Ball screw fault detection method, device and computer equipment

Abstract

The application relates to a ball screw fault detection method, device and computer equipment. The method comprises the following steps: collecting real-time motor torque of a servo motor connected with a target ball screw; determining a real-time torque order sequence according to the real-time motor torque; determining real-time screw meshing frequency, real-time nut meshing frequency and real-time rolling element meshing frequency according to real-time screw rotation speed, real-time screw nominal diameter, real-time ball nominal diameter, real-time screw thread angle and real-time contact angle of the target ball screw; inputting the real-time torque order sequence, the real-time screw meshing frequency, the real-time nut meshing frequency and the real-time rolling element meshing frequency into a pre-trained fault detection model to determine fault information of the target ball screw. The above scheme can improve the ball screw fault detection efficiency, improve the ball screw fault detection accuracy and reduce the ball screw fault detection cost.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of device detection, in particular to a ball screw fault detection method and device and computer equipment. BACKGROUND

[0002] Ball screws convert rotary motion into precise linear motion, and are widely used in numerical control machine tools, automation equipment, precision machinery and nuclear industry, bearing the responsibility of power transmission and positioning. However, in the long-term work, the ball screw may be abnormal due to factors such as wear, insufficient lubrication, overload, manufacturing defects, etc., such as vibration, noise and temperature rise, etc. Serious problems may cause equipment failure, affecting production efficiency and accuracy. At present, the abnormal state monitoring method of ball screw is detected by using vibration sensor, temperature sensor, current or visual detection, etc. Although these methods can monitor some abnormal conditions, they often require more complex hardware equipment and installation space, resulting in high cost of ball screw fault detection, and cannot timely and accurately reflect the working state of the ball screw. Therefore, how to improve the fault detection efficiency and accuracy of the ball screw, and save the fault detection cost of the ball screw, is a problem to be solved. SUMMARY

[0003] Therefore, it is necessary to provide a ball screw fault detection method, device and computer equipment capable of improving the fault detection efficiency and accuracy of the ball screw.

[0004] In a first aspect, the present application provides a ball screw fault detection method, which comprises:

[0005] Collecting the real-time motor torque of the servo motor connected to the target ball screw, and determining the real-time torque order sequence according to the real-time motor torque;

[0006] According to the real-time screw rotation speed, the real-time screw nominal diameter, the real-time ball nominal diameter, the real-time screw thread angle and the real-time contact angle of the target ball screw, the real-time screw engagement frequency, the real-time nut engagement frequency and the real-time rolling body engagement frequency are determined;

[0007] The real-time torque order sequence, the real-time screw engagement frequency, the real-time nut engagement frequency and the real-time rolling body engagement frequency are input into the pre-trained fault detection model to determine the fault information of the target ball screw.

[0008] In one embodiment, the real-time torque order sequence, the real-time screw engagement frequency, the real-time nut engagement frequency and the real-time rolling body engagement frequency are input into the pre-trained fault detection model to determine the fault information of the target ball screw, comprising:

[0009] inputting the real-time torque order sequence, the real-time screw engagement frequency, the real-time nut engagement frequency and the real-time rolling element engagement frequency into a pre-trained fault detection model to make the fault detection model determine a convolution kernel operator according to the real-time torque order sequence, the real-time screw engagement frequency, the real-time nut engagement frequency and the real-time rolling element engagement frequency;

[0010] determining a real-time fault detection index based on the convolution kernel operator, the real-time screw engagement frequency, the real-time nut engagement frequency and the real-time rolling element engagement frequency through a convolution layer of the fault detection model; the real-time fault detection index comprises a real-time screw engagement order, a real-time nut engagement order and a real-time rolling element engagement order;

[0011] determining a target fault perception feature according to the real-time fault detection index through an attention network layer of the fault detection model;

[0012] determining fault information of the target ball screw according to the target fault perception feature through a multi-layer perceptron of the fault detection model.

[0013] In one of the embodiments, determining a target fault perception feature according to the real-time fault detection index through an attention network layer of the fault detection model comprises:

[0014] determining a target input matrix according to the real-time fault detection index, determining a query vector matrix, a key vector matrix and a value vector matrix according to the target input matrix through an attention network layer of the fault detection model;

[0015] determining an attention matrix according to the query vector matrix and the key vector matrix, and determining an intermediate representation vector according to the attention matrix and the value vector matrix;

[0016] weighting the real-time fault detection index based on the intermediate representation vector to determine a fault perception feature.

[0017] In one of the embodiments, determining fault information of the target ball screw according to the target fault perception feature through a multi-layer perceptron of the fault detection model comprises:

[0018] determining a probability corresponding to a fault category of the target ball screw by processing the target fault perception feature based on a normalized activation function through a multi-layer perceptron of the fault detection model; the fault category comprises a screw fault, a nut fault, a rolling element fault and no fault;

[0019] determining the fault information of the target ball screw according to the probability corresponding to the fault category.

[0020] In one of the embodiments, the real-time screw meshing frequency, the real-time nut meshing frequency and the real-time rolling element meshing frequency are determined according to the real-time screw rotation speed, the real-time screw nominal diameter, the real-time ball nominal diameter, the real-time screw thread pitch angle and the real-time contact angle, comprising:

[0021] The real-time screw meshing frequency is determined according to the real-time screw rotation speed, the real-time screw nominal diameter, the real-time ball nominal diameter, the real-time screw thread pitch angle and the real-time contact angle;

[0022] The real-time nut meshing frequency is determined according to the real-time screw rotation speed, the real-time screw nominal diameter, the real-time ball nominal diameter, the real-time screw thread pitch angle and the real-time contact angle;

[0023] The real-time rolling element meshing frequency is determined according to the real-time screw rotation speed, the real-time screw nominal diameter, the real-time ball nominal diameter and the real-time contact angle.

[0024] In one of the embodiments, the above-mentioned ball screw fault detection method further comprises:

[0025] The sample motor torque of the servo motor connected to the sample ball screw is determined, the sample torque order sequence is determined according to the sample motor torque, and the sample state information of the sample ball screw is determined; the sample state information includes screw fault state, nut fault state, rolling element fault state and normal state;

[0026] The neural network model is trained according to the sample torque order sequence, the sample screw meshing frequency, the sample nut meshing frequency, the sample rolling element meshing frequency and the sample state information to determine the fault detection model.

[0027] In one embodiment, the neural network model is trained according to the sample torque order sequence, the sample screw meshing frequency, the sample nut meshing frequency, the sample rolling element meshing frequency and the sample state information to determine the fault detection model, comprising:

[0028] The neural network model is trained according to the sample torque order sequence, the sample screw meshing frequency, the sample nut meshing frequency, the sample rolling element meshing frequency and the sample state information, and the cross-entropy loss function of the neural network model is determined in real time during the training process;

[0029] The model parameters of the neural network model are updated based on the cross-entropy loss function by using the gradient descent method until the neural network model converges, and the trained neural network model is used as the fault detection model.

[0030] In a second aspect, the application also provides a ball screw fault detection device, the device comprising:

[0031] a torque order sequence determination module, configured to collect real-time motor torque of a servo motor connected to the target ball screw, and determine a real-time torque order sequence according to the real-time motor torque;

[0032] a meshing frequency determination module, configured to determine real-time screw meshing frequency, real-time nut meshing frequency and real-time rolling element meshing frequency according to real-time screw rotating speed, real-time screw nominal diameter, real-time ball nominal diameter, real-time screw thread angle and real-time contact angle of the target ball screw;

[0033] a fault information determination module, configured to input the real-time torque order sequence, the real-time screw meshing frequency, the real-time nut meshing frequency and the real-time rolling element meshing frequency into a pre-trained fault detection model, and determine fault information of the target ball screw.

[0034] In a third aspect, the present application further provides a computer device, which comprises a memory and a processor, the memory stores a computer program, and the processor realizes the following steps when executing the computer program:

[0035] collecting real-time motor torque of a servo motor connected to the target ball screw, and determining a real-time torque order sequence according to the real-time motor torque;

[0036] determining real-time screw meshing frequency, real-time nut meshing frequency and real-time rolling element meshing frequency according to real-time screw rotating speed, real-time screw nominal diameter, real-time ball nominal diameter, real-time screw thread angle and real-time contact angle of the target ball screw;

[0037] inputting the real-time torque order sequence, the real-time screw meshing frequency, the real-time nut meshing frequency and the real-time rolling element meshing frequency into a pre-trained fault detection model, and determining fault information of the target ball screw.

[0038] In a fourth aspect, the present application further provides a computer readable storage medium, which stores a computer program, and the computer program realizes the following steps when executed by a processor:

[0039] collecting real-time motor torque of a servo motor connected to the target ball screw, and determining a real-time torque order sequence according to the real-time motor torque;

[0040] determining real-time screw meshing frequency, real-time nut meshing frequency and real-time rolling element meshing frequency according to real-time screw rotating speed, real-time screw nominal diameter, real-time ball nominal diameter, real-time screw thread angle and real-time contact angle of the target ball screw;

[0041] input the real-time torque order sequence, the real-time screw engagement frequency, the real-time nut engagement frequency and the real-time rolling element engagement frequency into a pre-trained fault detection model to determine the fault information of the target ball screw.

[0042] The ball screw fault detection method, device and computer equipment described above collect the real-time motor torque of a servo motor connected to the target ball screw, determine a real-time torque order sequence according to the real-time motor torque, determine a real-time screw engagement frequency, a real-time nut engagement frequency and a real-time rolling element engagement frequency according to the real-time screw rotation speed, the real-time screw nominal diameter, the real-time ball nominal diameter, the real-time screw thread angle and the real-time contact angle of the target ball screw, input the real-time torque order sequence, the real-time screw engagement frequency, the real-time nut engagement frequency and the real-time rolling element engagement frequency into a pre-trained fault detection model to determine the fault information of the target ball screw. The problem that the fault detection cost of the ball screw is high and the fault information of the ball screw cannot be obtained in time and accurately is solved. In the above scheme, the ball screw rotation speed, the screw nominal diameter, the ball nominal diameter, the screw thread angle, the contact angle and the working state of the connected servo motor are comprehensively considered when the ball screw is detected, the real-time screw engagement frequency, the real-time nut engagement frequency and the real-time rolling element engagement frequency are determined according to the real-time screw rotation speed, the real-time screw nominal diameter, the real-time ball nominal diameter, the real-time screw thread angle and the real-time contact angle of the target ball screw, the real-time torque order sequence of the servo motor, and the fault information of the ball screw is determined through the pre-trained fault detection model, without the need to install large fault detection equipment, the fault information of the target ball screw can be determined, the fault detection efficiency of the ball screw can be improved, the fault detection precision of the ball screw can be improved, and the fault detection cost of the ball screw can be reduced. The servo motor torque and speed signals are used to construct a ball screw pair fault diagnosis model combining data driving and mechanism, without external sensors, the monitoring cost is low, and false alarms caused by abnormal data of external sensors can be reduced. BRIEF DESCRIPTION OF DRAWINGS

[0043] Figure 1 It is an application environment diagram of the ball screw fault detection method in one embodiment;

[0044] Figure 2 It is a flowchart of the ball screw fault detection method in one embodiment;

[0045] Figure 3 It is a flowchart of the target ball screw fault information determination method in one embodiment;

[0046] Figure 4 It is a flowchart of the ball screw fault detection method in another embodiment;

[0047] Figure 5 A structural block diagram of a ball screw fault detection device in an embodiment;

[0048] Figure 6 An internal structure diagram of a computer device in an embodiment. DETAILED DESCRIPTION

[0049] In order to make the purposes, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and not to limit the present application.

[0050] The ball screw fault detection method provided by the embodiments of the present application can be applied to the application environment as shown in Figure 1 The terminal 102 communicates with the server 104 through a network. The data storage system can store data required to be processed by the server 104. The data storage system can be integrated on the server 104, or placed on a cloud or other network server. The server 104 collects real-time motor torque of a servo motor connected to a target ball screw, determines a real-time torque order sequence according to the real-time motor torque, determines real-time screw meshing frequency, real-time nut meshing frequency and real-time rolling element meshing frequency according to real-time screw rotation speed, real-time screw nominal diameter, real-time ball nominal diameter, real-time screw thread angle and real-time contact angle of the target ball screw, inputs the real-time torque order sequence, the real-time screw meshing frequency, the real-time nut meshing frequency and the real-time rolling element meshing frequency into a pre-trained fault detection model, determines fault information of the target ball screw, and sends the fault information of the target ball screw to the terminal 102 through a communication network. The terminal 102 can be, but is not limited to, various personal computers, notebook computers, smart phones, tablet computers, Internet of Things devices and portable wearable devices. The Internet of Things device can be a smart speaker, a smart television, a smart air conditioner, a smart vehicle device, etc. The portable wearable device can be a smart watch, a smart bracelet, a head-mounted device, etc. The server 104 can be implemented by an independent server or a server cluster composed of multiple servers.

[0051] In an embodiment, as shown in Figure 2 A ball screw fault detection method is provided. The embodiment takes the method applied to a terminal as an example for illustration. It can be understood that the method can also be applied to a server, and can also be applied to a system including a terminal and a server, and is implemented through the interaction of the terminal and the server. In the embodiment, the method includes the following steps:

[0052] S210, collecting real-time motor torque of a servo motor connected to a target ball screw, and determining a real-time torque order sequence according to the real-time motor torque.

[0053] Wherein, the ball screw is a mechanical transmission device composed of a screw, a nut and balls, and the conversion between rotary motion and linear motion is realized through the rolling of steel balls. The ball screw realizes high-precision linear motion control through the rolling of internal balls, and the servo motor provides driving power. The two work together to meet the requirements of precision and efficiency in high-precision and high-efficiency scenarios such as numerical control machine tools and automatic equipment. The target ball screw refers to the ball screw to be detected for failure. The real-time torque order sequence refers to a signal sequence composed of a fixed number of motor torque signals collected when the servo motor rotates one circle. For example, a fixed number of motor torques collected when the servo motor rotates one circle can be used as a torque order signal, and the torque order signal collected when the servo motor rotates n circles can be used as a real-time torque order sequence, where n is an integer greater than or equal to 1.

[0054] Specifically, the real-time motor torque of the servo motor connected to the target ball screw is collected by a PLC (Programmable Logic Controller), and the real-time motor torque is converted into a real-time torque order sequence sampled at equal angles based on the speed of the servo motor. Wherein, the equal angle torque collection is a measurement method for synchronously collecting torque data by fixing the number of sampling points per revolution, which is suitable for dynamic torque measurement of rotating machinery. The core is to ensure that the number of sample points collected per revolution is the same at different speeds, thereby eliminating the signal aliasing and spectrum tailing problems caused by speed fluctuations.

[0055] For example, the real-time torque order sequence can be represented by formula (1):

[0056] (1)

[0057] Wherein, is the real-time torque order sequence, is the torque order signal of the servo motor collected when the servo motor rotates to the ith circle, 1≤i≤n; the servo motor can collect servo motor motor torque signals at equal angles per revolution, that is, contains data, is an integer greater than or equal to 1; the real-time torque order sequence contains n torque order signals.

[0058] S220, determining the real-time screw meshing frequency, the real-time nut meshing frequency and the real-time rolling body meshing frequency according to the real-time screw rotation speed, the real-time screw nominal diameter, the real-time ball nominal diameter, the real-time screw thread angle and the real-time contact angle of the target ball screw.

[0059] wherein the screw rotation speed refers to the number of rotations of the ball screw per minute, the unit is revolutions per minute, and is an important indicator for measuring the speed of the screw. The nominal diameter of the screw is the virtual ideal diameter at the position of the screw; the nominal diameter of the ball refers to the virtual ideal diameter at the position of the ball center, that is, the diameter of the connecting line of all ball centers, which is used to determine the stiffness and accuracy of the screw and other performances. The thread angle of the screw refers to the angle between the tangent of the helix on the middle diameter cylinder of the screw and the plane perpendicular to the thread axis. The middle diameter cylinder of the screw refers to the diameter of the imaginary cylinder at the position where the groove and the protrusion width of the thread axial section are equal, which is usually used to control the fitting accuracy and strength of the thread. The contact angle of the screw usually refers to the angle between the common normal line of the contact surface of the ball screw and the diameter line of the ball center, which is mainly used to describe the mechanical properties of the ball screw transmission. The meshing frequency of the screw refers to the number of meshing times or the corresponding periodic frequency generated by the screw in the rotation process, which is determined by the rotation frequency of the screw and the number of teeth of the screw. The number of teeth of the screw usually refers to the number of threads of the screw, that is, the number of threads, which represents the density of the thread. The target rolling body meshing is explained below; the rolling body is a ball; the real-time screw nominal diameter, the real-time ball nominal diameter, the real-time screw thread angle, the real-time contact angle and the real-time screw rotation speed can be obtained by collecting the corresponding sensors. The meshing frequency is the periodic vibration frequency caused by the meshing of the gear transmission system, which is determined by the product of the rotation frequency of the gear and the number of teeth.

[0060] For example, according to the real-time screw rotation speed, the real-time screw nominal diameter, the real-time ball nominal diameter, the real-time screw thread angle and the real-time contact angle of the target ball screw, the real-time screw meshing frequency, the real-time nut meshing frequency and the real-time rolling body meshing frequency are determined, comprising:

[0061] According to the real-time screw rotation speed, the real-time screw nominal diameter, the real-time ball nominal diameter, the real-time screw thread angle and the real-time contact angle, the real-time screw meshing frequency is determined; according to the real-time screw rotation speed, the real-time screw nominal diameter, the real-time ball nominal diameter, the real-time screw thread angle and the real-time contact angle, the real-time nut meshing frequency is determined; according to the real-time screw rotation speed, the real-time screw nominal diameter, the real-time ball nominal diameter and the real-time contact angle, the real-time rolling body meshing frequency is determined.

[0062] Specifically, the calculation formula of the real-time screw meshing frequency is shown in formula (2):

[0063] (2)

[0064] wherein n is the real-time screw rotation speed, d0 is the real-time screw nominal diameter, d b is the real-time ball nominal diameter, is the real-time screw thread angle, is the real-time contact angle, and f sThe real-time screw engagement frequency.

[0065] The calculation formula of the real-time nut engagement frequency is shown in formula (3):

[0066] (3)

[0067] Wherein, f n The real-time nut engagement frequency.

[0068] The calculation formula of the real-time rolling body engagement frequency is shown in formula (4):

[0069] (4)

[0070] Wherein, f b The real-time rolling body engagement frequency.

[0071] The above scheme provides a real-time screw engagement frequency, a real-time nut engagement frequency and a real-time rolling body engagement frequency, which can improve the accuracy of the real-time screw engagement frequency, the real-time nut engagement frequency and the real-time rolling body engagement frequency.

[0072] S230, input the real-time torque order sequence, the real-time screw engagement frequency, the real-time nut engagement frequency and the real-time rolling body engagement frequency into the pre-trained fault detection model, and determine the fault information of the target ball screw.

[0073] Wherein, the fault information of the target ball screw includes the probability corresponding to the possible fault category corresponding to the target ball screw. The possible fault category corresponding to the target ball screw can include screw failure, nut failure, rolling body failure and no failure.

[0074] The fault detection method of the ball screw comprises the following steps: collecting real-time motor torque of a servo motor connected to the target ball screw, determining a real-time torque order sequence according to the real-time motor torque; determining real-time screw meshing frequency, real-time nut meshing frequency and real-time rolling element meshing frequency according to real-time screw rotating speed, real-time screw nominal diameter, real-time ball nominal diameter, real-time screw thread angle and real-time contact angle of the target ball screw; inputting the real-time torque order sequence, the real-time screw meshing frequency, the real-time nut meshing frequency and the real-time rolling element meshing frequency into a pre-trained fault detection model to determine fault information of the target ball screw. The problem that the fault detection cost of the ball screw is high and the fault information of the ball screw cannot be obtained in time and accurately is solved. In the above scheme, the rotating speed of the ball screw, the screw nominal diameter, the ball nominal diameter, the screw thread angle, the contact angle and the working state of the connected servo motor are comprehensively considered when the ball screw is detected for faults. The real-time screw meshing frequency, the real-time nut meshing frequency and the real-time rolling element meshing frequency determined according to the real-time screw rotating speed, the real-time screw nominal diameter, the real-time ball nominal diameter, the real-time screw thread angle and the real-time contact angle of the target ball screw, and the real-time torque order sequence of the servo motor are input into the pre-trained fault detection model to determine the fault information of the ball screw. Without the need of installing large-scale fault detection equipment, the fault information of the target ball screw can be determined, the fault detection efficiency of the ball screw is improved, the fault detection precision of the ball screw is improved, and the fault detection cost of the ball screw is reduced. The servo motor torque and speed signals are used to construct a ball screw pair fault diagnosis model combining data driving and mechanism, without external sensors, the monitoring cost is low, and false alarms caused by abnormal data of external sensors can be reduced.

[0075] In one embodiment, as shown in Figure 3 The real-time torque order sequence, the real-time screw meshing frequency, the real-time nut meshing frequency and the real-time rolling element meshing frequency are input into a pre-trained fault detection model to determine the fault information of the target ball screw.

[0076] S310, the real-time torque order sequence, the real-time screw meshing frequency, the real-time nut meshing frequency and the real-time rolling element meshing frequency are input into a pre-trained fault detection model, so that the fault detection model determines a convolution kernel operator according to the real-time torque order sequence, the real-time screw meshing frequency, the real-time nut meshing frequency and the real-time rolling element meshing frequency.

[0077] The convolution kernel operator comprises a convolution kernel operator corresponding to the real-time screw meshing frequency, a convolution kernel operator corresponding to the real-time nut meshing frequency and a convolution kernel operator corresponding to the real-time rolling element meshing frequency.

[0078] For example, the calculation formula of the convolution kernel operator corresponding to the real-time screw meshing frequency is shown in formula (5):

[0079] (5)

[0080] wherein, is the convolution kernel operator corresponding to the real-time screw engagement frequency, n is the number of torque order signals in the real-time torque order sequence, and j is an imaginary number.

[0081] The calculation formula of the convolution kernel operator corresponding to the real-time nut engagement frequency is shown in formula (6):

[0082] (6)

[0083] wherein, is the convolution kernel operator corresponding to the real-time nut engagement frequency.

[0084] The calculation formula of the convolution kernel operator corresponding to the real-time rolling body engagement frequency is shown in formula (7):

[0085] (7)

[0086] wherein, is the convolution kernel operator corresponding to the real-time rolling body engagement frequency.

[0087] S320, determining a real-time fault detection index based on the convolution kernel operator, the real-time screw engagement frequency, the real-time nut engagement frequency and the real-time rolling body engagement frequency through the convolution layer of the fault detection model.

[0088] The real-time fault detection index includes: a real-time screw engagement order, a real-time nut engagement order and a real-time rolling body engagement order.

[0089] Specifically, the real-time screw engagement order is determined according to the real-time screw engagement frequency based on the convolution kernel operator corresponding to the real-time screw engagement frequency through the convolution layer of the fault detection model; the real-time nut engagement order is determined according to the real-time nut engagement frequency based on the convolution kernel operator corresponding to the real-time nut engagement frequency; and the real-time rolling body engagement order is determined according to the real-time rolling body engagement frequency based on the convolution kernel operator corresponding to the real-time rolling body engagement frequency.

[0090] For example, the expression of the real-time screw engagement order is shown in formula (8):

[0091] (8)

[0092] wherein, is the real-time screw engagement order.

[0093] The expression of the real-time nut engagement order is shown in formula (9):

[0094] (9)

[0095] The real-time nut engagement order.

[0096] The expression of the real-time rolling body engagement order is shown in formula (10):

[0097] (10)

[0098] The real-time nut engagement order is shown in formula (10): The real-time rolling body engagement order.

[0099] S330, through the attention network layer of the fault detection model, determine the target fault perception feature according to the real-time fault detection index.

[0100] For example, the method for determining the target fault perception feature can be: determining a target input matrix according to the real-time fault detection index, determining a query vector matrix, a key vector matrix and a value vector matrix through the attention network layer of the fault detection model according to the target input matrix; determining an attention matrix according to the query vector matrix and the key vector matrix, and determining an intermediate representation vector according to the attention matrix and the value vector matrix; weighting the real-time fault detection index based on the intermediate representation vector to determine the real-time fault perception feature.

[0101] The query vector matrix is the Query matrix in the attention mechanism, the key vector matrix is the Key matrix in the attention mechanism, and the value vector matrix is the Value matrix in the attention mechanism.

[0102] Specifically, the expression of the target input matrix is shown in formula (11):

[0103] (11)

[0104] The target input matrix is F.

[0105] The target input matrix is processed through the three independent parameter matrices in the attention network layer of the fault detection model to generate the query vector matrix, the key vector matrix and the value vector matrix. The determination method of the attention matrix is shown in formula (12):

[0106] (12)

[0107] The attention matrix is shown in formula (12): The attention matrix is shown in formula (12): The query vector matrix is shown in formula (12): The transpose matrix of the key vector matrix is shown in formula (12), and d is the feature dimension corresponding to the attention network layer, which can be set according to actual needs.

[0108] The determination method of the intermediate representation vector is shown in formula (13):

[0109] (13)

[0110] wherein, is an intermediate representation vector, V is a value vector matrix.

[0111] The real-time fault detection index is weighted based on the intermediate representation vector, and a fault awareness feature is determined.

[0112] For example, the determination method of the fault awareness feature is shown in formula (14):

[0113] (14)

[0114] wherein, M is a fault awareness feature.

[0115] The above scheme provides a method for determining a fault awareness feature based on an attention mechanism, which can fully mine the correlation between the torque feature and the fault mode of the ball screw, thereby improving the accuracy of the subsequent determined fault information of the target ball screw.

[0116] S340, determining the fault information of the target ball screw according to the target fault awareness feature through the multi-layer perceptron of the fault detection model.

[0117] For example, the target fault awareness feature is processed based on a normalized activation function through the multi-layer perceptron of the fault detection model, and the probability corresponding to the fault category of the target ball screw is determined; the fault category includes screw failure, nut failure, rolling element failure and no failure; the fault information of the target ball screw is determined according to the probability corresponding to the fault category.

[0118] The normalized activation function is a softmax activation function, which is mainly used in the output layer of a multi-classification task to convert the output of the model into a probability distribution, i.e. the output value corresponding to each category is compressed to between 0 and 1, and the sum of all output values is 1.

[0119] Specifically, the target fault awareness feature is processed based on a normalized activation function through the multi-layer perceptron of the fault detection model, and the probability corresponding to the fault category of the target ball screw is determined, if there is a fault category corresponding to a probability greater than a preset probability threshold, it is determined that the target ball screw has a fault corresponding to the fault category.

[0120] For example, if the preset probability threshold is 0.4, the probability corresponding to the fault category of the target ball screw is: the probability corresponding to the screw failure is 0.1, the probability corresponding to the nut failure is 0.6, the probability corresponding to the rolling element failure is 0.2, and the probability corresponding to the no failure is 0.1, then the fault information of the target ball screw is that the target ball screw has a nut failure.

[0121] The above scheme processes the target fault perception feature based on a normalized activation function, determines the probability corresponding to the fault category of the target ball screw, can determine the probability of all possible faults of the target ball screw, and positions the fault of the target ball screw according to the probability corresponding to the fault category of the target ball screw, thereby improving the fault detection accuracy and fault positioning efficiency of the target ball screw.

[0122] The above method can avoid the influence of motor speed fluctuation on the fault detection result of the target ball screw and improve the stability of the model by extracting the fault order amplitude through the convolution kernel operator when determining the fault information of the target ball screw through the fault detection model.

[0123] In one embodiment, as shown in Figure 4 The above ball screw fault detection method further comprises:

[0124] S410, determining the sample motor torque of the servo motor connected to the sample ball screw, determining the sample torque order sequence according to the sample motor torque, and determining the sample state information of the sample ball screw.

[0125] The sample state information includes screw fault state, nut fault state, rolling element fault state and normal state.

[0126] The sample ball screw includes a ball screw in a screw fault state, a ball screw in a nut fault state, a ball screw in a rolling element fault state, and a ball screw in a normal state.

[0127] Specifically, the sample motor torque of the servo motor connected to the sample ball screw is collected through the PLC, the sample motor torque is converted into an equal-angle-sampled sample torque order sequence based on the speed of the servo motor, and the sample state information of the sample ball screw is determined.

[0128] S420, training the neural network model according to the sample torque order sequence, the sample screw engagement frequency, the sample nut engagement frequency, the sample rolling element engagement frequency and the sample state information, and determining the fault detection model.

[0129] For example, the model training data and the model test data can be determined according to the sample torque order sequence, the sample screw engagement frequency, the sample nut engagement frequency, the sample rolling element engagement frequency and the sample state information, and the neural network model is trained using the model training data to determine the candidate detection model. The candidate detection model is tested using the model test data, the cross-entropy loss function of the candidate detection model is determined, and if the cross-entropy loss function of the candidate detection model is less than the preset loss function threshold, the candidate detection model is determined as the fault detection model.

[0130] The scheme provides a model training method of a fault detection model, adopts the fault detection model to detect the ball screw fault, and can improve the fault detection efficiency of the ball screw.

[0131] In one embodiment, the neural network model is trained according to the sample torque order sequence, the sample screw engagement frequency, the sample nut engagement frequency, the sample rolling body engagement frequency and the sample state information, and a fault detection model is determined, including:

[0132] The neural network model is trained according to the sample torque order sequence, the sample screw engagement frequency, the sample nut engagement frequency, the sample rolling body engagement frequency and the sample state information, and the cross-entropy loss function of the neural network model is determined in real time during the training; the model parameters of the neural network model are updated based on the cross-entropy loss function by using the gradient descent method until the neural network model converges, and the trained neural network model is used as the fault detection model.

[0133] The gradient descent method is a first-order optimization algorithm for adjusting parameters to minimize the objective function through iteration, which adjusts the parameters step by step in the opposite direction of the gradient by calculating the gradient, i.e. the first derivative, of the objective function, so that the function value converges to a local minimum. The gradient descent method mainly includes three steps of initializing parameters, calculating gradients and iterative updating. The initialization of parameters includes random selection of a starting point; the calculation of gradients includes determination of the steepest descent direction of the current point; the iterative updating includes adjusting the parameters along the negative gradient direction by a step size until the neural network model converges.

[0134] The above scheme uses the gradient descent method to update the model parameters of the neural network model based on the cross-entropy loss function until the neural network model converges, which can improve the model training efficiency.

[0135] For example, on the basis of the above embodiment, the ball screw fault detection method includes:

[0136] The sample motor torque of the sample ball screw connected servo motor is collected by the PLC, the sample motor torque is converted into the sample torque order sequence of equal-angle sampling based on the speed of the servo motor, and the sample state information of the sample ball screw is determined. The neural network model is trained according to the sample torque order sequence and the sample state information, and the cross-entropy loss function of the neural network model is determined in real time during the training; the model parameters of the neural network model are updated based on the cross-entropy loss function by using the gradient descent method until the neural network model converges, and the trained neural network model is used as the fault detection model.

[0137] The real-time motor torque of the servo motor connected with the target ball screw is collected by a PLC (Programmable Logic Controller), and the real-time motor torque is converted into a real-time torque order sequence based on the rotation speed of the servo motor.

[0138] The real-time screw meshing frequency is determined according to the real-time screw rotation speed, the real-time screw nominal diameter, the real-time ball nominal diameter, the real-time screw thread angle and the real-time contact angle; the real-time nut meshing frequency is determined according to the real-time screw rotation speed, the real-time screw nominal diameter, the real-time ball nominal diameter, the real-time screw thread angle and the real-time contact angle; and the real-time rolling body meshing frequency is determined according to the real-time screw rotation speed, the real-time screw nominal diameter, the real-time ball nominal diameter and the real-time contact angle.

[0139] The real-time torque order sequence, the real-time screw meshing frequency, the real-time nut meshing frequency and the real-time rolling body meshing frequency are input into the pre-trained fault detection model, so that the fault detection model determines a convolution kernel operator according to the real-time torque order sequence, the real-time screw meshing frequency, the real-time nut meshing frequency and the real-time rolling body meshing frequency. Through the convolution layer of the fault detection model, the real-time screw meshing order is determined according to the real-time screw meshing frequency based on the convolution kernel operator corresponding to the real-time screw meshing frequency; the real-time nut meshing order is determined according to the real-time nut meshing frequency based on the convolution kernel operator corresponding to the real-time nut meshing frequency; and the real-time rolling body meshing order is determined according to the real-time rolling body meshing frequency based on the convolution kernel operator corresponding to the real-time rolling body meshing frequency. The target input matrix is determined according to the real-time fault detection index, the query vector matrix, the key vector matrix and the value vector matrix are determined through the attention network layer of the fault detection model according to the target input matrix, the attention matrix is determined according to the query vector matrix and the key vector matrix, and the intermediate representation vector is determined according to the attention matrix and the value vector matrix. The real-time fault detection index is weighted based on the intermediate representation vector to determine the fault perception feature. The target fault perception feature is processed based on the normalized activation function through the multilayer perceptron of the fault detection model to determine the probability corresponding to the fault category of the target ball screw, and if the probability corresponding to a fault category is greater than a preset probability threshold, it is determined that the target ball screw has a fault corresponding to the fault category.

[0140] In the ball screw fault detection method, the real-time motor torque of the servo motor connected to the target ball screw is collected, and the real-time torque order sequence is determined according to the real-time motor torque. The real-time screw meshing frequency, the real-time nut meshing frequency and the real-time rolling element meshing frequency are determined according to the real-time screw rotating speed, the real-time screw nominal diameter, the real-time ball nominal diameter, the real-time screw thread angle and the real-time contact angle of the target ball screw. The real-time torque order sequence, the real-time screw meshing frequency, the real-time nut meshing frequency and the real-time rolling element meshing frequency are input into the pre-trained fault detection model to determine the fault information of the target ball screw. The problem that the fault detection cost of the ball screw is high and the fault information of the ball screw cannot be obtained in time and accurately is solved. In the above scheme, when the ball screw is detected, the rotating speed of the ball screw, the screw nominal diameter, the ball nominal diameter, the screw thread angle, the contact angle and the working state of the connected servo motor are comprehensively considered. The real-time screw meshing frequency, the real-time nut meshing frequency and the real-time rolling element meshing frequency determined according to the real-time screw rotating speed, the real-time screw nominal diameter, the real-time ball nominal diameter, the real-time screw thread angle and the real-time contact angle of the target ball screw, and the real-time torque order sequence of the servo motor are input into the pre-trained fault detection model to determine the fault information of the ball screw. Without installing large fault detection equipment, the fault information of the target ball screw can be determined, the fault detection efficiency of the ball screw can be improved, the fault detection precision of the ball screw can be improved, and the fault detection cost of the ball screw can be reduced. The servo motor torque and speed signal are used to construct a ball screw pair fault diagnosis model combining data driving and mechanism. Without external sensors, the monitoring cost is low, and false alarms caused by abnormal external sensor data can be reduced.

[0141] It should be understood that, although each step in the flowchart involved in each embodiment as described above is displayed in sequence according to the arrow, these steps are not necessarily executed in the order indicated by the arrow. Unless otherwise specified herein, the execution of these steps is not strictly limited in sequence, and these steps can be executed in other orders. Moreover, at least part of the steps in the flowchart involved in each embodiment as described above can include multiple steps or stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily sequential, but can be alternately or alternately executed with at least part of other steps or steps or stages in other steps.

[0142] Based on the same inventive concept, the application further provides a ball screw fault detection device for implementing the ball screw fault detection method described above. The implementation scheme of the device for solving the problem is similar to the implementation scheme described in the above method, so the specific limitations in one or more ball screw fault detection device embodiments provided below can refer to the limitations of the ball screw fault detection method in the foregoing, which will not be repeated here.

[0143] In one embodiment, as shown in Figure 5 A ball screw fault detection device is provided, comprising: a torque order sequence determination module 501, a meshing frequency determination module 502, and a fault information determination module 503, wherein:

[0144] The torque order sequence determination module 501 is configured to collect real-time motor torque of a servo motor connected to a target ball screw, and determine a real-time torque order sequence based on the real-time motor torque.

[0145] The meshing frequency determination module 502 is configured to determine real-time screw meshing frequency, real-time nut meshing frequency, and real-time rolling element meshing frequency based on real-time screw rotation speed, real-time screw nominal diameter, real-time ball nominal diameter, real-time screw thread angle, and real-time contact angle of the target ball screw.

[0146] The fault information determination module 503 is configured to input the real-time torque order sequence, the real-time screw meshing frequency, the real-time nut meshing frequency, and the real-time rolling element meshing frequency into a pre-trained fault detection model to determine fault information of the target ball screw.

[0147] For example, the fault information determination module 503 is specifically configured to:

[0148] Input the real-time torque order sequence, the real-time screw meshing frequency, the real-time nut meshing frequency, and the real-time rolling element meshing frequency into the pre-trained fault detection model, so that the fault detection model determines a convolution kernel operator based on the real-time torque order sequence, the real-time screw meshing frequency, the real-time nut meshing frequency, and the real-time rolling element meshing frequency.

[0149] Determine a real-time fault detection index based on the convolution kernel operator, the real-time screw meshing frequency, the real-time nut meshing frequency, and the real-time rolling element meshing frequency through a convolution layer of the fault detection model; the real-time fault detection index includes real-time screw meshing order, real-time nut meshing order, and real-time rolling element meshing order.

[0150] Determine a target fault perception feature based on the real-time fault detection index through an attention network layer of the fault detection model.

[0151] determine, by the multi-layer perceptron of the fault detection model, the fault information of the target ball screw according to the target fault perception feature.

[0152] Further, the fault information determination module 503 is also specifically used for:

[0153] determine a target input matrix according to the real-time fault detection index, determine a query vector matrix, a key vector matrix and a value vector matrix according to the target input matrix through an attention network layer of the fault detection model;

[0154] determine an attention matrix according to the query vector matrix and the key vector matrix, and determine an intermediate representation vector according to the attention matrix and the value vector matrix;

[0155] weight the real-time fault detection index based on the intermediate representation vector to determine a fault perception feature.

[0156] Further, the fault information determination module 503 is also specifically used for:

[0157] determine a probability corresponding to a fault category of the target ball screw based on a normalized activation function by processing the target fault perception feature through the multi-layer perceptron of the fault detection model; the fault category includes a screw fault, a nut fault, a rolling element fault and a no fault;

[0158] determine the fault information of the target ball screw according to the probability corresponding to the fault category.

[0159] Exemplarily, the meshing frequency determination module 502 is specifically used for:

[0160] determine a real-time screw meshing frequency according to the real-time screw rotating speed, the real-time screw nominal diameter, the real-time ball nominal diameter, the real-time screw thread angle and the real-time contact angle;

[0161] determine a real-time nut meshing frequency according to the real-time screw rotating speed, the real-time screw nominal diameter, the real-time ball nominal diameter, the real-time screw thread angle and the real-time contact angle;

[0162] determine a real-time rolling element meshing frequency according to the real-time screw rotating speed, the real-time screw nominal diameter, the real-time ball nominal diameter and the real-time contact angle.

[0163] Exemplarily, the above-mentioned ball screw fault detection device further comprises:

[0164] The sample data determination module is configured to determine sample motor torque of a servo motor connected with a sample ball screw, determine a sample torque order sequence according to the sample motor torque, and determine sample state information of the sample ball screw; the sample state information includes a screw fault state, a nut fault state, a rolling element fault state, and a normal state.

[0165] The model training module is configured to train a neural network model according to the sample torque order sequence, the sample screw engagement frequency, the sample nut engagement frequency, the sample rolling element engagement frequency, and the sample state information, and determine a fault detection model.

[0166] For example, the model training module is specifically configured to:

[0167] The model training module is configured to train the neural network model according to the sample torque order sequence, the sample screw engagement frequency, the sample nut engagement frequency, the sample rolling element engagement frequency, and the sample state information, and determine a cross-entropy loss function of the neural network model in real time during the training.

[0168] The gradient descent method is used to update model parameters of the neural network model based on the cross-entropy loss function until the neural network model converges, and the trained neural network model is used as the fault detection model.

[0169] The above-mentioned modules in the ball screw fault detection device can be realized by software, hardware, or a combination thereof. The above-mentioned modules can be embedded in or independent of a processor in a computer device in hardware form, or stored in a memory in a computer device in software form, so as to be called and executed by a processor to perform operations corresponding to the above-mentioned modules.

[0170] In one embodiment, a computer device is provided, which can be a terminal, and an internal structure diagram of the computer device can be as shown in Figure 6As shown in the figure. The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit and an input device. Among them, the processor, the memory and the input / output interface are connected through the system bus, and the communication interface, the display unit and the input device are connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capability. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operating system and the computer program in the non-volatile storage medium to run. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used for wired or wireless communication with external terminals, and wireless communication can be realized through WIFI, mobile cellular network, NFC (near field communication) or other technologies. The computer program is executed by the processor to realize a ball screw fault detection method. The display unit of the computer device is used to form a visually visible picture, which can be a display screen, a projection device or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer overlaid on the display screen, or a key, trackball or touchpad arranged on the shell of the computer device, or an external keyboard, touchpad or mouse, etc.

[0171] Those skilled in the art can understand that, Figure 6 The skilled in the art can understand that,

[0172] In one embodiment, a computer device is provided, comprising a memory and a processor, the memory stores a computer program, and the processor executes the computer program to realize the following steps:

[0173] Step one, collecting the real-time motor torque of the servo motor connected to the target ball screw, and determining the real-time torque order sequence according to the real-time motor torque;

[0174] Step two, determining the real-time screw engagement frequency, the real-time nut engagement frequency and the real-time rolling body engagement frequency according to the real-time screw rotation speed, the real-time screw nominal diameter, the real-time ball nominal diameter, the real-time screw thread angle and the real-time contact angle of the target ball screw;

[0175] Step three, inputting the real-time torque order sequence, the real-time screw engagement frequency, the real-time nut engagement frequency and the real-time rolling body engagement frequency into the pre-trained fault detection model to determine the fault information of the target ball screw.

[0176] In one embodiment, a computer readable storage medium is provided, having stored thereon a computer program which, when executed by a processor, implements the following steps:

[0177] Step one, collect the real-time motor torque of the servo motor connected to the target ball screw, and determine the real-time torque order sequence according to the real-time motor torque;

[0178] Step two, determine the real-time screw engagement frequency, real-time nut engagement frequency and real-time rolling body engagement frequency according to the real-time screw rotation speed, real-time screw nominal diameter, real-time ball nominal diameter, real-time screw thread angle and real-time contact angle of the target ball screw;

[0179] Step three, input the real-time torque order sequence, real-time screw engagement frequency, real-time nut engagement frequency and real-time rolling body engagement frequency into the pre-trained fault detection model to determine the fault information of the target ball screw.

[0180] In one embodiment, a computer program product is provided, comprising a computer program which, when executed by a processor, implements the following steps:

[0181] Step one, collect the real-time motor torque of the servo motor connected to the target ball screw, and determine the real-time torque order sequence according to the real-time motor torque;

[0182] Step two, determine the real-time screw engagement frequency, real-time nut engagement frequency and real-time rolling body engagement frequency according to the real-time screw rotation speed, real-time screw nominal diameter, real-time ball nominal diameter, real-time screw thread angle and real-time contact angle of the target ball screw;

[0183] Step three, input the real-time torque order sequence, real-time screw engagement frequency, real-time nut engagement frequency and real-time rolling body engagement frequency into the pre-trained fault detection model to determine the fault information of the target ball screw.

[0184] It should be noted that the user information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties, and the collection, use and processing of related data need to comply with relevant laws, regulations and standards of relevant countries and regions.

[0185] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when the computer program is executed, the processes of the above-mentioned embodiments of the methods can be included. Any reference to memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical storage, high-density embedded non-volatile memory, resistive memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration but not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The database involved in the embodiments provided in the present application can include at least one of a relational database and a non-relational database. The non-relational database can include a distributed database based on a block chain, etc., without being limited thereto. The processor involved in the embodiments provided in the present application can be a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., without being limited thereto.

[0186] Any combination of the technical features of the above embodiments can be made. In order to make the description simple, all possible combinations of the technical features in the above embodiments are not described, however, as long as the combination of the technical features does not exist, it should be considered as the scope of the present application.

[0187] The above embodiments only express several implementation manners of the present application, and the description is more specific and detailed, but it should not be understood as a limitation on the scope of the patent of the present application. It should be pointed out that for ordinary skilled in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are within the scope of protection of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. A method of detecting a failure of a ball screw, characterized by, The method comprises the following steps: Collecting real-time motor torque of a servo motor connected to a target ball screw, and determining a real-time torque order sequence according to the real-time motor torque; Determining real-time screw meshing frequency, real-time nut meshing frequency and real-time rolling element meshing frequency according to real-time screw rotation speed, real-time screw nominal diameter, real-time ball nominal diameter, real-time screw thread angle and real-time contact angle of the target ball screw; Inputting the real-time torque order sequence, the real-time screw meshing frequency, the real-time nut meshing frequency and the real-time rolling element meshing frequency into a pre-trained fault detection model to determine fault information of the target ball screw.

2. The method of claim 1, wherein, The method for determining fault information of a target ball screw by inputting a real-time torque order sequence, a real-time screw meshing frequency, a real-time nut meshing frequency and a real-time rolling element meshing frequency into a pre-trained fault detection model comprises the following steps: Inputting the real-time torque order sequence, the real-time screw meshing frequency, the real-time nut meshing frequency and the real-time rolling element meshing frequency into the pre-trained fault detection model to make the fault detection model determine a convolution kernel operator according to the real-time torque order sequence, the real-time screw meshing frequency, the real-time nut meshing frequency and the real-time rolling element meshing frequency; Determining a real-time fault detection index based on the convolution kernel operator, the real-time screw meshing frequency, the real-time nut meshing frequency and the real-time rolling element meshing frequency through a convolution layer of the fault detection model; the real-time fault detection index comprises a real-time screw meshing order, a real-time nut meshing order and a real-time rolling element meshing order; Determining a target fault perception feature according to the real-time fault detection index through an attention network layer of the fault detection model; Determining fault information of the target ball screw according to the target fault perception feature through a multi-layer perceptron of the fault detection model.

3. The method of claim 2, wherein, The method for determining a target fault perception feature according to a real-time fault detection index through an attention network layer of a fault detection model comprises the following steps: Determining a target input matrix according to the real-time fault detection index, and determining a query vector matrix, a key vector matrix and a value vector matrix according to the target input matrix through the attention network layer of the fault detection model; Determining an attention matrix according to the query vector matrix and the key vector matrix, and determining an intermediate representation vector according to the attention matrix and the value vector matrix; Performing weighted processing on the real-time fault detection index based on the intermediate representation vector to determine a fault perception feature.

4. The method of claim 2, wherein, The method for determining fault information of a target ball screw according to a target fault perception feature through a multi-layer perceptron of a fault detection model comprises the following steps: Performing processing on the target fault perception feature based on a normalized activation function through the multi-layer perceptron of the fault detection model to determine a probability corresponding to a fault category of the target ball screw; the fault category comprises screw fault, nut fault, rolling element fault and no fault; Determining fault information of the target ball screw according to the probability corresponding to the fault category.

5. The method of claim 1, wherein, According to the real-time screw rotation speed, the real-time screw nominal diameter, the real-time ball nominal diameter, the real-time screw thread angle and the real-time contact angle of the target ball screw, the real-time screw engagement frequency, the real-time nut engagement frequency and the real-time rolling element engagement frequency are determined, comprising: According to the real-time screw rotation speed, the real-time screw nominal diameter, the real-time ball nominal diameter, the real-time screw thread angle and the real-time contact angle, the real-time screw engagement frequency is determined; According to the real-time screw rotation speed, the real-time screw nominal diameter, the real-time ball nominal diameter, the real-time screw thread angle and the real-time contact angle, the real-time nut engagement frequency is determined; According to the real-time screw rotation speed, the real-time screw nominal diameter, the real-time ball nominal diameter and the real-time contact angle, the real-time rolling element engagement frequency is determined.

6. The method of claim 1, wherein, Further comprising: The sample motor torque of the servo motor connected to the sample ball screw is determined, the sample torque order sequence is determined according to the sample motor torque, and the sample state information of the sample ball screw is determined; The sample state information includes screw fault state, nut fault state, rolling element fault state and normal state; The neural network model is trained according to the sample torque order sequence, the sample screw engagement frequency, the sample nut engagement frequency and the sample rolling element engagement frequency and the sample state information, and a fault detection model is determined.

7. The method of claim 6, wherein, The neural network model is trained according to the sample torque order sequence, the sample screw engagement frequency, the sample nut engagement frequency and the sample rolling element engagement frequency and the sample state information, and a fault detection model is determined, comprising: The neural network model is trained according to the sample torque order sequence, the sample screw engagement frequency, the sample nut engagement frequency and the sample rolling element engagement frequency and the sample state information, and the cross entropy loss function of the neural network model is determined in real time during the training process; The model parameters of the neural network model are updated based on the cross entropy loss function by using gradient descent method until the neural network model converges, and the trained neural network model is used as the fault detection model.

8. A ball screw failure detection device characterized by comprising: The fault detection device of the ball screw comprises: A torque order sequence determination module is configured to collect real-time motor torque of a servo motor connected to a target ball screw, and determine a real-time torque order sequence according to the real-time motor torque; An engagement frequency determination module is configured to determine real-time screw engagement frequency, real-time nut engagement frequency and real-time rolling element engagement frequency according to real-time screw rotation speed, real-time screw nominal diameter, real-time ball nominal diameter, real-time screw thread angle and real-time contact angle of the target ball screw; A fault information determination module is configured to input the real-time torque order sequence, the real-time screw engagement frequency, the real-time nut engagement frequency and the real-time rolling element engagement frequency into a pre-trained fault detection model, and determine fault information of the target ball screw. 9.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-8 when the computer program is executed by the processor. The processor executes the computer program to realize the steps of the method of any one of claims 1 to 7.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to realize the steps of the method of any one of claims 1 to 7.

Images