A method for elevator fault diagnosis and positioning based on big data

Abstract

The present application relates to the technical field of elevator fault diagnosis, and specifically to a method for diagnosing and locating elevator faults based on big data, including: obtaining elevator speed, load data, and vibration acceleration in the vertical and horizontal directions; dividing each elevator operation phase into starting, constant speed, and braking phases, obtaining the amplitude variation coefficient and amplitude frequency coefficient of the starting and braking phases of each elevator operation, and obtaining the vibration anomaly coefficient of the starting and braking phases of each elevator operation; calculating the vibration consistency coefficient affected by the imbalance of the elevator traction wheel during each elevator operation to obtain the vertical vibration anomaly value during each elevator operation; calculating the natural frequency approximate coefficient of the horizontal vibration acceleration during each elevator operation to obtain the elevator fault significance value for each monitoring period; diagnosing the elevator fault condition; and locating the fault in combination with a classification model. The present application can improve the accuracy of elevator fault diagnosis and location.

Description

Technical Field

[0001] The present application relates to the technical field of elevator fault diagnosis, and specifically to a method for diagnosing and locating elevator faults based on big data. Background Art

[0002] Elevator systems have complex structures, and their operation relies on the coordinated operation of numerous mechanical components. Vibration acceleration can directly and in real time reflect the operating status of these mechanical components. Conventional elevator fault diagnosis methods perform spectral analysis on the different components of the elevator. When the amplitude of a frequency in the spectral data exceeds its corresponding set threshold, a fault alarm is issued. However, elevator operating load conditions are relatively variable, and vibration characteristics will vary under different loads and operating speeds, which can easily lead to misjudgments. In addition, the vibration signals generated by different components may interact with each other, making it difficult to accurately locate the fault. Therefore, due to the failure to fully consider the different load conditions of elevator operation and the specific fault characteristics of each component, it is easy to cause misjudgments in elevator fault diagnosis and inaccurate fault location.

[0003] Publication No. CN109264521B describes an elevator fault diagnosis device. Sensors installed on the car and speed governor detect vibration signals and obtain corresponding spectrum data. If the amplitude of a frequency in the spectrum exceeds a set threshold, the main controller issues an alarm signal to diagnose the fault. However, due to the complex operating conditions of elevators, this method is prone to misjudgment and inaccurately locates faults. Summary of the Invention

[0004] In order to solve the above technical problems, the present application provides an elevator fault diagnosis and positioning method based on big data to solve the existing problems.

[0005] The present application discloses a method for diagnosing and locating elevator faults based on big data, which adopts the following technical solutions:

[0006] An embodiment of the present application provides a method for diagnosing and locating elevator faults based on big data, comprising the following steps:

[0007] Obtain the elevator's speed, load data, and vibration acceleration in the vertical and horizontal directions;

[0008] Each elevator operation stage is divided into starting, constant speed, and braking stages according to the speed change of each elevator operation. The amplitude variation coefficient of the starting and braking stages of each elevator operation is obtained by the variation trend of the maximum and minimum values ​​of the vibration acceleration in the starting and braking stages of each elevator operation. The amplitude frequency coefficient of the starting and braking stages of each elevator operation is obtained based on the difference between adjacent extreme values ​​of the vibration acceleration in the starting and braking stages of each elevator operation. In combination with the amplitude variation coefficient, the vibration abnormality coefficient of the starting and braking stages of each elevator operation is obtained.

[0009] Analyze the differences in vibration acceleration data between the uniform speed stage and the starting and braking stages of each elevator operation, obtain the vibration consistency coefficient affected by the imbalance of the elevator traction sheave, and then combine it with the vibration anomaly coefficient to obtain the vertical vibration anomaly value of each elevator operation;

[0010] Based on the correlation between the horizontal vibration acceleration of different components of the elevator during operation and the corresponding frequency domain data, the natural frequency approximate coefficient of the horizontal vibration acceleration of each elevator operation is obtained, and the correlation between the elevator load data in each monitoring period and the vertical vibration abnormal value and the natural frequency approximate coefficient is analyzed to obtain the elevator fault significance value in each monitoring period;

[0011] The elevator fault significance value is used to diagnose the elevator fault situation in each monitoring period, and the fault location is performed in combination with the classification model.

[0012] Preferably, the method of dividing each elevator operation stage into starting, uniform speed and braking stages is: performing a linear fitting on the running speed of each elevator operation, taking the stage corresponding to a slope greater than zero as the starting stage, the stage corresponding to a slope of zero as the uniform speed stage, and the stage corresponding to a slope less than zero as the braking stage.

[0013] Preferably, the method for obtaining the amplitude variation coefficient in the starting phase of each elevator operation is as follows: extracting all extreme points of the vibration acceleration in the starting phase of the elevator operation, arranging all the maximum values ​​and minimum values ​​in the starting phase in ascending time order to obtain a maximum value sequence and a minimum value sequence, respectively, using the Sen's Slope test algorithm to obtain the trend slopes of the maximum value and minimum value sequences, respectively, and taking the average of the absolute values ​​of the two trend slopes as the amplitude variation coefficient in the starting phase of each elevator operation.

[0014] Preferably, the method for obtaining the amplitude frequency coefficient of the starting phase of each elevator operation is: calculating the difference between all adjacent extreme values ​​of the vibration acceleration in the starting phase of each elevator operation, and summing up all the differences as the amplitude frequency coefficient of the starting phase of each elevator operation.

[0015] Preferably, the calculation method of the vibration abnormality coefficient during the starting phase of each elevator operation is: , where are the vibration abnormality coefficient, amplitude variation coefficient, and amplitude frequency coefficient of the starting stage of the i-th elevator operation.

[0016] Preferably, the method for obtaining the vibration consistency coefficient affected by the imbalance of the elevator traction sheave each time the elevator runs is: respectively calculating the DTW distance between the vibration acceleration data corresponding to the uniform speed stage and the starting and braking stages of each elevator run, and taking the average of the two DTW distances as the vibration consistency coefficient affected by the imbalance of the elevator traction sheave each time the elevator runs.

[0017] Preferably, the method for calculating the abnormal value of vertical vibration during each elevator operation is:

[0018] , where is the abnormal value of vertical vibration during the i-th elevator operation, are the vibration anomaly coefficients of the starting and braking stages of the i-th elevator operation, It is the vibration consistency coefficient caused by the imbalance of the elevator traction wheel during the i-th elevator operation.

[0019] Preferably, the method for obtaining the natural frequency approximate coefficient of the horizontal vibration acceleration each time the elevator runs is:

[0020] Perform modal decomposition on the horizontal vibration acceleration of different components during elevator operation, extract the frequency spectrum of each modal component of the horizontal vibration acceleration of different components, and use the frequency with the maximum amplitude in the frequency spectrum of each modal component as the natural frequency of each modal component;

[0021] The natural frequencies of all modal components of the horizontal vibration acceleration of each component are arranged in ascending order according to the modal component serial number to obtain the natural frequency sequence corresponding to each component. The mean of the maximum mutual information coefficient between all any two natural frequency sequences is used as the natural frequency approximation coefficient of the horizontal vibration acceleration during elevator operation.

[0022] Preferably, the calculation method of the elevator fault significance value in each monitoring period is:

[0023] ,in, is the elevator fault significance value in the tth monitoring period, 、 are the correlation coefficients between the load sequence of the t-th monitoring period and the vertical vibration abnormal value sequence and the natural frequency approximate sequence, respectively. exp() represents the exponential function with the natural constant as the base;

[0024] Among them, the preset time length is used as a monitoring period, and all load data, vertical vibration abnormal values, and natural frequency approximation coefficients in each monitoring period are arranged in ascending order to obtain the load sequence, vertical vibration abnormal value sequence, and natural frequency approximation sequence of each monitoring period.

[0025] Preferably, the diagnosis of elevator fault conditions in each monitoring period further includes: normalizing the elevator fault significance value and setting a fault diagnosis threshold. When the normalized result of the elevator fault significance value is less than or equal to the fault diagnosis threshold, there is no fault in the elevator operation during the corresponding monitoring period; otherwise, there is a fault in the elevator operation during the corresponding monitoring period.

[0026] This application has at least the following beneficial effects:

[0027] This application deeply analyzes the abnormal characteristics of the oscillation amplitude changes corresponding to the vibration acceleration in the vertical direction and the similar characteristics of the natural frequency of the vibration acceleration in the horizontal direction, considers the degree to which the vibration abnormalities in different directions are affected by the changes in load conditions, and calculates the elevator fault significance value. Its advantage is that it can reduce the interference of variable operating load conditions and the interaction of component vibrations, perform more accurate diagnosis and analysis based on the fault significance value, and combine the classification model to locate the fault, so as to make up for the defects of the existing method that is prone to misjudgment and inaccurate positioning. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present application or the prior art, the following is a brief introduction to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0029] Figure 1 This application provides a flowchart of the steps of a method for diagnosing and locating elevator faults based on big data. DETAILED DESCRIPTION

[0030] To further illustrate the technical means and effectiveness of this application to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, describes in detail the specific implementation, structure, features, and effectiveness of a big data-based elevator fault diagnosis and location method proposed in this application. In the following description, different references to "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.

[0031] Unless otherwise defined, terms such as "comprises," "comprising," or any other variants thereof are intended to encompass non-exclusive inclusion, such that a circuit structure, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such article or device. In the absence of further restrictions, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the article or device comprising the element. In addition, the term "and\or" as used herein includes any and all combinations of one or more related listed items. All technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application pertains.

[0032] The following describes in detail a specific solution of an elevator fault diagnosis and positioning method based on big data provided by this application with reference to the accompanying drawings.

[0033] An embodiment of the present application provides a method for diagnosing and locating elevator faults based on big data. For details, please refer to Figure 1 , including the following steps:

[0034] Step 1: Obtain the elevator's speed, load data, and vibration acceleration in the vertical and horizontal directions.

[0035] Elevator operation relies on the coordinated operation of numerous mechanical components, such as the hoist, guide rails, car, and door operator. Faults such as wear, looseness, and deformation in these components inevitably cause changes in vibration characteristics. For example, wear in the hoist's bearings can increase unbalanced forces during rotation, generating abnormal vibration signals. Vibration acceleration can directly and in real time reflect the operating status of these mechanical components. The vibration characteristics of different components and directions vary. Therefore, this embodiment installs vibration sensors in the elevator car to collect vertical vibration acceleration data and the elevator's operating speed during each run. Simultaneously, vibration sensors are installed on the car, guide rails, and hoist to collect horizontal vibration acceleration data. Furthermore, the degree of elevator vibration abnormality varies with load. Therefore, to fully account for elevator vibration characteristics under different load conditions, a pressure sensor is used to obtain elevator load data. The vibration acceleration is collected at a frequency of 100 Hz, and the elevator load data is the load data collected before the elevator runs.

[0036] Step 2: Divide each elevator operation stage into starting, constant speed and braking stages according to the change of vibration acceleration data during each elevator operation. Obtain the amplitude variation coefficient of the starting and braking stages of each elevator operation through the change trend of the maximum and minimum values ​​of the vibration acceleration during the starting and braking stages of each elevator operation. Obtain the amplitude frequency coefficient of the starting and braking stages of each elevator operation based on the difference between adjacent extreme values ​​of the vibration acceleration during the starting and braking stages of each elevator operation. Combined with the amplitude variation coefficient, obtain the vibration abnormality coefficient of the starting and braking stages of each elevator operation.

[0037] The dynamic characteristics analysis of elevator vibration includes transient response analysis and modal response analysis. Transient response analysis is the analysis of the instantaneous response of the elevator system to external influences or component abnormalities during operation, and modal analysis is the calculation and analysis of the natural frequency of free vibration during elevator operation. Due to the complex and changeable load conditions of elevator operation and the influence of the interaction between the vibrations of different components, it is difficult to fully capture and accurately identify potential fault information by simply analyzing the abnormalities of a certain feature of vibration acceleration. Therefore, this application further fully considers the influence of load status by obtaining the transient response and modal response of vibration acceleration, and then diagnoses elevator faults.

[0038] First, regarding transient response, many factors influence elevator vertical vibration, such as inertia during starting and braking, and the impact of elevator traction sheave imbalance on vertical vibration. Under the influence of different faults, the elevator's vertical vibration acceleration data exhibits different abnormal vibration characteristics. Taking the vibration acceleration data collected during the i-th elevator run as an example, since elevator system operation is divided into three phases: starting, constant speed, and braking, the vibration acceleration data, under the influence of inertia, exhibits particularly pronounced fluctuations during the starting acceleration and braking deceleration phases. However, the vertical vibration fluctuations during the constant speed phase are relatively small. Furthermore, during starting and braking, the vibration acceleration data exhibits large wave-like oscillations around zero, with the amplitude of the oscillation rapidly decreasing. Based on these characteristics, the following analysis is performed.

[0039] During elevator operation, the speed variation corresponding to the starting and braking stages is significantly larger, while the speed during the uniform speed stage is relatively stable. Therefore, in this embodiment, each elevator operation process is divided into three stages: starting, uniform speed, and braking, based on the acceleration, uniform speed, and deceleration intervals of the elevator operation speed. In actual application scenarios, the implementer can make the demarcation at will, and this embodiment does not impose any special restrictions on this. Preferably, in this embodiment, during each elevator operation, a linear fit is performed on the operating speed of each elevator operation, and the stage corresponding to the slope greater than zero is regarded as the starting stage, the stage corresponding to the slope of zero is regarded as the uniform speed stage, and the stage corresponding to the slope less than zero is regarded as the braking stage. Thus, the starting, uniform speed, and braking stages of the elevator operation can be obtained respectively.

[0040] Furthermore, all extreme points in the vibration acceleration during the starting and braking stages of the elevator operation are extracted. Taking the starting stage as an example, all the maximum values ​​in the starting stage are arranged in ascending order of time to obtain the corresponding maximum value sequence, and the Sen's Slope test algorithm is used to obtain the change trend characteristics of the maximum value sequence, and the corresponding trend slope is output. The larger the absolute value of the estimated slope is, the faster the rate of change is represented. The above method is used to obtain the corresponding trend slope for all the minimum values ​​in the starting stage, and the average of the absolute values ​​of the two trend slopes is calculated to obtain the amplitude variation coefficient with the characteristic of rapid reduction of oscillation amplitude. The amplitude variation coefficient of the starting stage of the i-th elevator operation is recorded as , income The larger the value is, the faster the rate of decrease of the oscillation amplitude of the vibration acceleration in the starting stage of the i-th elevator operation is.

[0041] In addition, the overall amplitude and frequency of the oscillation can reflect the degree of influence of the inertial force, and then the difference between the adjacent extreme values ​​of the vibration acceleration in the starting stage of each elevator operation is calculated. The cumulative sum of all the differences is then used as the amplitude frequency coefficient of the starting stage. The amplitude frequency coefficient of the starting stage of the i-th elevator operation is recorded as , income The larger the value is, the higher the oscillation amplitude and frequency of the vibration acceleration in the starting phase of the i-th elevator operation are.

[0042] Therefore, according to the amplitude variation coefficient and the amplitude frequency coefficient of the starting phase of each elevator operation, the vibration abnormality coefficient of the starting phase of each elevator operation is obtained. In this embodiment, the specific calculation formula is:

[0043] , where They are respectively the vibration abnormality coefficient, amplitude variation coefficient, and amplitude frequency coefficient of the starting stage of the i-th elevator operation. The larger the value is, the greater the oscillation amplitude and frequency will be during the elevator startup process, and the characteristic of rapid amplitude reduction will be more obvious.

[0044] Correspondingly, for the vibration acceleration in the braking phase, the above process of this embodiment is repeated, and the same calculation steps are used to obtain the vibration abnormality coefficient in the braking phase of each elevator operation. The vibration abnormality coefficient in the braking phase of the i-th elevator operation is recorded as .

[0045] Step 3: Analyze the data differences between the vibration acceleration in the uniform speed stage and the starting and braking stages during each elevator operation, obtain the vibration consistency coefficient affected by the imbalance of the elevator traction wheel during each elevator operation, and then combine it with the vibration anomaly coefficient to obtain the vertical vibration anomaly value during each elevator operation.

[0046] Furthermore, considering that the greater the impact of the elevator traction sheave imbalance during operation, the more likely it is to intensify the oscillation degree of the entire vibration acceleration, the waveform characteristics of the vibration acceleration in the uniform speed stage are closer to those in the starting and braking stages. In view of this, for the vibration acceleration data of the i-th elevator operation, the DTW distances between the vibration acceleration data corresponding to the uniform speed stage and the starting and braking stages are calculated respectively. The average of the two DTW distances is taken as the vibration consistency coefficient affected by the elevator traction sheave imbalance during the i-th elevator operation, which is recorded as . Income It reflects the consistency between the curves of the uniform speed phase and the starting and braking phases in the vibration acceleration. In order to comprehensively reflect the abnormal characteristics of the vertical vibration acceleration in the inertial force and the imbalance of the elevator traction sheave during the operation of the elevator, this embodiment will combine the vibration abnormality coefficients of the starting and braking phases of each elevator operation and the vibration consistency coefficient affected by the imbalance of the elevator traction sheave during the elevator operation to calculate the vertical vibration abnormality value of each elevator operation. Preferably, the specific calculation formula is:

[0047] , where is the abnormal value of vertical vibration during the i-th elevator operation, are the vibration anomaly coefficients of the starting and braking stages of the i-th elevator operation, is the vibration consistency coefficient due to the imbalance of the elevator traction wheel during the i-th elevator operation. It can be understood that the obtained The larger the value is, the more significant the abnormal characteristics of vertical vibration caused by inertia and imbalance of the elevator traction wheel are during the operation of the elevator.

[0048] Step 4: Based on the correlation between the horizontal vibration acceleration of different components during elevator operation and the corresponding frequency domain data, obtain the natural frequency approximate coefficient of the horizontal vibration acceleration of each elevator operation, analyze the correlation between the elevator load data in each monitoring period and the vertical vibration abnormal value and the natural frequency approximate coefficient, and obtain the elevator fault significance value in each monitoring period.

[0049] Furthermore, the horizontal vibration frequency of an elevator system is slightly lower than its vertical vibration frequency. However, the similar natural vibration frequencies of different components in the horizontal direction are prone to cause resonance anomalies, which in turn exacerbates the horizontal vibration of the elevator system. Therefore, in terms of modal response analysis, this embodiment analyzes the frequency difference characteristics of the horizontal vibration acceleration of the car, guide rails, and traction machine.

[0050] In an elevator system, the vibration states of different components vary. However, the closer their corresponding frequency states are, the more likely they are to cause abnormal resonance in the elevator system. To analyze the differences in the frequency states of different components, this embodiment uses the variational modal decomposition algorithm to perform modal decomposition on the horizontal vibration acceleration of the car, guide rails, and traction machine, taking the i-th elevator operation as an example. Setting the number of modal components to 5, the modal components corresponding to the horizontal vibration acceleration of the car, guide rails, and traction machine are obtained.

[0051] Furthermore, the fast discrete Fourier transform is used to obtain the spectrum of each modal component. The frequency with the largest amplitude in the spectrum is the natural frequency of the corresponding modal component. For the horizontal vibration acceleration of the car, guide rail, and traction machine, the corresponding five natural frequency values ​​can be obtained respectively. Therefore, taking the horizontal vibration acceleration of the car as an example, all its natural frequencies are arranged in ascending order according to the modal component serial number to obtain a natural frequency sequence. The corresponding natural frequency sequence can also be obtained for the horizontal vibration acceleration of the guide rail and the traction machine. Based on the characteristic that the closer the natural frequencies are, the easier it is to cause abnormal resonance, the maximum mutual information coefficient (MIC) between any two natural frequency sequences is calculated respectively, and the average of all the maximum mutual information coefficients is used as the natural frequency approximation coefficient of the horizontal vibration acceleration of the elevator operation, and the natural frequency approximation coefficient of the horizontal vibration acceleration during the i-th elevator operation is recorded as . Income The larger the value, the closer the natural frequencies of the horizontal vibration accelerations of different components are during elevator operation, which is more likely to intensify the horizontal vibration of the elevator system. It should be noted that the calculation process of the maximum mutual information coefficient is a prior art and will not be described in detail in this embodiment.

[0052] Thus, according to the above process of this embodiment, the abnormal vibration characteristics in the vertical and horizontal directions during the i-th elevator operation can be obtained.

[0053] However, the amplitude and natural frequency of elevator vibration are not constant, but will change with changes in load. Specifically, as the load increases, the relative motion and interaction between the various components in the elevator system will also change, which may lead to changes in the vibration mode. For example, when lightly loaded, the vibration of the elevator may mainly manifest as the up and down vertical vibration of the car, so the corresponding vertical vibration anomaly is more significant; when heavily loaded, in addition to vertical vibration, the car may also experience horizontal shaking and torsional vibration, which are more complex vibration modes. This is because when heavily loaded, the center of gravity of the car may change, and the connections and constraints between the various components will also be affected, causing the vibration propagation path and distribution to change, and the corresponding horizontal vibration anomaly is more significant.

[0054] In view of this, this embodiment analyzes the changing characteristics of vibration anomalies in different directions as the load changes during multiple elevator operations. The preset time length is used as a monitoring period. Specifically, in this embodiment, 12 hours is used as a monitoring period. All load data, vertical vibration anomaly values, and natural frequency approximation coefficients in each monitoring period are arranged in ascending order to obtain the load sequence, vertical vibration anomaly value sequence, and natural frequency approximation sequence of each monitoring period. According to the above analysis, the more obvious the negative correlation feature between the obtained load sequence and the vertical vibration anomaly value sequence, or the more obvious the positive correlation feature between the load sequence and the natural frequency approximation sequence, the greater the possibility of a fault in the elevator operation during the monitoring period.

[0055] Therefore, in this embodiment, for the t-th monitoring period, the Spearman correlation coefficients between the load sequence of the t-th monitoring period and the vertical vibration abnormal value sequence and the natural frequency approximate sequence are calculated and recorded as 、 Based on this, the formula for calculating the elevator fault significance value in each monitoring period is: ,in, is the elevator fault significance value in the tth monitoring period, 、 are the Spearman correlation coefficients between the load sequence, the vertical vibration anomaly sequence, and the natural frequency approximation sequence for the tth monitoring period, respectively. exp() represents an exponential function with a natural constant as its base. The larger the value M, the greater the likelihood of an elevator failure.

[0056] Step 5: Use the elevator fault significance value to diagnose the elevator fault situation in each monitoring period, and combine it with the classification model to locate the fault.

[0057] In this embodiment, by deeply analyzing the abnormal characteristics of the oscillation amplitude change corresponding to the vibration acceleration in the vertical direction and the similar characteristics of the natural frequency of the vibration acceleration in the horizontal direction, the degree to which the vibration abnormalities in different directions are affected by the changes in load conditions is considered respectively, so as to obtain the elevator fault significance value in each monitoring period, which is used to reflect the possibility of elevator operation faults in each monitoring period.

[0058] Furthermore, fault diagnosis is performed based on the elevator fault significance value, the elevator fault significance value is normalized, and a fault diagnosis threshold is set. When the normalized result of the elevator fault significance value is less than or equal to the fault diagnosis threshold, there is no fault in the elevator operation during the corresponding monitoring period; otherwise, there is a fault in the elevator operation during the corresponding monitoring period.

[0059] Preferably, in this embodiment, the sigmoid function is used to normalize the elevator fault significance value, and the fault diagnosis threshold is 0.6. When the normalized result of the elevator fault significance value is in [0, 0.6], the elevator operation status is good during the corresponding monitoring period and no fault problem occurs; when the normalized result of the elevator fault significance value is in (0.6, 1], the elevator operation has a fault during the corresponding monitoring period. So far, according to the above process of this embodiment, the elevator operation status in each monitoring period can be diagnosed.

[0060] Since faults at different elevator locations will affect the vibration response of the car, causing the car vibration acceleration to exhibit different vibration characteristics, this embodiment adopts the PSO-LSSVM classification model to locate the fault based on the vibration characteristics. The loss function of the classification model is set to the cross entropy loss function, and the optimizer is the particle swarm optimization algorithm.

[0061] Preferably, in this embodiment, the vertical vibration anomaly values ​​corresponding to each elevator operation, the natural frequency approximation coefficients, and all elements in the natural frequency sequence corresponding to the car are combined to form a characteristic sequence. This characteristic sequence is input into the classification model, and the output is the corresponding elevator fault location. It should be noted that the specific process of training the classification model and classification is well known to those skilled in the art and will not be detailed in this embodiment.

[0062] It is understood that references to "one embodiment" or "some embodiments" in the present specification mean that one or more embodiments of the present application include a particular feature, structure, or characteristic described in conjunction with that embodiment. Thus, if "in one embodiment," "in some embodiments," "in other embodiments," or "in other embodiments" appear in different places in this specification, they do not necessarily refer to the same embodiment, but rather mean "one or more but not all embodiments," unless otherwise specifically emphasized. The terms "including," "comprising," "having," and their variations all mean "including but not limited to," unless otherwise specifically emphasized.

[0063] It should be noted that the above-mentioned sequence of the embodiments of the present application is for description only and does not represent the advantages and disadvantages of the embodiments. The above description is of a specific embodiment of this specification. 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, multi-tasking and parallel processing are also possible or may be advantageous. At the same time, the size of the sequence number of each step in the embodiment does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments in this specification.

[0064] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present application, and should all be included in the scope of protection of the present application.

Claims

1. A method for elevator fault diagnosis and location based on big data, characterized in that: The following steps are involved: Obtain the elevator's speed, load data, and vibration acceleration in the vertical and horizontal directions; Each elevator operation stage is divided into starting, constant speed, and braking stages according to the speed change of each elevator operation. The amplitude variation coefficient of the starting and braking stages of each elevator operation is obtained by the variation trend of the maximum and minimum values ​​of the vibration acceleration in the starting and braking stages of each elevator operation. The amplitude frequency coefficient of the starting and braking stages of each elevator operation is obtained based on the difference between adjacent extreme values ​​of the vibration acceleration in the starting and braking stages of each elevator operation. In combination with the amplitude variation coefficient, the vibration abnormality coefficient of the starting and braking stages of each elevator operation is obtained. Analyze the differences in vibration acceleration data between the uniform speed stage and the starting and braking stages of each elevator operation, obtain the vibration consistency coefficient affected by the imbalance of the elevator traction sheave, and then combine it with the vibration anomaly coefficient to obtain the vertical vibration anomaly value of each elevator operation; Based on the correlation between the horizontal vibration acceleration of different components of the elevator during operation and the corresponding frequency domain data, the natural frequency approximate coefficient of the horizontal vibration acceleration of each elevator operation is obtained, and the correlation between the elevator load data in each monitoring period and the vertical vibration abnormal value and the natural frequency approximate coefficient is analyzed to obtain the elevator fault significance value in each monitoring period; The elevator fault significance value is used to diagnose the elevator fault situation in each monitoring period, and the fault location is performed in combination with the classification model.

2. The elevator fault diagnosis and location method based on big data according to claim 1, characterized in that: The method for dividing each elevator operation stage into starting, uniform speed and braking stages is as follows: a linear fitting is performed on the running speed of each elevator operation, and the stage corresponding to a slope greater than zero is taken as the starting stage, the stage corresponding to a slope of zero is taken as the uniform speed stage, and the stage corresponding to a slope less than zero is taken as the braking stage.

3. The elevator fault diagnosis and location method based on big data according to claim 1, characterized in that: The method for obtaining the amplitude variation coefficient during the starting phase of each elevator operation is as follows: all extreme points of the vibration acceleration during the starting phase of the elevator operation are extracted, all the maximum values ​​and minimum values ​​of the starting phase are arranged in ascending time order to obtain a maximum value sequence and a minimum value sequence, and the Sen's Slope test algorithm is used to obtain the trend slopes of the maximum value sequence and the minimum value sequence, respectively. The average of the absolute values ​​of the two trend slopes is used as the amplitude variation coefficient during the starting phase of each elevator operation.

4. The elevator fault diagnosis and location method based on big data according to claim 1, characterized in that: The method for obtaining the amplitude frequency coefficient of the starting phase of each elevator operation is as follows: calculating the difference between all adjacent extreme values ​​of the vibration acceleration in the starting phase of each elevator operation, and summing up all the differences as the amplitude frequency coefficient of the starting phase of each elevator operation.

5. The elevator fault diagnosis and location method based on big data according to claim 1, characterized in that: The calculation method of the vibration abnormality coefficient during the starting phase of each elevator operation is: , where are the vibration abnormality coefficient, amplitude variation coefficient, and amplitude frequency coefficient of the starting stage of the i-th elevator operation.

6. The elevator fault diagnosis and location method based on big data according to claim 1, characterized in that: The method for obtaining the vibration consistency coefficient due to the imbalance of the elevator traction sheave during each elevator operation is as follows: the DTW distances between the vibration acceleration data corresponding to the uniform speed stage and the starting and braking stages of each elevator operation are respectively calculated, and the average of the two DTW distances is used as the vibration consistency coefficient due to the imbalance of the elevator traction sheave during each elevator operation.

7. The elevator fault diagnosis and location method based on big data according to claim 1, characterized in that: The calculation method of the vertical vibration abnormal value during each elevator operation is: , where is the abnormal value of vertical vibration during the i-th elevator operation, are the vibration anomaly coefficients of the starting and braking stages of the i-th elevator operation, It is the vibration consistency coefficient caused by the imbalance of the elevator traction wheel during the i-th elevator operation.

8. The elevator fault diagnosis and location method based on big data according to claim 1, characterized in that: The method for obtaining the approximate coefficient of the natural frequency of the horizontal vibration acceleration during each elevator operation is as follows: Perform modal decomposition on the horizontal vibration acceleration of different components during elevator operation, extract the frequency spectrum of each modal component of the horizontal vibration acceleration of different components, and use the frequency with the maximum amplitude in the frequency spectrum of each modal component as the natural frequency of each modal component; The natural frequencies of all modal components of the horizontal vibration acceleration of each component are arranged in ascending order according to the modal component serial number to obtain the natural frequency sequence corresponding to each component. The mean of the maximum mutual information coefficient between all any two natural frequency sequences is used as the natural frequency approximation coefficient of the horizontal vibration acceleration during elevator operation.

9. The elevator fault diagnosis and location method based on big data according to claim 1, characterized in that: The calculation method of the elevator fault significance value in each monitoring period is as follows: ,in, is the elevator fault significance value in the tth monitoring period, 、 are the correlation coefficients between the load sequence of the t-th monitoring period and the vertical vibration abnormal value sequence and the natural frequency approximate sequence, respectively. exp() represents the exponential function with the natural constant as the base; Among them, the preset time length is used as a monitoring period, and all load data, vertical vibration abnormal values, and natural frequency approximation coefficients in each monitoring period are arranged in ascending order to obtain the load sequence, vertical vibration abnormal value sequence, and natural frequency approximation sequence of each monitoring period.

10. The elevator fault diagnosis and location method based on big data according to claim 1, characterized in that: The diagnosis of the elevator fault conditions in each monitoring period further includes: normalizing the elevator fault significance value and setting a fault diagnosis threshold. When the normalized result of the elevator fault significance value is less than or equal to the fault diagnosis threshold, the elevator operation in the corresponding monitoring period has no fault; otherwise, the elevator operation in the corresponding monitoring period has a fault.

Images