amplitude, frequency, etc. will be extracted. Through such screening and extraction, a set of data corresponding to each predetermined monitoring indicator is integrated together to form a first historical monitoring parameter set. For example, the first historical monitoring parameter set contains a temperature value of 80°C, a vibration amplitude of 5 mm per second, a current of 10 amperes, etc. at a certain moment. Through such a fine and comprehensive matching analysis process, a targeted and valuable first historical monitoring parameter set is successfully obtained from the complex first historical operation monitoring information, which provides solid data support for subsequent research, analysis, model training, etc., and helps to understand and evaluate the operation status and safety performance of the fan more deeply and accurately.

[0053] Support Vector Machine (SVM) is an algorithm with wide application and excellent performance in the field of machine learning. Based on the principle of support vector machine, supervised learning and verification of the first training data set are performed to obtain a safety classifier. First, the first training data set has been constructed. This data set consists of the first historical monitoring parameter set and the corresponding first historical safety or non-safety identification. Each data sample is like an information-rich data packet. It not only contains the specific values ​​of various detailed monitoring parameters such as temperature, vibration, current, etc., but also comes with an extremely critical identification, which clearly indicates whether the operation of the wind turbine is "safe" or "unsafe" under the circumstances at that time. When supervised learning is started based on the principle of support vector machine, its goal is to find an optimal separation interface, that is, a hyperplane, in this multidimensional space composed of many data points. This hyperplane should be able to separate the data points marked as "safe" and the data points marked as "unsafe" in different areas with the greatest accuracy. In order to find this ideal hyperplane, the position and influence of each data point in space will be carefully measured. Those data points that are located near the boundary of the two types of data and play a key role in determining the hyperplane are called support vectors. During the calculation process, SVM will try to maximize the distance between the hyperplane and the support vector. The advantage of this is that the hyperplane has better robustness and generalization ability for the classification of new data points, and can make more accurate judgments when facing unseen data. While learning, a part of the data in the first training data set is retained and does not participate in the learning process. This part of the data will be used to test the performance of the model. When the learning phase is completed and the preliminary model is obtained, these hidden test data are input into the model. The model will classify and predict these new data based on the rules and patterns learned before, and give a judgment of "safe" or "unsafe". Compare the prediction results of the model with the original real safety labels of these test data. By calculating a series of evaluation indicators, such as accuracy, recall rate, F1 value, etc., the performance of the model on the classification task is measured. If the performance of the model on the test data is not ideal, there may be overfitting (overfitting the training data, resulting in poor adaptability to new data) or underfitting (not fully learning the rules in the data). At this time, you need to adjust the parameters of the SVM algorithm, such as the type of kernel function, the size of the penalty coefficient, etc., or try to preprocess the data, such as feature scaling, data cleaning, etc., and then re-learn and test. This repeated adjustment and optimization process will continue until the model's performance on the test data reaches the preset standard, such as an accuracy rate of more than 90%, or other strict indicators set according to specific application requirements.Finally, a safety classifier was obtained, which has the ability to quickly and accurately determine whether the fan operation is safe based on the input fan monitoring parameters, providing strong technical support for ensuring the reliable operation of the fan.

[0054] In a possible implementation, the first target prediction operation parameter acquisition module 30 further includes:

[0055] A first operation scatter plot is generated based on the first operation parameter time series, wherein the first operation scatter plot includes a plurality of scatter points.

[0056] A first scattered point group is extracted from the plurality of scattered points by using the consistency principle of random sampling, and a polynomial regression fitting is performed on the first scattered point group to obtain a first spline curve.

[0057] The plurality of scattered points are compared with the first scatter point group to obtain a first verification scatter point group, and a fitness test is performed on the first spline curve based on the first verification scatter point group to obtain a first test result.

[0058] When the first test result satisfies the test constraint, the first spline curve is used as a target curve, and a target fitting formula of the target curve is obtained.

[0059] The target time is input into the target fitting formula to obtain the first target prediction operation parameter.

[0060] Specifically, a series of first operating parameters arranged in chronological order are obtained. These parameters are numerical values ​​related to the operating status of a system, equipment or process, such as the change value of temperature over time, the change value of speed over time, etc. Then, according to the time sequence of these parameters and the corresponding numerical values, the corresponding points are drawn in a two-dimensional coordinate system. The horizontal axis usually represents time, and the vertical axis represents the specific numerical value of the operating parameter. Since there are multiple parameter values ​​at different time points, multiple scattered points will be generated in the coordinate system, and these scattered points together constitute the first operating scatter plot. For example, if the working current is measured at different time points, the time is 1 second, 2 seconds, 3 seconds..., and the corresponding current values ​​are 5 amperes, 7 amperes, 4 amperes..., then points such as (1, 5) (2, 7) (3, 4) will be marked in the coordinate diagram in sequence to form a scatter plot. Through such a scatter plot, the changing trend and distribution of the operating parameters over time can be intuitively observed.

[0061] The random sampling consistency principle is a method for estimating model parameters from data containing outliers. Among multiple scattered points, when using the random sampling consistency principle, some scattered points are randomly selected as a subset (i.e., the first scattered point group), and then a mathematical model is attempted to be established based on this subset (in this case, polynomial regression is performed). Among the multiple randomly selected subsets, the subset that best fits most of the scattered points is selected as the final first scattered point group. After obtaining the first scattered point group, it is fitted using the polynomial regression method. Polynomial regression is to find a suitable polynomial function so that this function is as close to these scattered points as possible. After fitting, the result is the first spline curve. This curve can reflect the overall distribution trend and law of the scattered points to a certain extent.

[0062] There are multiple scatter points initially generated, which represent the operating parameter data of the object under study at different times or conditions. Then, the first scatter point group is extracted by using the random sampling consistency principle. Next, the original multiple scatter points are compared with the scatter points in the first scatter point group one by one. For each original scatter point, if it is not in the first scatter point group, it is selected and put into the first verification scatter point group. For example, assuming that there are 500 scatter points initially, the first scatter point group extracted by random sampling contains 300 scatter points, and the remaining 200 scatter points constitute the first verification scatter point group. After obtaining the first verification scatter point group, the fitness of the first style curve should be tested based on this verification scatter point group. The specific test method is: for each scatter point in the first verification scatter point group, calculate its vertical distance to the first style curve. By setting a distance limit, comparing the vertical distance with the distance limit, dividing the scattered points into inner points or outer points, and taking the ratio of the number of inner points to the number of outer points as the first test result, we can finally get a first test result that can fully reflect the fitness of the first spline curve. This result is a specific numerical indicator.

[0063] When the first test result is obtained, it will be compared with the pre-set test constraints. Assuming that the test constraint stipulates that the value of the first test result must be greater than 5, and the calculated first test result is 7.5, it is considered that the first test result meets the test constraint. Once the first test result meets these strict test constraints, the first spline curve obtained previously is determined as the target curve. This means that this curve is recognized as being able to effectively represent the trends and patterns of the data. Subsequently, in order to more conveniently apply and analyze this target curve, its target fitting formula will be obtained. This target fitting formula is usually a mathematical expression, such as a polynomial function of the form y=ax³+bx²+cx+d. Among them, the coefficients a, b, c, d, etc. are specific values ​​accurately calculated through the previous fitting process.

[0064] The target fitting formula is determined, which can reflect the changing law of a certain operating parameter over time and other factors. Then, there is a specific target time value. Substitute this target time value into the target fitting formula, and by performing corresponding mathematical calculations, the corresponding operating parameter prediction value at this specific time can be obtained. This prediction value is the first target prediction operating parameter. For example, if the target fitting formula is y=2x²+3x+1, and the x value corresponding to the target time is 5, then substitute x=5 into this formula: y=2×5²+3×5+1=66, so when the target time is 5, the value of the first target prediction operating parameter y is 66. Through a series of operations such as generating a scatter plot based on the first operating parameter time series, extracting the scatter group fitting curve, and performing fitness tests, the target fitting formula is obtained when the test result meets the constraints, and the target time is input into it to obtain the first target prediction operating parameter, so as to achieve effective analysis and prediction of the operating parameters.

[0065] In a possible implementation, the first target prediction operation parameter acquisition module 30 further includes:

[0066] A first verification scatter point in the first verification scatter point group is extracted, and a first vertical distance from the first verification scatter point to the first spline curve is obtained.

[0067] When the first vertical distance is within the distance limit, the first verification scattered point is added to the inner point set; when the first vertical distance is not within the distance limit, the first verification scattered point is added to the outer point set.

[0068] The ratio of the number of scattered points of the inner point set to the number of scattered points of the outer point set is used as the first test result.

[0069] Specifically, a scatter point is taken out from the first verification scatter point group one by one, and it is defined as the first verification scatter point, and then the vertical distance from the first verification scatter point to the first spline curve is calculated, and this distance is called the first vertical distance. Through such an operation, the deviation between the first verification scatter point and the first spline curve can be quantified.

[0070] The first vertical distance from the first test scatter point to the first spline curve is obtained. Then, the first vertical distance is compared with the pre-set distance limit. If the first vertical distance is less than or equal to the distance limit, it means that the first test scatter point is relatively close to the first spline curve and meets the preset proximity standard. At this time, the first test scatter point will be added to the inner point set. On the contrary, if the first vertical distance is greater than the distance limit, it indicates that the first test scatter point deviates from the first spline curve to a large extent and does not meet the preset proximity standard. In this case, the first test scatter point will be added to the outer point set. For example, the distance limit is set to 5 unit lengths. If the first vertical distance from a first test scatter point to the first spline curve is 4, then it is added to the inner point set; if the first vertical distance is 6, it is added to the outer point set. The number of scattered points in the inner point set and the number of scattered points in the outer point set are calculated respectively, and then the number of scattered points in the inner point set is divided by the number of scattered points in the outer point set. The quotient obtained is the so-called first test result. For example, if there are 30 scattered points in the inner point set and 10 scattered points in the outer point set, then the first test result is 30÷10=3. This ratio can be used to evaluate the fitting effect of the first spline on the data or the distribution characteristics of the data. The larger the ratio of the number of scattered points in the inner point set to the outer point set, the better. A larger ratio means that more test scattered points are included in the inner point set, that is, more scattered points are closer to the first spline, which indicates that the first spline fits the data better and can better represent the overall trend of the data.

[0071] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0072] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

[0073] This specification and the drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.

Claims

1. External rotor fan safety braking system based on automatic sensing, It is characterized in that include: A first operation status information acquisition module, the first operation status information acquisition module is used to dynamically acquire target operation status information of the target outer rotor fan through a multi-source sensing device, and extract first operation status information from the target operation status information, wherein the first operation status information includes a plurality of operation characteristic parameters with time marks; A target characteristic parameter acquisition module, the target characteristic parameter acquisition module is used to acquire the target characteristic parameter from the plurality of operation characteristic parameters with time marks, and determine whether the target characteristic parameter meets the parameter threshold constraint; A first target prediction operation parameter acquisition module, the first target prediction operation parameter acquisition module is used to predict and analyze the first operation parameter time series generated based on the multiple operation characteristic parameters with time marks if it meets the requirements, so as to obtain the first target prediction operation parameter under the target time; A first target prediction operation parameter judgment module, the first target prediction operation parameter judgment module is used to judge whether the first target prediction operation parameter meets the parameter threshold constraint; A predictive braking analysis result acquisition module, wherein if the condition does not meet the requirement, the predictive braking analysis result acquisition module is used to retrieve the predictive braking decision rule to perform predictive braking analysis on the target outer rotor fan to obtain a predictive braking analysis result; A safety braking module, the safety braking module is used to safely brake the target outer rotor fan according to the predictive braking analysis result; Wherein, the system further comprises: When the target characteristic parameter does not meet the parameter threshold constraint, triggering a predetermined safety braking mechanism to perform safety braking on the target outer rotor fan; Wherein, the system further comprises: Performing dynamic operation monitoring on the target outer rotor fan by multi-source monitoring equipment to obtain target operation monitoring information; Extracting a first monitoring indicator from the predetermined monitoring indicators, and matching a first indicator monitoring data set of the first monitoring indicator in the target operation monitoring information; Taking the mean value of the first indicator monitoring data set as the first monitoring parameter of the first monitoring indicator; Inputting a target monitoring parameter set formed based on the first monitoring parameter into a security classifier to obtain a target security classification result, wherein the security classifier is an intelligent model trained based on the support vector machine principle; Performing trigger verification of the predetermined safety braking mechanism according to the target safety classification result; Wherein, the system further comprises: Obtaining historical safety braking records of fans of the same type as the target outer rotor fan, and extracting a first historical record from the historical safety braking records, wherein the first historical record has an identification of whether the first historical record is safe or not; Obtaining a first historical monitoring parameter set based on matching analysis of the predetermined monitoring indicator in the first historical operation monitoring information in the first historical record; Based on the support vector machine principle, supervised learning and verification are performed on a first training data set formed by the first historical monitoring parameter set and the first historical safety identification to obtain the safety classifier.

2. According to claim 1, the outer rotor fan safety braking system based on automatic induction, It is characterized in that The multi-source sensing device includes a torque sensor and a speed sensor.

3. According to claim 1, the outer rotor fan safety braking system based on automatic induction, It is characterized in that The predetermined monitoring indicators include temperature, vibration, current, pressure, flow, sound and acceleration.

4. According to claim 1, the outer rotor fan safety braking system based on automatic induction, It is characterized in that include: generating a first operation scatter plot based on the first operation parameter time series, wherein the first operation scatter plot includes a plurality of scatter points; Extracting a first scattered point group from the plurality of scattered points by using a random sampling consistency principle, and performing polynomial regression fitting on the first scattered point group to obtain a first spline curve; Comparing the plurality of scattered points with the first scatter point group to obtain a first verification scatter point group, and performing a fitness test on the first spline curve based on the first verification scatter point group to obtain a first test result; When the first test result satisfies the test constraint, taking the first spline curve as a target curve, and obtaining a target fitting formula of the target curve; The target time is input into the target fitting formula to obtain the first target prediction operation parameter.

5. According to claim 4, the outer rotor fan safety braking system based on automatic induction, It is characterized in that include: Extracting a first verification scatter point from the first verification scatter point group, and acquiring a first vertical distance from the first verification scatter point to the first spline curve; When the first vertical distance is within the distance limit, adding the first check scattered point to the inner point set, and when the first vertical distance is not within the distance limit, adding the first check scattered point to the outer point set; The ratio of the number of scattered points of the inner point set to the number of scattered points of the outer point set is used as the first test result.

Images