Blower maintenance system based on fault analysis

Abstract

The invention relates to the technical field of blower maintenance, and discloses a blower maintenance system based on fault analysis, which comprises a fault prediction module, a test maintenance module and a maintenance auxiliary module, wherein the fault prediction module is used for predicting the occurrence probability of each potential fault of the air feeder; the test and maintenance module is used for formulating a test and maintenance plan of the air feeder according to the occurrence probability of each potential fault of the air feeder; and the maintenance auxiliary module is used for assisting in executing the test maintenance plan, including providing test maintenance suggestions, assisting spare part allocation and recording test maintenance data. According to the invention, the intelligent and predictable maintenance of the air feeder is realized, the equipment failure rate is reduced, the test and maintenance efficiency is improved, and the operation cost is saved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of air supply machine maintenance, and particularly relates to an air supply machine maintenance system based on fault analysis. BACKGROUND

[0002] With the acceleration of industrialization process and the continuous improvement of modern production management level, as an indispensable key equipment in various industrial production environments, air supply machines play a vital role in many fields such as electric power, metallurgy, chemical industry, environmental protection and the like. The air supply machine is responsible for delivering necessary air or other gas media to the system to ensure the stable operation of the entire process flow. However, due to its long-time high-load working characteristics, the maintenance and test maintenance of the air supply machine are particularly important, which not only relates to the extension of the service life of the equipment, but also directly affects the continuity and safety of the overall production.

[0003] At present, the traditional air supply machine operation and maintenance management mode generally has the following problems: the current operation and maintenance management of the air supply machine still depends on regular inspection and emergency response after sudden failure, lacks scientific and effective fault prediction means, even if some online monitoring systems are implemented, but they are often limited to single parameter threshold alarm, and cannot comprehensively and accurately capture the complex potential fault mode and its development dynamics inside the air supply machine; the existing test maintenance plan is made based on fixed period or cumulative running time, and the actual running state, working condition change and component wear degree of the air supply machine are not fully considered, which may lead to over-maintenance or delayed maintenance, increase unnecessary cost expenditure, and may also miss the best maintenance opportunity, increase the risk of major failure; although modern air supply machines are generally equipped with data acquisition systems, the massive operation data are not fully utilized, especially in the aspects of multi-dimensional and deep fault diagnosis and preventive maintenance, and the data-driven intelligent management strategy still has great development space.

[0004] The patent application with publication number CN112697424A discloses a fan gear lubrication system fault diagnosis method based on an improved decision tree, which includes the following steps: S10. Collecting multiple sets of generator set gear oil data; S20. Normalizing the gear oil data in step S10, and dividing the normalized data into a training set and a test set according to the proportion; S30. Improving the decision tree model, and training the improved decision tree model with the training set in step S20 to obtain a decision tree integrated model; inputting the test set in step S20 into the decision tree integrated model to verify the training result; then inputting the to-be-tested data into the trained decision tree for prediction, and evaluating the fault level according to the typical wear failure of the fan gear box. The invention can improve the prediction accuracy, reduce misjudgment and omission, and predict and judge possible hidden faults; but it still has the problems of the background technology: it fails to fully consider the actual operation state of the air supply fan, the change of working conditions, and the wear degree of the components to develop a test and maintenance plan, which may cause delay in maintenance.

[0005] The information disclosed in this BACKGROUND section is only intended to increase an understanding of the general context in which the present application can be practiced and is not admitted to be prior art. SUMMARY

[0006] The technical problem to be solved by the present application is to overcome the defects of the prior art, provide an air supply fan maintenance system based on fault analysis, realize the digitization, intelligentization and predictability of air supply fan operation and maintenance management, reduce the equipment failure rate, improve the maintenance efficiency, save the operation cost, and ensure the continuity and safety of industrial production.

[0007] To solve the above technical problems, the present application provides the following technical solutions: The air supply fan maintenance system based on fault analysis comprises a fault prediction module, a test and maintenance module, and a maintenance assistance module; wherein: The fault prediction module is used to predict the occurrence probability of each potential fault of the air supply fan, and the method is to collect the operation parameters of the air supply fan and train a multi-modal health assessment model, and to perform fault prediction through the multi-modal health assessment model; The test and maintenance module is used to develop a test and maintenance plan for the air supply fan according to the occurrence probability of each potential fault of the air supply fan, including determining the test and maintenance time and the test and maintenance task sequence; The maintenance assistance module is used to assist in executing the test and maintenance plan, including providing test and maintenance suggestions, assisting in spare parts allocation, and recording test and maintenance data.

[0008] As a preferred scheme of the air supply machine maintenance system based on fault analysis, wherein: the fault prediction module comprises a data acquisition unit, a feature extraction unit, a preprocessing unit, a model training unit, and a fault prediction unit; the data acquisition unit is configured to acquire air supply machine data; The feature extraction unit extracts operation parameters of the air supply machine based on the air supply machine data. The preprocessing unit is configured to preprocess the operation parameters. The model training unit is configured to establish the multi-modal health assessment model and train, optimize, and test the model. The fault prediction unit is configured to deploy and apply the multi-modal health assessment model to predict faults of the air supply machine based on the preprocessed operation parameters.

[0009] As a preferred scheme of the air supply machine maintenance system based on fault analysis, wherein: the air supply machine data comprises vibration signals, noise signals, energy consumption data, air supply volume data, rotation speed data, torque data, and air supply machine temperature data.

[0010] As a preferred scheme of the air supply machine maintenance system based on fault analysis, wherein: the operation parameters comprise amplitude peak values, vibration fundamental frequencies, distortion degrees, noise intensities, average powers, energy consumption efficiency coefficients, average rotation speeds, maximum torques, and temperature fluctuation indexes; wherein: The amplitude peak value is the maximum vibration amplitude of the vibration signal. The vibration fundamental frequency is the maximum single frequency component on a frequency spectrum of the vibration signal. The distortion degree is calculated according to the following formula: ; wherein D represents the distortion degree, represents the amplitude of the fundamental frequency of the vibration signal. represents the amplitude of the i-th harmonic of the vibration signal, and i is in the range of 2, 3, …, m, where m is the highest order of the harmonic participating in the calculation. The noise intensity is represented by a sound pressure level and is calculated based on the noise signal. The average power is the energy consumed by the air supply machine per unit time. The energy consumption efficiency coefficient is calculated according to the following method: Record the energy consumption data within a period of time and fit the energy consumption curve, with the horizontal coordinate representing time and the vertical coordinate representing the energy consumption of the air supply machine. Record the air supply volume data within the same period of time and fit the air supply volume curve, with the horizontal coordinate representing time and the vertical coordinate representing the air supply volume of the air supply machine. Select n time points, read the energy consumption and air supply volume of the air supply fan at each time point through the energy consumption curve and the air supply volume curve, and calculate the energy consumption efficiency coefficient, the formula is as follows: ; Wherein, The energy consumption efficiency coefficient is represented by E, The energy consumption of the air supply fan at the kth time point is represented by Ek, The air supply volume of the air supply fan at the kth time point is represented by Qk; The average rotating speed is the average value of the rotating speed data of the main shaft of the air supply fan; The maximum torque is the maximum value in the torque data of the air supply fan; The calculation method of the temperature fluctuation index comprises: extracting continuous N air supply fan temperature values from the air supply fan temperature data at a fixed time interval, and calculating the temperature fluctuation index, the formula is as follows: ; Wherein, The temperature fluctuation index is represented by T, The qth air supply fan temperature value is represented by Tq, The q-1th air supply fan temperature value is represented by Tq-1, and q takes the value range of 2, 3, …, N; The time interval between any two continuous air supply fan temperature values is represented by Δt.

[0011] As a preferred scheme of the air supply fan maintenance system based on fault analysis, wherein: the calculation method of the temperature fluctuation index further comprises: extracting M air supply fan temperature values from the air supply fan temperature data, and calculating the temperature fluctuation index, the formula is as follows: ; Wherein, The jth air supply fan temperature value is represented by Tj, The j-1th air supply fan temperature value is represented by Tj-1, and j takes the value range of 2, 3, …, M; The time interval between and is calculated according to the following formula: ; Wherein, The basic sampling interval is represented by Δt0; The sampling correction factor is represented by f; The j-2th air supply fan temperature value is represented by Tj-2.

[0012] As a preferred scheme of the air supply machine maintenance system based on fault analysis, the multi-modal health assessment model is one of a convolutional neural network, a long short-term memory network, a gated recurrent unit and a deep feedforward network, the input is preprocessed operation parameters, and the output is the occurrence probability of each potential fault of the air supply machine.

[0013] As a preferred scheme of the air supply machine maintenance system based on fault analysis, the test maintenance module comprises an environment monitoring unit, a maintenance time unit and a maintenance task unit. The environment monitoring unit is used for measuring the environmental temperature and humidity of the working environment of the air supply machine in real time. The maintenance time unit is used for calculating the test maintenance time. The maintenance task unit is used for sorting the test maintenance tasks.

[0014] As a preferred scheme of the air supply machine maintenance system based on fault analysis, the calculation formula of the test maintenance time is as follows: ; Wherein, represents the time interval between the next test maintenance time and the current time; represents the standard reference period; represents the current time, represents the last test maintenance time; F represents the comprehensive risk index; represents the average environmental temperature of the day, represents the standard environmental temperature of the working air supply machine; represents the average environmental humidity of the day, represents the standard environmental humidity of the working air supply machine; 、 are weight coefficients; The calculation method of the comprehensive risk index F is as follows: set a risk threshold for each fault type of the air supply machine; count the occurrence probability of each potential fault of the air supply machine output by the multi-modal health assessment model, and calculate the comprehensive risk index F, the formula is as follows: ; Wherein, represents the occurrence probability of the pth potential fault whose occurrence probability exceeds the corresponding risk threshold, represents the risk threshold of the corresponding potential fault of ; Q is the number of types of potential faults whose occurrence probability exceeds the risk threshold.

[0015] As a preferred scheme of the air supply machine maintenance system based on fault analysis, the test maintenance module comprises an environment monitoring unit, a maintenance time unit and a maintenance task unit. The method for sequencing the test maintenance tasks is as follows: obtaining Q kinds of faults with occurrence probabilities exceeding a risk threshold, and arranging a test maintenance task for each kind of fault; assigning a priority to the corresponding test maintenance task based on the occurrence probability of each potential fault; wherein, The higher the occurrence probability of the potential fault, the higher the priority of the corresponding test maintenance task; sequencing the test maintenance tasks in order from high to low priority, and generating a test maintenance task sequence.

[0016] As a preferred scheme of the air supply machine maintenance system based on fault analysis, wherein: the maintenance task unit is further configured with an adjustment strategy for adjusting the sequencing of the test maintenance tasks; the adjustment strategy specifically includes: setting a prediction period; at the beginning of each prediction period, triggering the prediction of the occurrence probability of each potential fault; calculating the growth rate of the occurrence probability of each potential fault based on the occurrence probability of each potential fault in the last at least two prediction periods; The maintenance task unit is further configured with a probability risk threshold; if the growth rate of the occurrence probability of any potential fault is greater than the probability risk threshold, the priority of the corresponding test maintenance task is increased; If the priority of at least one test maintenance task changes, resequence the test maintenance tasks and update the test maintenance task sequence.

[0017] As a preferred scheme of the air supply machine maintenance system based on fault analysis, wherein: the adjustment strategy further includes: determining the associated operating parameters of each potential fault; the associated operating parameters of any potential fault include at least one of the operating parameters; The maintenance task unit is further configured with a reference interval for each operating parameter; the abnormal operating parameters are identified based on the reference interval; If there is an abnormal operating parameter in the associated operating parameters of any potential fault, the priority of the corresponding test maintenance task is adjusted, specifically including: calculating the proportion of the abnormal operating parameter in the associated operating parameters; increasing the priority based on the proportion, and the larger the proportion, the greater the adjusted priority.

[0018] As a preferred scheme of the air supply machine maintenance system based on fault analysis, wherein: the maintenance auxiliary module includes a spare parts allocation unit, a maintenance guidance unit, a maintenance data unit, and a visualization unit; wherein: The maintenance guidance unit is used to provide test maintenance suggestions; the method is to receive the test maintenance task sequence generated by the maintenance task unit, and provide test maintenance suggestions for each test maintenance task; The spare parts allocation unit is used for generating a spare parts allocation list; the method is to count the required spare parts and tools of all test maintenance tasks based on the maintenance suggestions provided by the maintenance guide unit for each test maintenance task, and generate a spare parts allocation list.

[0019] The maintenance data unit is used for recording test maintenance data. The visualization unit is used for visualizing the running parameters of the air blower, the fault prediction results, the test maintenance task progress, and the maintenance suggestions of each test maintenance task.

[0020] Compared with the prior art, the present application has the following beneficial effects: The multi-modal health assessment model is used for fault prediction of the air blower, and by collecting and analyzing the running parameters of the air blower in real time, the occurrence probability of various potential faults can be effectively identified and predicted, and the accuracy and foresight of fault early warning are significantly improved, so that equipment maintenance is changed from passive response to active prevention.

[0021] According to the fault occurrence probability output by the fault prediction module, the test maintenance module can formulate a more scientific and reasonable test maintenance plan, including accurate calculation of the test maintenance time point and optimization of the test maintenance task sequence. At the same time, the addition of the environmental monitoring unit further improves the adaptability of the plan, ensuring that the plan can be adjusted in time according to the on-site environmental conditions. In summary, the present application integrates big data analysis, artificial intelligence algorithms and Internet of Things technology, and comprehensively improves the intelligent level of air blower maintenance, reduces the equipment failure rate, and improves the efficiency of test maintenance. BRIEF DESCRIPTION OF DRAWINGS

[0022] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor. Among them: Figure 1 The structure diagram of the air blower maintenance system based on fault analysis provided by the present application; Figure 2 The working mode diagram of the air blower maintenance system based on fault analysis provided by the present application. DETAILED DESCRIPTION

[0023] The technical solutions of the present application will be described in detail below by means of the drawings and specific embodiments. It should be understood that the embodiments of the present application and the specific features in the embodiments are detailed descriptions of the technical solutions of the present application, and are not limitations of the technical solutions of the present application. In the case of no conflict, the technical features in the embodiments of the present application and the embodiments can be combined with each other.

[0024] The embodiment introduces a blower maintenance system based on fault analysis, referring to Figure 1 The system comprises a fault prediction module, a test maintenance module and a maintenance assistance module.

[0025] The fault prediction module is used to predict the occurrence probability of each potential fault of the blower; the method is to collect the operation parameters of the blower and train a multi-modal health assessment model, and to make fault prediction through the multi-modal health assessment model; including a data acquisition unit, a feature extraction unit, a preprocessing unit, a model training unit and a fault prediction unit; The data acquisition unit is used to acquire blower data; the blower data includes blower vibration signal, noise signal, energy consumption data, air supply volume data, rotating speed data, torque data and blower temperature data; a plurality of high-precision sensors are deployed on the blower to acquire multi-element data such as vibration signal and noise signal in real time. These sensors have high precision and stability, and can monitor the equipment operation state stably for a long time. For example, sensors are installed at the shaft end of the power source of the blower, i.e. the motor, to monitor the vibration caused by unbalance of the motor rotor, bearing wear, shaft misalignment, etc.; sensors are installed at the coupling part between the motor and the blower to monitor the vibration caused by misalignment, wear and fracture of the coupling, etc.

[0026] The feature extraction unit extracts the operation parameters of the blower based on the blower data; the operation parameters include amplitude peak value, vibration fundamental frequency, distortion degree, noise intensity, average power, energy consumption efficiency coefficient, average rotating speed, maximum torque and temperature fluctuation index.

[0027] The preprocessing unit is used to preprocess the operation parameters; the preprocessing includes abnormal value detection and replacement, data standardization processing.

[0028] The model training unit is used to establish the multi-modal health assessment model and to train, optimize and test the model; the multi-modal health assessment model is one of convolutional neural network, long short-term memory network, gated recurrent unit and deep feedforward network, the input is the preprocessed operation parameters, and the output is the occurrence probability of each fault.

[0029] Collect historical fault data, each piece of historical fault data including fault type and operation parameters one week before the fault occurs; encode the fault type and preprocess the operation parameters; divide the historical fault data into training set, validation set and test set, the ratio being 8:1:1; First, the training set is used for the training of the multi-modal health assessment model; that is, the optimization algorithm such as gradient descent is performed on the data set to update the model parameters so that the model can fit the rules in the data as much as possible; then the validation set is used for the tuning of the multi-modal health assessment model; try different architectures (such as different number of layers, number of nodes, activation function, etc.), use the performance (such as accuracy, AUC, logloss, etc.) on the validation set as the evaluation standard, and select the optimal model; after determining the model architecture, adjust the hyperparameters (such as learning rate, regularization strength, batch size, etc.) of the model, observe the performance changes on the validation set, and select the best hyperparameter combination; finally, the test set is used for the testing of the multi-modal health assessment model; after the model training and tuning are completed, the test set is used to evaluate the performance of the model on new data that does not participate in the training and tuning process, so as to obtain the expected performance of the model in actual application. The tested model is saved to the fault prediction unit.

[0030] The fault prediction unit is configured to deploy the multi-modal health assessment model and perform fault prediction on the air supply machine based on the preprocessed operating parameters.

[0031] The amplitude peak is the maximum vibration amplitude of the vibration signal; it is commonly used to judge the abnormality and severity of vibration; The vibration fundamental frequency is the largest single frequency component on the frequency spectrum of the vibration signal; it can help to judge problems such as shaft imbalance, bearing damage, and gear meshing.

[0032] The distortion degree is calculated according to the following formula: where D represents the distortion degree, represents the amplitude of the fundamental frequency of the vibration signal; represents the amplitude of the i-th harmonic of the vibration signal, and i takes the value of 2, 3, …, m, where m is the highest order of the harmonic participating in the calculation; the distortion degree reflects the distortion degree of the vibration signal and can reveal potential operating problems such as friction, impact, and resonance to some extent.

[0033] The frequency spectrum of the vibration signal is obtained by Fourier transform or fast Fourier transform (FFT), which can display the amplitudes of each frequency component, thereby identifying features such as fundamental frequency and harmonic components. These features are often associated with specific fault modes. For example, if the distortion degree is large and there are obvious harmonic components, the air supply machine may have faults such as shaft bending and rotor imbalance.

[0034] The noise intensity is represented by sound pressure level and calculated based on the noise signal; The average power is the energy consumed by the air supply machine per unit time; the energy consumption data of one day or one week can be counted, and the average power is obtained by dividing the energy by the time; ​The calculation method of the energy consumption efficiency coefficient is as follows: Record the energy consumption data in a period of time, and fit the energy consumption curve, with time as the abscissa and the energy consumption of the air supply fan as the ordinate. Record the air supply volume data in the same period of time, and fit the air supply volume curve, with time as the abscissa and the air supply volume of the air supply fan as the ordinate. Select n time points (n is a positive integer), read the energy consumption and air supply volume of the air supply fan at each time point through the energy consumption curve and the air supply volume curve, and calculate the energy consumption efficiency coefficient, the formula is as follows: ; Among them, The energy consumption efficiency coefficient is represented by C, The energy consumption of the air supply fan at the kth time point is represented by Ek, The air supply volume of the air supply fan at the kth time point is represented by Vk.

[0035] By studying the average power and the energy consumption efficiency coefficient, the possibility of energy utilization rate and efficiency reduction can be analyzed, and the performance degradation or other potential problems of the equipment can be indirectly inferred. For example, if the energy consumption efficiency coefficient is too low and the vibration amplitude and noise intensity are too large, the air supply fan may have blade damage or air duct blockage failure.

[0036] The average rotating speed is the average value of the rotating speed data of the main shaft of the air supply fan; the rotating speed directly affects the air volume output and working efficiency of the air supply fan, and is closely related to the mechanical wear of the equipment and the stability of the electrical system. Common types of rotating speed sensors include Hall effect sensors, photoelectric encoders, magneto-electric sensors, etc.

[0037] The maximum torque is the maximum value in the torque data of the air supply fan; the torque reflects the ability of the air supply fan to overcome resistance to do work, which is closely related to the air pressure generated by the air supply fan. If the torque of the air supply fan is abnormal, it may be caused by internal resistance increase (such as filter screen blockage, impeller scaling, bearing wear, etc.), drive system problem or motor failure, etc. Torque sensors mainly include resistance strain gauge type, magnetoelastic type, optical fiber Bragg grating sensor type, and eddy current type.

[0038] The calculation method of the temperature fluctuation index includes: extracting N consecutive air supply fan temperature values at a fixed time interval from the air supply fan temperature data, and calculating the temperature fluctuation index, the formula is as follows: ; Among them, The temperature fluctuation index is represented by T, The qth air supply fan temperature value is represented by Tq, The q-1th air supply fan temperature value is represented by Tq-1, and q takes the value of 2, 3, …, N; The time interval between any two consecutive temperature values of the air supply fan. The temperature fluctuation index can reflect whether there is overheating, insufficient cooling or other problems that may cause the performance of the equipment to decline or fail during operation.

[0039] As a preferred scheme of the present application, another calculation method of the temperature fluctuation index is as follows: M temperature values of the air supply fan are extracted from the air supply fan temperature data, and the temperature fluctuation index is calculated, the formula is as follows: ; Wherein, represents the jth temperature value of the air supply fan, represents the j-1th temperature value of the air supply fan, and j takes the value of 2, 3, …, M; represents the time interval between and , and the calculation formula is as follows: ; Wherein, represents the basic sampling interval, which is set by the person skilled in the art according to experience; represents the sampling correction factor, which is determined by the person skilled in the art through a large number of experiments; represents the j-2th temperature value of the air supply fan.

[0040] Since the temperature change of the air supply fan is usually continuous, the sampling frequency of the temperature sensor can be appropriately reduced to reduce the working pressure and data processing amount of the system. The basic sampling interval is set as the basis for subsequent adjustment of the sampling interval, and the sampling correction factor is adjusted to ensure that the value of the sampling interval is within a reasonable range; when the gap between two adjacent temperature data is large, it indicates that the temperature change trend of the air supply fan is obvious, and at this time, reducing the sampling frequency of the temperature data has no effect on the calculation of the temperature fluctuation index; when the gap between two adjacent temperature data is small, the temperature change trend of the air supply fan is not obvious, and there may be a turning point of temperature. If the turning point is missed by sampling, it will have a negative impact on the calculation accuracy of the temperature fluctuation index, so the sampling frequency should be increased at this time. Therefore, the present method can balance the simplification of data processing amount and the guarantee of calculation accuracy.

[0041] The test and maintenance module formulates a test and maintenance plan for the air supply fan based on the occurrence probability of each potential fault of the air supply fan, including determining the test and maintenance time and the test and maintenance task sequence; including an environment monitoring unit, a maintenance time unit, and a maintenance task unit; The environment monitoring unit is used to measure the environmental temperature and humidity of the working environment of the air supply fan in real time; The overhaul time unit is used to calculate the test overhaul time, and the formula is as follows: ; Wherein, represents the time interval between the time of the next test overhaul and the current time; represents the standard reference period, which is determined according to the industry standard of the air supply machine; represents the current time, represents the time of the last test overhaul; F represents the comprehensive risk index, which is calculated based on the failure prediction result of the multi-modal health assessment model; the larger the value of F is, the higher the failure risk of the air supply machine is; represents the average ambient temperature of the day, represents the standard ambient temperature of the air supply machine working; represents the average ambient humidity of the day, represents the standard ambient humidity of the air supply machine working; , are weight coefficients; they are respectively used to adjust the influence of ambient temperature and ambient humidity on the overhaul period, which are set by the person skilled in the art according to the specific circumstances.

[0042] The calculation method of the comprehensive risk index F is as follows: set a risk threshold for each failure type of the air supply machine; when the occurrence risk of any failure type is greater than its risk threshold, it is considered that the influence on the working of the air supply machine is large, and the overhaul needs to be carried out as soon as possible; the occurrence probability of each potential failure of the air supply machine output by the multi-modal health assessment model is counted, and the comprehensive risk index F is calculated, and the formula is as follows: ; Wherein, represents the occurrence probability of the pth potential failure whose occurrence probability exceeds the corresponding risk threshold, represents the risk threshold of the potential failure corresponding to ; Q is the number of types of potential failures whose occurrence probability exceeds the risk threshold.

[0043] The overhaul task unit is used to sort the test overhaul tasks, and the method is as follows: Get Q types of failures whose occurrence probability exceeds the risk threshold, and arrange test overhaul tasks for each failure; Based on the occurrence probability of each potential failure, the priority of the corresponding test overhaul task is allocated; wherein, The higher the occurrence probability of the potential failure is, the higher the priority of the corresponding test overhaul task is; The test overhaul tasks are sorted in order from high to low priority, and a test overhaul task sequence is generated.

[0044] The repair task unit is further configured with an adjustment strategy for adjusting the sequence of the test repair tasks; the adjustment strategy specifically includes: a prediction cycle is set; at the beginning of each prediction cycle, prediction of the occurrence probability of each potential fault is triggered; a growth rate of the occurrence probability of each potential fault is calculated based on the occurrence probability of each potential fault in the last at least two prediction cycles; The repair task unit is further configured with a probability risk threshold; if the growth rate of the occurrence probability of any potential fault is greater than the probability risk threshold, the priority of the corresponding test repair task is increased; If the priority of at least one test repair task changes, the reordering of the test repair task is triggered and the test repair task sequence is updated.

[0045] The application introduces the growth rate of the occurrence probability of the fault as a new dynamic criterion, not only considers the static risk level of the potential fault, but also pays attention to the change trend of the fault risk, significantly improves the detection sensitivity and response timeliness of the potential fault burst. Through periodic dynamic updating of the sequence, the maintenance plan has higher adaptability, and the fault risk caused by delay in prediction or risk aggravation is reduced.

[0046] The adjustment strategy further includes: determining the associated operating parameter of each potential fault; the associated operating parameter of any potential fault includes at least one of the operating parameters; optionally, the associated operating parameter of each potential fault is determined according to expert experience or historical data or the multi-modal health assessment model; for example, according to historical data, whether each operating parameter is abnormal when each potential fault actually occurs is counted; if the frequency of any operating parameter being abnormal when the potential fault actually occurs is greater than a preset reference frequency, the operating parameter is the associated operating parameter of the potential fault. Optionally, through the multi-modal health assessment model, the specified operating parameter is manually set to be abnormal and input into the model; if the occurrence probability of any potential fault significantly increases, the operating parameter is the associated operating parameter of the corresponding potential fault.

[0047] The repair task unit is further configured with a reference interval of each operating parameter; the abnormal operating parameter is identified based on the reference interval; for example, if the value of any operating parameter is outside the corresponding reference interval, it is marked as an abnormal operating parameter.

[0048] If there is an abnormal operating parameter in the associated operating parameter of any potential fault, the priority of the corresponding test repair task is adjusted, specifically including: calculating the proportion of the abnormal operating parameter in the associated operating parameter; the priority is increased based on the proportion, and the greater the proportion, the greater the adjusted priority.

[0049] The application further integrates actual working condition data by incorporating multiple operating parameters into the priority adjustment logic, realizes cross verification of actual fault performance and prediction model results, and improves the fine control of test maintenance task priority.

[0050] This module integrates information from different sensors and data sources, plans and adjusts the test maintenance time for the performance changes of the air supply fan under different environmental conditions, makes the system more intelligent, and makes the maintenance arrangement more reasonable.

[0051] The maintenance assistance module is used to assist in executing the test maintenance plan; including a spare parts allocation unit, a maintenance guidance unit, a maintenance data unit, and a visualization unit. The maintenance guidance unit is used to provide test maintenance suggestions; receives the test maintenance task sequence generated by the maintenance task unit, and provides test maintenance suggestions for each test maintenance task; integrates various fault diagnosis manuals, maintenance instructions, and historical maintenance cases, and recommends the optimal maintenance scheme for each fault type. For example, for a test maintenance task with a bearing fault type, the maintenance suggestion provided is to check whether the bearing is worn, fatigued, or poorly lubricated, replace the bearing and improve the lubrication system if necessary, adjust the appropriate pre-tightening force, and ensure that the bearing installation is well centered; for a test maintenance task with a motor fault type, the maintenance suggestion provided is to check the motor winding insulation, joint connection status, verify whether there is an overload, short circuit or ground fault, replace the damaged motor parts, and reasonably control the motor load; for a test maintenance task with a transmission system fault type, the maintenance suggestion provided is to check the tension of the belt or chain, the concentricity of the coupling, and confirm whether the transmission components such as the gear box are worn, damaged or misaligned, and replace or repair the damaged parts if necessary.

[0052] The spare parts allocation unit is used to generate a spare parts allocation list; based on the maintenance suggestions provided by the maintenance guidance unit for each test maintenance task, statistics all the spare parts and tools required by all test maintenance tasks, and generate a spare parts allocation list.

[0053] The maintenance data unit is used to record test maintenance data; specifically including test maintenance task content, consumable usage, and work hour consumption, forming a data analysis basis for subsequent optimization of maintenance strategies and improvement of work efficiency.

[0054] The visualization unit is used to visualize the operation parameters of the air supply machine, the fault prediction results, the test and maintenance task progress, and the maintenance suggestion of each test and maintenance task; the test and maintenance task execution situation is updated in real time in the form of a chart, a list, etc., including the current completion degree, the remaining workload, the predicted completion time, and other key information, so as to facilitate the management layer and the relevant personnel to master the maintenance progress. Meanwhile, the air supply machine operation data, the fault prediction results, the maintenance records, and other information are graphically displayed, helping the user to intuitively understand the state evolution and the maintenance effectiveness of the air supply machine.

[0055] The maintenance assistance module is an important component of the system, and can improve the maintenance efficiency, ensure the maintenance quality, and realize the reasonable scheduling of resources. The cooperative working mode among the fault prediction module, the test and maintenance module, and the maintenance assistance module is as shown in Figure 2

[0056] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can adopt a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer usable storage media containing computer usable program codes (including but not limited to disk storage, CD-ROM, optical storage, etc.).

[0057] The embodiments of the present application are described above in combination with the drawings, but the present application is not limited to the above-described specific embodiments, and the above-described specific embodiments are merely illustrative rather than limiting. Those skilled in the art can make many forms under the inspiration of the present application without departing from the purpose of the present application and the scope protected by the claims, and these all belong to the protection of the present application.​

Claims

1. A blower maintenance system based on fault analysis, characterized in that: It includes a fault prediction module, a testing and repair module, and a repair assistance module; among which: The fault prediction module is used to predict the probability of occurrence of each potential fault of the blower. The method is to collect the operating parameters of the blower and train a multimodal health assessment model, and then perform fault prediction through the multimodal health assessment model. The test and maintenance module is used to formulate a test and maintenance plan for the blower based on the probability of occurrence of each potential fault of the blower, including determining the test and maintenance time and the order of test and maintenance tasks. The maintenance assistance module is used to assist in the execution of the test and maintenance plan, including providing test and maintenance suggestions, assisting in spare parts allocation, and recording test and maintenance data.

2. The blower maintenance system based on fault analysis as described in claim 1, characterized in that: The fault prediction module includes a data acquisition unit, a feature extraction unit, a preprocessing unit, a model training unit, and a fault prediction unit; wherein: the data acquisition unit is used to collect data from the blower; The feature extraction unit extracts the operating parameters of the blower based on the blower data; The preprocessing unit is used to preprocess the operating parameters; The model training unit is used to establish the multimodal health assessment model and to train, optimize, and test the model. The fault prediction unit is used to deploy and apply the multimodal health assessment model to predict blower faults based on preprocessed operating parameters. The input of the multimodal health assessment model is the preprocessed operating parameters, and the output is the probability of occurrence of potential faults for each type of blower.

3. The blower maintenance system based on fault analysis as described in claim 2, characterized in that: The data acquisition unit collects the following data about the blower: blower vibration signal, noise signal, energy consumption data, air volume data, speed data, torque data, and blower temperature data. The operating parameters extracted by the feature extraction unit include peak amplitude, fundamental frequency, distortion, noise intensity, and average power; wherein: The peak amplitude is the maximum vibration amplitude of the vibration signal; The fundamental frequency of vibration is the largest single frequency component on the spectrum of the vibration signal. The formula for calculating the distortion is as follows: ； Where D represents the distortion degree, The amplitude of the fundamental frequency of the vibration signal; This represents the amplitude of the i-th harmonic of the vibration signal, where i ranges from 2, 3, ..., m, and m is the highest order of the harmonic involved in the calculation. The noise intensity is expressed as sound pressure level and is calculated based on the noise signal. The average power is the energy consumed by the blower per unit time.

4. The blower maintenance system based on fault analysis as described in claim 3, characterized in that: The operating parameters extracted by the feature extraction unit also include the energy efficiency coefficient, which is calculated as follows: Record energy consumption data over a period of time and fit an energy consumption curve, with time on the horizontal axis and energy consumption of the blower on the vertical axis. Record the air volume data within the same time period and fit an air volume curve, with the horizontal axis representing time and the vertical axis representing the air volume delivered by the blower. Select n time points, and read the energy consumption and air volume of the blower at each time point using the energy consumption curve and air volume curve, and calculate the energy efficiency coefficient as follows: ； in, Indicates the energy efficiency coefficient. This represents the energy consumption of the blower at the k-th time point. This represents the air volume delivered by the blower at the k-th time point.

5. The blower maintenance system based on fault analysis as described in claim 4, characterized in that: The operating parameters extracted by the feature extraction unit also include average rotational speed, maximum torque, and temperature fluctuation index; wherein: The average rotational speed is the average value of the rotational speed data of the blower's main shaft; The maximum torque is the maximum value in the torque data of the blower; The method for calculating the temperature fluctuation index includes: extracting N consecutive blower temperature values ​​from the blower temperature data at fixed time intervals, and calculating the temperature fluctuation index, as shown in the following formula: ； in, Indicates the temperature fluctuation index; This represents the temperature value of the q-th blower. This represents the temperature value of the (q-1)th blower, where q ranges from 2, 3, ..., N; This represents the time interval between any two consecutive blower temperature values.

6. The blower maintenance system based on fault analysis as described in claim 5, characterized in that: The method for calculating the temperature fluctuation index further includes: extracting M blower temperature values ​​from the blower temperature data and calculating the temperature fluctuation index, as shown in the following formula: ； in, This represents the temperature value of the j-th blower. This represents the temperature value of the (j-1)th blower, where j ranges from 2, 3, ..., M; express and The time interval between them is calculated using the following formula: ； in, Indicates the basic sampling interval; Indicates the sampling correction factor; This represents the temperature value of the (j-2)th blower.

7. The blower maintenance system based on fault analysis as described in claim 6, characterized in that: The testing and maintenance module includes an environmental monitoring unit, a maintenance time unit, and a maintenance task unit; wherein: The environmental monitoring unit is used to measure the ambient temperature and humidity of the working environment of the blower in real time. The maintenance time unit is used to calculate the test maintenance time; The maintenance task unit is used to sort test and maintenance tasks.

8. The blower maintenance system based on fault analysis as described in claim 7, characterized in that: The formula for calculating the test maintenance time in the maintenance time unit is as follows: ； in, This indicates the time interval between the next test and maintenance scheduled time and the current time. Indicates the standard reference period; Indicates the current moment. Indicates the date of the last test and maintenance; F represents the overall risk index; This indicates the average ambient temperature for the day. Indicates the standard ambient temperature at which the blower operates; This indicates the average ambient humidity for the day. The standard ambient humidity indicating the operation of the blower; , All are weighting coefficients; The comprehensive risk index F is calculated as follows: A risk threshold is set for each type of failure of the blower; the probability of occurrence of each potential failure of the blower output by the multimodal health assessment model is statistically analyzed, and the comprehensive risk index F is calculated using the following formula: ； in, This represents the probability of the p-th type of potential failure occurring if its probability exceeds the corresponding risk threshold. Indicates and The corresponding risk threshold for potential faults; Q is the number of types of potential faults with a probability of occurrence exceeding the risk threshold.

9. The blower maintenance system based on fault analysis as described in claim 8, characterized in that: The method for sequencing test and maintenance tasks by the maintenance task unit is as follows: Identify Q types of faults whose probability of occurrence exceeds a risk threshold, and assign test and maintenance tasks to each type of fault. Priority is assigned to the corresponding test and maintenance tasks based on the probability of occurrence of each potential fault; among which, The higher the probability of a potential fault occurring, the higher the priority of the corresponding testing and maintenance task. The test and maintenance tasks are sorted in descending order of priority to generate a test and maintenance task sequence.

10. The blower maintenance system based on fault analysis as described in claim 9, characterized in that: The maintenance task unit is also configured with an adjustment strategy for adjusting the order of the test and maintenance tasks; the adjustment strategy specifically includes: Set the prediction period; at the beginning of each prediction period, trigger the prediction of the probability of occurrence for each potential failure. Calculate the growth rate of the probability of occurrence of each potential fault based on the probability of occurrence of each potential fault in at least the last two prediction periods; The maintenance task unit is also equipped with a probability risk threshold; if the growth rate of the probability of any potential fault exceeds the probability risk threshold, the priority of the corresponding test and maintenance task is increased. If the priority of at least one test and maintenance task changes, the test and maintenance tasks are reordered and the test and maintenance task sequence is updated.

11. The blower maintenance system based on fault analysis as described in claim 10, characterized in that: The adjustment strategy further includes: determining the associated operating parameters for each potential fault; the associated operating parameters for any potential fault include at least one of the operating parameters; The maintenance task unit is also equipped with a reference range for each operating parameter; abnormal operating parameters are identified based on the reference range. If any of the associated operating parameters of a potential fault contain abnormal operating parameters, the priority of the corresponding test and maintenance task will be adjusted. Specifically, this includes: calculating the proportion of abnormal operating parameters to associated operating parameters; increasing the priority based on the proportion, and the larger the proportion, the greater the adjusted priority.

12. The blower maintenance system based on fault analysis as described in claim 11, characterized in that: The maintenance assistance module includes a spare parts allocation unit, a maintenance guidance unit, a maintenance data unit, and a visualization unit; wherein: The maintenance guidance unit is used to provide test and maintenance suggestions for each test and maintenance task; The spare parts allocation unit is used to generate a spare parts allocation list; The maintenance data unit is used to record test and maintenance data; The visualization unit is used to visually display the operating parameters of the blower, fault prediction results, test and maintenance task progress, and maintenance suggestions for each test and maintenance task.

Images