CN108627340B - filling production line roller bearing fault prediction method - Google Patents

Abstract

The invention discloses a fault prediction method for a roller bearing of a filling production line, which comprises the steps of measuring a vibration signal of a roller bearing of any sample by using an accelerometer, calculating by using the amplitude of the vibration signal to obtain a prediction factor, further calculating to obtain four degradation state functions, measuring the amplitude of the vibration signal of the roller bearing to be detected on the filling production line by using the accelerometer to obtain an observation sequence of the roller bearing to be detected, calculating a state probability function of the roller bearing to be detected, and further predicting whether the roller bearing to be detected has a fault.

Description

filling production line roller bearing fault prediction method

Technical Field

The invention relates to a fault prediction method for workpieces, in particular to a fault prediction method for a roller bearing of filling production lines.

Background

The of the filling production line can be widely applied to realize the batch production and the quick filling of liquid products in the fields of food, medicine, chemical industry and the like, and provides an important solution for meeting the requirements of the industry and the civilian on the liquid products.

However, at present, fault detection of the roller bearing of the filling line mainly depends on real-time diagnosis, that is, the running data of the roller bearing is monitored in real time to judge the health condition (that is, whether a fault occurs) of the roller bearing, so that detailed analysis on the working state of the roller bearing is lacked, and a fault prediction function is not provided, so that the problem that the normal running of the filling line is possibly affected cannot be eliminated in advance, which reduces the running safety of the filling line.

Disclosure of Invention

In order to solve the problems, the invention provides fault prediction methods for the roller bearings of the filling production line, which can describe the health states of the roller bearings to different degrees and realize the advanced and accurate prediction of the faults of the roller bearings by calculating the degradation state function and the state probability function, thereby improving the operation stability and reliability of the filling production line.

The technical scheme of the invention comprises the following steps:

1) measuring a vibration signal of any sample roller bearing in a full life cycle by using an accelerometer, calculating by using the amplitude of the vibration signal to obtain a prediction factor, and further calculating to obtain four degradation state functions;

the sample roller bearing is a roller bearing which is the same as the roller bearing to be tested on the filling production line, is type and is in batch with .

2) Measuring a vibration signal of a roller bearing to be measured on a filling production line by using an accelerometer to obtain an observation sequence of the roller bearing to be measured;

3) and calculating the state probability functions of the roller bearing to be tested at the time t and the time t +1, and further predicting the fault condition of the roller bearing to be tested.

Preferably, the accelerometer used in the steps 1) and 2) measures vibration signals of the roller bearing in a full life cycle.

th, second and third prediction factors k in the step 1)₁、k₂、k₃Using the following formulaAnd (3) calculating:

wherein x is_mThe amplitude of the vibration signal of the roller bearing at the mth sampling point is represented by m 1, …, 400.

The 400 sampling points are obtained by sampling at equal intervals in the whole life cycle of the sample roller bearing.

Four degradation state functions S in the step 1)₁、S₂、S₃、S₄The following formula is used for calculation:

wherein S is₁Representing a normal state, S₂Representing a minor state of degradation, S₃Representing a moderate degenerative state, S₄Representing a fault condition, k₁、k₂、k₃Respectively , a second predictor and a third predictor.

The observation sequence of the roller bearing to be measured in the step 2) is X ═ { X ═ X_t-2,x_t-1,x_t}, wherein: x is the number of_t-2、x_t-1、x_tThe amplitudes of vibration signals of the roller bearing to be detected at t-2, t-1 and t moments are respectively; t represents time, and t is more than or equal to 0.

The state probability function q (t) of the roller bearing to be tested in the step 3) at the time t and the state probability function q (t +1) of the roller bearing to be tested at the time t +1 are calculated by adopting the following formulas:

wherein: a is_ijIs the state transition probability; i, j ═ 1,2,3, 4; x is the number of_t+1The amplitude of the vibration signal of the roller bearing to be tested at the t +1 th moment is obtained.

And 3) judging whether the roller bearing to be detected fails at the moment t +1 by adopting the following modes:

when q (t +1) ═ q (t), the roller bearing to be tested is in a normal state;

when q (t +1) -q (t) is more than 0 and less than or equal to 0.2, the roller bearing to be tested is in a micro-degradation state;

when q (t +1) -q (t) is more than 0.2 and less than or equal to 0.5, the roller bearing to be detected is in a medium degradation state;

and when q (t +1) -q (t) is more than 0.5, the roller bearing to be tested is in a fault state.

The invention has the beneficial effects that:

the invention describes the health state of the roller bearing in a grading way according to the health degree, not only can realize the detailed division of the running condition of the roller bearing, but also can provide a basis for the formulation of the corresponding measures of the roller bearing.

The method can accurately and quickly predict whether the roller bearing has faults or not, thereby providing an important basis for predicting the working state of the filling production line and further providing an important guarantee for improving the operation reliability of the filling production line.

Drawings

FIG. 1 is an experimental screenshot comparing the fault prediction and the real-time fault diagnosis of the roller bearing according to the embodiment.

Detailed Description

The invention is further described in detail in connection with the figures and the embodiments.

Step 1) of the method provides an important basis for accurately predicting the health degree of the roller bearing by calculating the detailed description of the prediction factor and the health state of the roller bearing of the solid-line filling production line of the degradation state function.

In the step 2), the working state of the roller bearing to be tested is described by utilizing the observation sequence, and an important prediction basis is provided for whether the roller bearing fails or not.

In the step 3), the state probability functions of the roller bearings to be tested at the adjacent moments are calculated and compared, so that the health states of the roller bearings to be tested can be accurately and quickly predicted, and the working stability and reliability of the filling production line are improved.

The specific embodiment of the invention:

the roller bearing fault prediction method provided by the invention is tested on a filling production line test platform, and specifically comprises the following steps:

1) the radial vibration signal of the full life cycle of any sample roller bearing is measured by an accelerometer, a prediction factor is obtained by utilizing the amplitude calculation of the vibration signal, and then four degradation state functions are obtained by calculation, wherein the sample roller bearing is a roller bearing which is the same as the type of the roller bearing to be measured on the filling production line and has the same batch as .

2) Measuring a vibration signal of a roller bearing to be measured on a filling production line by using an accelerometer to obtain an observation sequence of the roller bearing to be measured;

3) and calculating the state probability functions of the roller bearing to be tested at the time t and the time t +1, and further predicting the fault condition of the roller bearing to be tested.

Experimental data are acquired and processed through a filling production line monitoring system, and the experimental data obtained by adopting the prediction method provided by the invention are as follows: the failure prediction time is less than 6.5 multiplied by 10⁴min, much less than fault diagnosis time (8 × 10)⁴min)。

The experimental screenshot is as follows (the black solid line in the figure is a real-time fault diagnosis method, and the black dotted line is a fault prediction method provided by the invention):

as can be seen from fig. 1: the real-time fault diagnosis method needs to consume 8 multiplied by 10 times⁴min can detect the fault state of the roller bearing, and a more detailed description of the health state of the roller bearing is lacked; compared with the real-time fault diagnosis method, the fault prediction method provided by the invention can be 6.2 multiplied by 10⁴The fault state of the roller bearing is predicted in advance at min, and can be described more carefully, which can provide important basis for the efficient management and operation of the filling production line.

The foregoing detailed description is intended to illustrate and not limit the invention, which is intended to be within the spirit and scope of the appended claims, and any changes and modifications that fall within the true spirit and scope of the invention are intended to be covered by the following claims.

Claims (4)

1, kinds of filling line roller bearing fault prediction methods, characterized by including the following steps:

1) measuring a vibration signal of any sample roller bearing in a full life cycle by using an accelerometer, calculating by using the amplitude of the vibration signal to obtain a prediction factor, and further calculating to obtain four degradation state functions;

four degradation state functions S in the step 1)₁、S₂、S₃、S₄The following formula is used for calculation:

wherein S is₁Representing a normal state, S₂Representing a minor state of degradation, S₃Representing a moderate degenerative state, S₄Representing a fault condition, k₁、k₂、k₃, a second predictor and a third predictor respectively;

2) measuring a vibration signal of a roller bearing to be measured on a filling production line by using an accelerometer to obtain an observation sequence of the roller bearing to be measured;

the observation sequence of the roller bearing to be measured in the step 2) is X ═ { X ═ X_t-2,x_t-1,x_t}, wherein: x is the number of_t-2、x_t-1、x_tThe amplitudes of vibration signals of the roller bearing to be detected at t-2, t-1 and t moments are respectively; t represents a time;

3) calculating state probability functions of the roller bearing to be tested at the time t and the time t +1 according to the degradation state function and the observation sequence, and predicting the fault condition of the roller bearing to be tested;

the state probability function q (t) of the roller bearing to be tested in the step 3) at the time t and the state probability function q (t +1) of the roller bearing to be tested at the time t +1 are calculated by adopting the following formulas:

wherein: a is_ijIs the state transition probability; i, j ═ 1,2,3, 4; x is the number of_t+1The amplitude of the vibration signal of the roller bearing to be tested at the t +1 th moment is obtained.

2. The method for predicting roller bearing failure in bottling line according to claim 1, wherein the accelerometer in steps 1) and 2) is used to measure vibration signals of the roller bearing during its full life cycle.

3. The method for predicting roller bearing failure in bottling line according to claim 1, wherein the , the second and the third prediction factors k in step 1)₁、k₂、k₃The following formula is used for calculation:

wherein x is_mThe amplitude of the vibration signal of the roller bearing at the mth sampling point is represented by m 1, …, 400.

4. The method for predicting the roller bearing fault of the bottling line according to claim 1, wherein the step 3) is performed by determining whether the roller bearing to be tested has a fault at the time t +1 in the following manner:

when q (t +1) ═ q (t), the roller bearing to be tested is in a normal state;

when q (t +1) -q (t) is more than 0 and less than or equal to 0.2, the roller bearing to be tested is in a micro-degradation state;

when q (t +1) -q (t) is more than 0.2 and less than or equal to 0.5, the roller bearing to be detected is in a medium degradation state;

and when q (t +1) -q (t) is more than 0.5, the roller bearing to be tested is in a fault state.