CN103971025B - A kind of fault of numerical control machine tool correlationship Dynamic Variation Analysis method - Google Patents

Abstract

The invention discloses a method for analyzing the dynamic change of the correlation relationship between CNC machine tools; it aims to overcome the problem of the correlation degree between multiple systems with intricate correlations in the fault chain of the mutual interference fault (IF) that cannot be determined in the prior art , the steps are: Step 1: Use FMECA analysis technology to process the fault data, divide the fault parts of each subsystem of the machine tool, and sort out the data with related faults among each subsystem; Step 2: Analyze the relevant data, summarize the relevant sub-systems The interaction form between systems defines the types of fault chains and the elements of fault chains; Step 3: For different related fault chains, use the dependence of independent failure rate, related failure rate and comprehensive failure rate to obtain related fault subsystems The comprehensive failure rate, respectively establish the correlation coefficient calculation model of all failure chains, and form the correlation coefficient model system of related failures; Step 4: Analysis of maintenance strategies considering related failures.

Description

A Method for Analyzing Dynamic Changes of Correlation Relationships between CNC Machine Tool Faults

technical field

The invention belongs to the technical field of numerical control machine tools, and relates to a method for analyzing related faults of numerically controlled machine tools, in particular to a method for analyzing dynamic changes of related relations of numerically controlled machine tools.

Background technique

The existence of faults related to CNC equipment has caused changes in its inherent failure rate. If the reliability design and allocation stage ignores the fault correlation between subsystems, the reliability design value will inevitably have a large error with the actual production value. At the same time, the existence of related faults, if predictive maintenance is carried out with the reliability of independent faults, will also cause a delay in the maintenance time node. There are four most typical correlation faults in system faults: series faults, negative correlation faults, common cause failures and mutual interference faults. The present invention relates to the type of mutual interference failures (Interactive failures, referred to as I.F). The common feature of this type of related failures is: when component A fails, it will accelerate or cause the failure of component B. Sometimes this destructive effect is mutual. As a result of this interaction, the failure rate of the faulty system increases, and the magnitude of the increase is related to the degree of interaction between subsystems. Improper design, manufacturing, installation and operation of machine tools may cause such related failures, so the probability of occurrence is relatively high and the hazards are relatively strong. With domestic and foreign scholars paying attention to related faults in recent years, relevant research has been deepened continuously, but there are no reports on the types of mutual interference faults that commonly exist between subsystems on CNC machine tools.

The degree of mutual interference between subsystems is defined as the correlation coefficient. There are few literatures that focus on the degree of correlation, and the only literature is to roughly estimate it as a parameter in order to achieve other reliability indicators, ignoring its importance and rigor in reliability calculations. Although the existing analysis methods and models have realized the determination of the correlation coefficient to a certain extent, they all have certain theoretical limitations. The main analysis methods are: 1. Mathematical statistics method, using the relevant failure rate as the correlation coefficient, so the number of samples collected determines the change of the correlation coefficient, and the error is large; 2. Experimental method, using the test method to obtain a large number of The experimental data determine the correlation coefficient model. This method is highly pertinent, and the implementation cost is high, and it is not universal; 3. Subjective assignment method, experienced experts score the degree of correlation between related systems, and calculate the comprehensive score value as the correlation coefficient. This method is highly subjective, and the assignment error is relatively large; 4. Copula function method, the Copula function describes the correlation between variables, and the function that connects the variable joint cumulative distribution function with the variable marginal cumulative distribution function can be used The relevant fault data of the subsystems calculates the correlation coefficient between the subsystems, and this correlation coefficient is common and unique. The calculation of this method is relatively complicated, and the interaction relationship and direction of action between subsystems cannot be clarified, so the model cannot realize the analysis of the intricate correlation relationship between multiple systems and the establishment of multiple correlation coefficient models; 5. Narrow bound theory method , using the functional functions of the main failure modes of the relevant subsystems to calculate the correlation coefficient, this method is only suitable for the multi-mode correlation of the components themselves, because it is impossible to determine the correlation of the functional functions of the main failure modes between multiple systems, it is difficult to implement Calculation of the correlation coefficient between multiple systems; 6. Failure rate method, establish the relational expression of the failure rate between related subsystems, derive the correlation coefficient, and determine the degree of correlation of the subsystems. But the research of the existing failure rate method ends at the calculation of the correlation coefficient between the two subsystems.

Among the analysis methods for the degree of correlation between all relevant subsystems, the failure rate method uses the relationship between the independent failure rate and the related failure rate to reflect the degree of correlation more accurately, because the method not only has quantitative results, but also can clarify the relationship between the subsystems. The direction of interaction between systems. However, the existing analysis process has theoretical limitations. This analysis method can only determine the degree of correlation between two related subsystems. One, so the failure data of the same subsystem may come from different related subsystems, which makes further analysis difficult.

Contents of the invention

The purpose of the present invention is to overcome the problem that the existing technology cannot determine the correlation degree between multiple systems with intricate correlations in the fault chain of mutual interference faults (I.F), and to provide a dynamic change analysis method for the fault correlations of CNC machine tools .

In order to achieve the above object, the technical solution provided by the present invention is a method for analyzing the dynamic change of the fault correlation relationship of a CNC machine tool, comprising the following steps:

Step 1: Use the FMECA analysis method to process the fault data, conduct statistical analysis on the fault data of each subsystem of the machine tool, and sort out the data with related faults between each subsystem.

Step 2: Analyze the relevant fault data, summarize the interaction forms of the mutual interference fault I.F fault types between relevant subsystems, and define the types of fault chains and elements of fault chains.

Step 3: For different related fault chains, use the dependent relationship of independent failure rate, related failure rate and comprehensive failure rate, and use the comprehensive failure rate model of related fault subsystems to deduce the correlation coefficient calculation model of all fault chains, all The correlation coefficient model of the fault chain constitutes the correlation coefficient model system of related faults.

Step 4: Analysis of maintenance strategies considering relevant failures.

The types of fault chains described in the technical proposal include five:

The first type: the fault chain has only two subsystems, and they act in one direction. During the operation of the machine tool, the subsystems The adverse motion state will affect the subsystem , until the subsystem failure occurs while the subsystem pair subsystem have no effect;

The second type: the fault chain has only two subsystems, and the two subsystems interact, and the bad motion states of the two subsystems affect each other, forming a vicious circle until one of the systems terminates due to a fault;

The third type: the fault chain is composed of multiple subsystems, and the subsystems The motion state of the system affects multiple subsystems at the same time, but itself is not affected by other related subsystems;

The fourth type: the fault chain is composed of multiple subsystems, and the subsystems is affected by more than one subsystem and has no relevant effect on any other subsystem;

The fifth type: the fault chain is composed of more than three subsystems, which is a combination of part or all of the above four basic fault chains, at least one subsystem is affected by more than two subsystems, and there is a fault intermediate point subsystem , an intricate correlation relationship is formed among the related subsystems, which belongs to the type of complex correlation fault chain.

Defining the fault chain elements described in the technical solution refers to defining the fault subsystem according to the position and role of the fault subsystem in the fault chain during the fault occurrence process; the fault chain elements include:

(1) Related fault starting point:

In a fault chain with a correlation relationship, a subsystem that only affects other subsystems but is not affected by other subsystems is called a correlation fault origin;

(2) Related failure endpoints:

In a fault chain with a correlation relationship, the subsystem that is only affected by other subsystems but not other subsystems is called the correlation fault end point;

(3) Intermediate point of failure:

In the fault chain with correlation relationship, the subsystems that have both influence and affected relationships are called fault intermediate points.

According to the different related fault chains mentioned in the technical proposal, the correlation coefficient calculation models of all fault chains are deduced respectively by using the interdependence relationship of independent failure rate, related failure rate and comprehensive failure rate, and the comprehensive failure rate model of related fault subsystems , specifically based on the comprehensive failure rate calculation model:

(1)

: for the relevant fault subsystem The comprehensive failure rate of is calculated from the failure data in production;

: for the subsystem The independent failure rate of , which is determined by the inherent reliability, is obtained by testing or production data before the product leaves the factory. In the subsystem Without being affected by related faults, theoretically ;

: for the subsystem acceptor system The correlation coefficient of the effect, ,when , there is no correlation, that is, the subsystem failure will not cause failure occurs when When , it is completely related, that is, the subsystem Failure must cause malfunction;

: for the pair subsystem Subsystems that play a role relative failure rate;

According to the comprehensive failure rate model, according to the formula (1), according to the characteristics of the failure chain, the correlation coefficient calculation models of all the failure chains are respectively established, and the correlation coefficient calculation models of all the failure chains form the correlation coefficient model system of related faults.

For the first and third types of fault chains in the technical solution, the correlation relationship of the fault chain is one-way correlation, and the relevant fault terminal subsystem only one subsystem effect, the associated fault subsystem The comprehensive failure rate of is obtained by formula (1):

(2)

Then there are:

(3)

due to related subsystems It is the starting point of related faults and is not affected by other subsystems, so ;in:

: for the subsystem independent failure rate;

: for the relevant fault subsystem comprehensive failure rate;

: for the subsystem acceptor system The correlation coefficient of the effect.

For the fourth type of fault chain in the technical solution, the correlation relationship of the fault chain is multi-system unidirectional correlation, and the related fault end point Simultaneously affected by multiple subsystems, while is the relevant fault end point, the relevant fault end point The operating state does not affect other related subsystems; a subsystem comprehensive failure rate Depend on A correlation decision, for the computing subsystem acceptor system Correlation coefficient of action value, make the following assumptions:

(1) Subsystem and correlation coefficient There is no linear relationship between ;make When , formulas (2) and (3) are established, and formula (4) is deduced from formula (2):

(4)

for the subsystem Subject to the first The sub-failure rate of the relevant effects of each subsystem; by the subsystem faulty data removal Modeling related fault data of other subsystems.

(2) Subsystem Subject to the first Fractional Failure Rate and subsystem Comprehensive failure rate into a functional relationship:

(5)

subsystem Both are related fault starting points, for subsystems The relative failure rate of the subsystem is equal to The independent failure rate of ; then in formula (4) , , ,thus It can be found that according to formulas (1) and (4), formula (5) can be deduced as formula (6):

 

(6)

: for the subsystem acceptor system The correlation coefficient of the effect.

For the second and fifth types of failure chains in the technical solution, the correlation of the failure chains is multi-system correlation. When calculating the comprehensive failure rate of each subsystem, it is necessary to separately calculate the relationship between the subsystem and other related subsystems. Correlation failure rate and correlation coefficient; set the names of the subsystems as: correlation failure starting point , the associated failure endpoint , the associated failure intermediate point , then the relevant failure endpoint The comprehensive failure rate model of is as follows:

(7)

Subsystem obtained by formula (4) The fractional failure rate model of is:

(8)

: for acceptor system Affected Subsystems relative failure rate;

: for acceptor system Affected Subsystems relative failure rate;

: for the pair subsystem Subsystems that play a role relative failure rate;

: for the pair subsystem Subsystems that play a role relative failure rate;

for the subsystem acceptor system The correlation coefficient of the effect;

for the subsystem acceptor system The correlation coefficient of the effect;

, Calculate the fractional failure rate calculation as described; is the relative fault starting point, so ,but can be obtained.

The pair subsystem described in the technical proposal Subsystems that play a role The relative failure rate of The value analysis process of is as follows, taking the subsystem For the research object, the subsystem is the intermediate point of failure, for the subsystem affected by the subsystem influence, the subsystem The comprehensive failure rate model of is:

(9)

: for the subsystem comprehensive failure rate;

: for the subsystem independent failure rate;

: for the subsystem acceptor system The correlation coefficient of the effect;

: for the pair subsystem Subsystems that play a role relative failure rate;

because is the relevant fault starting point, so , Depend on The production failure data obtained, for the subsystem The independent failure rate of is known, It can be determined by formula (9); according to and relevant or not, for the subsystem acceptor system The correlation coefficient of the effect, There are two cases of values,

The first: , and irrelevant

The second type: , and relevant

The value is determined, According to the formula (8) can be obtained.

The analysis of the maintenance strategy considering the relevant faults mentioned in the technical proposal is specifically to use the determined correlation coefficient model system between the subsystems to revise the reliability of the subsystems, and calculate the predictive maintenance nodes;

The maintenance strategy of the system is formulated and optimized based on the reliability. When the reliability of the system or subsystem is less than the threshold value specified in the plan, it is necessary to perform preventive maintenance on the system or subsystem; if the subsystem There is no related fault in the fault, then its reliability is:

(11)

: for the subsystem Reliability models with only independent failure rates;

Consider subsystems with associated faults The reliability model of is:

(12)

F:

(13)

: for the subsystem The reliability model established by the comprehensive failure rate;

The reliability calculated by the independent failure rate may be too large, when the subsystem The reliability reaches the threshold specified by the subsystem When the time point calculated by the reliability model of is used as the maintenance node, since the time node calculated by the independent failure rate is greater than the time node calculated by the related failure rate, the subsystem calculated by formula (11) When the maintenance time node is used as the basis of the maintenance plan, it will inevitably lead to the delay of the preventive maintenance plan.

Compared with prior art, the beneficial effect of the present invention:

1. A kind of numerically controlled machine tool fault correlation dynamic change analysis method according to the present invention is based on investigating a large amount of fault data, seeks the law of the data change of I.F correlation fault, is based on the correlation form between subsystems to I.F Fault types are divided into fault chain types, which provides an important analysis basis for the analysis of related faults.

2. A kind of numerically controlled machine tool failure correlation dynamic change analysis method according to the present invention takes the failure rate method as the analysis means, and is based on the dependence between the independent failure rate, the correlation failure rate and the comprehensive failure rate, and aims at different I.F. According to the characteristics of the correlation fault chain, the correlation coefficient model between subsystems is determined, the correlation coefficient model system is formed, and the correlation relationship and correlation degree analysis method between complex multi-systems is established.

3. The method for analyzing the dynamic change of the correlation relationship between CNC machine tool failures described in the present invention breaks through the theoretical limitations of the prior art and expands the theoretical system of related failures. Using the analysis method of the present invention to optimize and predict the maintenance plan improves the accuracy of maintenance nodes and reduces the failure rate of equipment, thereby providing a method and a theoretical basis for increasing the reliability of equipment.

Description of drawings

Fig. 1 is a flow chart of a kind of numerically controlled machine tool fault correlation dynamic change analysis method according to the present invention;

Fig. 2 is the first kind of fault chain of the five kinds of fault chains described in the dynamic change analysis method of a kind of numerically controlled machine tool fault correlation relationship of the present invention, that is, the one-way action schematic diagram of two subsystems;

Fig. 3 is the second kind of failure chain of the five kinds of failure chains described in the dynamic change analysis method of a kind of numerically controlled machine tool fault correlation relationship of the present invention, that is, the interaction diagram of two subsystems;

Fig. 4 is a schematic diagram of the third fault chain of the five fault chains, that is, a single subsystem acting on multiple subsystems at the same time, in a method for analyzing dynamic changes in the fault correlation of a numerically controlled machine tool according to the present invention;

Fig. 5 is a schematic diagram of the fourth fault chain of the five fault chains described in the present invention in a method for analyzing dynamic changes in fault correlations of CNC machine tools, that is, a single subsystem is simultaneously affected by multiple subsystems;

Fig. 6 is a kind of schematic diagram of the fifth fault chain of the five fault chains described in the present invention, which is a complex correlation relationship;

Fig. 7 is a correlation diagram of the servo system, the hydraulic system and the tool rest system in a dynamic change analysis method of the fault correlation of a CNC machine tool.

Detailed ways

The following is a detailed description of a method for analyzing the dynamic change of the fault correlation relationship of a CNC machine tool in conjunction with the accompanying drawings:

The method for analyzing the dynamic change of the fault correlation relationship of a CNC machine tool according to the present invention is based on a large number of investigation fault data, explores the law of the data change of the relevant fault, and classifies the fault chain for the interaction form between the subsystems, according to Different failure chains Find out the interdependent relationship among independent failure rate, related failure rate and comprehensive failure rate, determine the correlation coefficient model between subsystems, use the failure rate method as the analysis basis, and establish the correlation relationship between complex multi-systems And the analysis method of correlation degree breaks the theoretical limitation of the existing method, and optimizes and predicts the maintenance plan with this analysis method.

A method for analyzing the dynamic change of the fault correlation relationship of a CNC machine tool includes the following steps:

Step 1. Use FMECA analysis technology to process the fault data, conduct statistical analysis on the fault data of each subsystem of the machine tool, and sort out the data with related faults between each subsystem.

1) According to the characteristics of the failure mode, the location of the failure and the cause of the failure, the fault belongs to, and the frequency of the failure of the subsystem and the severity and hazard of the failure mode are determined.

2) Calculate the failure interval time of the whole machine and the failure interval time of each subsystem.

3) Screen fault data with correlations to determine the interaction form between subsystems.

Step 2. Analyze relevant fault data, summarize the mutual interference fault (I.F) interaction form between related subsystems, and define the types of fault chains and elements of fault chains:

1) Define the type of fault chain:

Referring to Fig. 2, Fig. 3, Fig. 4, Fig. 5 and Fig. 6, the intricate correlation between the multiple subsystems of the CNC machine tool is removed and simplified, the complex surface phenomena are revealed, and five basic types of mutual interference faults are summarized. form. Assuming that the CNC machine tool consists of The subsystems are composed of , any complex related fault relationship can be composed of these five or several fault chains.

Refer to Figure 2, the first type: the fault chain has only two subsystems, and it acts in one direction, that is, the subsystems The adverse motion state will affect the subsystem , until failure occurs while the subsystem pair subsystem have no effect;

Refer to Figure 3, the second type: there are only two subsystems in the fault chain, and the two subsystems interact, and the bad motion states of the two subsystems affect each other, forming a vicious circle until one of the systems terminates due to a fault;

Refer to Figure 4, the third type: the fault chain is composed of multiple subsystems, and the subsystems The motion state of the system affects multiple subsystems at the same time, but itself is not affected by other related subsystems;

Refer to Figure 5, the fourth type: the fault chain is composed of multiple subsystems, and the subsystems Affected by multiple subsystems without relevant failure effects on any other system;

Refer to Figure 6, the fifth type: the failure chain is composed of three or more subsystems, which is the simplest combination form of the aforementioned four basic failure chains, and belongs to the type of complex correlation failure chain.

2) Define the fault chain elements:

The fault subsystem is defined according to its position and role in the fault occurrence process in the fault chain. The present invention is defined as follows:

1. Related fault starting point:

In the fault chain of the correlation relationship, the subsystem that only affects other systems and is not affected by other systems is called the relevant fault starting point, as shown in Figures 2, 4, 5 and 6 ;

2. Related failure endpoints:

In the failure chain of the correlation relationship, the subsystem that is only affected by other subsystems and does not affect other subsystems is called the relevant failure terminal, as shown in Figures 2, 4, 5, and 6 ;

3. Relevant failure intermediate points

In the failure chain of the correlation relationship, the subsystems that have the influence and the affected relationship at the same time are called the middle point of the correlation failure, as shown in Figure 4 , in Figure 3 , .

Step 3. Aiming at the characteristics of different related failure chains, the model system of correlation coefficient is established by using the dependence of independent failure rate, related failure rate and comprehensive failure rate:

1) According to the characteristics of different related failure chains, the model system of the correlation coefficient is established by using the interdependence of the independent failure rate, the related failure rate and the comprehensive failure rate. Specifically, it is based on the relevant failure rate calculation model:

(1)

: for the relevant fault subsystem The comprehensive failure rate of is calculated from the failure data in production;

: for the subsystem The independent failure rate of , determined by the inherent reliability, is usually obtained by testing or production data before the product leaves the factory, in the subsystem Without being affected by related faults, theoretically ;

for the subsystem acceptor system The correlation coefficient of the effect, ,when , there is no correlation, that is, the subsystem failure will not cause failure occurs when When , it is completely related, that is, the subsystem Failure must cause malfunction;

: for the pair subsystem Subsystems that play a role related failure rate.

2) According to the characteristics of different related failure chains, a model system of correlation coefficients is established by using the interdependence of independent failure rate, related failure rate and comprehensive failure rate. Specifically, according to formula (1), according to the characteristics of the five fault chains, the correlation coefficient calculation models of all fault chains are established respectively, and the correlation coefficient model system of related faults is formed as follows:

1. The correlation relationship of the fault chain is a simple one-way correlation, that is, the first and third types of the fault chain (see Figure 2 and Figure 4), and the related fault terminal subsystem only one subsystem , then the relevant failure terminal subsystem The comprehensive failure rate of is obtained by formula (1):

(2)

Then the correlation coefficient calculation model is:

(3)

due to related subsystems It is the starting point of related faults and is not affected by other subsystems, so , for the subsystem independent failure rate. for the subsystem acceptor system The correlation coefficient of the effect;

2. The correlation relationship of the fault chain is multi-system one-way correlation, that is, the fourth type of fault chain, as shown in Figure 5, the related fault end point Simultaneously affected by multiple subsystems, while is the relevant fault end point, the relevant fault end point The operating state does not affect other related subsystems. Comprehensive failure rate Depend on A correlation decision, as shown in formula (1). for computing subsystem acceptor system Correlation coefficient of action value, while satisfying the following two assumptions (1) and (2):

(1) Subsystem and correlation coefficient There is no linear relationship between . Immediately When formulas (2) and (3) are established, the correlation diagram in Figure 4 can be decomposed into A relational expression as shown in Figure 1, formula (4) is deduced from formula (2).

(4)

for the subsystem Subject to the first The failure rate of the relevant effects of each subsystem, based on the assumption (1), from the subsystem Exclude subsystems from fault data Subsystem-to-subsystem The relevant fault data modeling can be obtained.

(2) Subsystem Subject to the first Fractional Failure Rate and subsystem Comprehensive failure rate into a functional relationship.

(5)

Taking the relationship form in Figure 4 as an example, the subsystem Both are related fault starting points, for subsystems The relative failure rate of the subsystem is equal to independent failure rate. In formula (4) , , ,thus It can be found that according to formulas (1) and (4), formula (5) can be deduced as formula (6):

(6)

3. The correlation of fault chains is multi-system complex correlation, that is, the fifth and second fault chains, as shown in Figures 3 and 6 (Figure 3 can be regarded as a special case of Figure 6, and are regarded as fault intermediate points), and each subsystem in the figure has more than two correlations. When calculating the failure rate of each subsystem, it is necessary to obtain the relevant failure rate and correlation coefficient of this subsystem and other related subsystems. Taking the subsystem in Figure 6 Taking the failure rate calculation as an example, the names of the subsystems are respectively: the relevant failure starting point , the associated failure endpoint , the relevant intermediate point , according to formula (1) we have

(7)

By formula (4) get

(8)

: for acceptor system Affected Subsystems relative failure rate; : for acceptor system Affected Subsystems relative failure rate; : for the pair subsystem Subsystems that play a role relative failure rate; : for the pair subsystem Subsystems that play a role relative failure rate; : for the subsystem acceptor system The correlation coefficient of the effect; : for the subsystem acceptor system The correlation coefficient of the effect.

, Calculate the calculation of the fractional failure rate as described in formula (4); in formula (8) is the relative fault starting point, so ,but can be obtained; The determination of is relatively complicated because the subsystem is the middle point of failure, for while being affected by The influence of , its value analysis process is as follows, first of all with For the research object, the subsystem The combined failure rate is:

(9)

: subsystem comprehensive failure rate; : subsystem independent failure rate; for the subsystem acceptor system Correlation coefficient of action: : for the pair subsystem Subsystems that play a role related failure rate.

because is the relevant fault starting point, so , Depend on The production failure data obtained, for the subsystem The independent failure rate of is known, It can be determined by formula (9). according to and relevant or not, The value is divided into two cases

① , and irrelevant

② , and Related (10)

The value is determined, According to the formula (8) can be obtained.

So far, the correlation coefficient calculation model system between multiple subsystems of the IF fault type of the complex system has been established. The correlation coefficient is time function, the associated fault causes the subsystem After a fault occurs, the interaction stops, and after the fault is repaired, the CNC equipment continues to operate, and the related role starts again. This process repeats itself, so the correlation relationship exists during each fault interval, if and only if , , the correlation coefficient is a fixed value. in, : for the subsystem The reliability model established by the comprehensive failure rate.

Step 4. Analysis of maintenance strategies considering relevant faults

1) The analysis of the maintenance strategy considering related faults, specifically, using the determined correlation coefficients between the subsystems to revise the reliability of the subsystems, and calculate the predicted maintenance nodes.

The maintenance strategy of the system is formulated and optimized based on reliability. When the reliability of the system or subsystem is less than the threshold value specified in the plan, preventive maintenance of the system or subsystem is required. if the subsystem The faults are independent faults, then the reliability model is:

(11)

: for the subsystem The reliability model in the case of only independent failure rates;

The inevitability of the existence of related faults in actual production has led to large deviations in the calculation of reliability, and this deviation is more obvious when calculating the reliability of the whole machine. Taking the IF related faults studied by the present invention as an example, take The reliability of independent fault calculation has a tendency to be too large, considering the subsystems of related faults The reliability model of is:

(12)

: for the subsystem The reliability model established by the comprehensive failure rate;

F :

Since the reliability value calculated with independent faults is greater than the reliability value considering related faults, formula (11) leads to the delay of preventive maintenance planning.

The method of the present invention is through the analysis of a large number of CNC machine tool failure data, according to the action form of related failures, sums up the failure chain types of mutual interference failure models, and according to the dependent relationship of independent failure rate, related failure rate and comprehensive failure rate in related failures , establish a correlation failure rate model, and then according to the characteristics of each failure chain, establish a correlation coefficient solution model system between related subsystems, thereby realizing an analysis method for determining the degree of correlation between related subsystems of CNC machine tools, for A basis is provided for predictive maintenance planning of systems or subsystems that considers relevant failures.

Example:

The present invention takes the complex correlation analysis among the three subsystems of the CNC machine tool as an example.

Step 1: Use FMECA analysis technology to process the fault data, divide the fault parts of each subsystem of the machine tool, and sort out the data with related faults.

The data comes from the 13-month production failure tracking records of 175 machine tools of the same model. After the fault data is processed by FMECA analysis technology, the subsystems with correlation relationship are found. Refer to the correlation relationship in Figure 7 as an example.

After processing by FMECA method, the failure time points and related failure time points of each subsystem are obtained. The failure interval time value is calculated from the failure time points as the failure data. The daily working system of the machine tool is two shifts. Table 1 shows the data of the tool holder system process result. Its calculation formula is as follows:

(13)

(14)

: No. A fault interval time, the category is relevant, and the fault interval end point is a relevant fault time point, such as 28.60 ^* of the tool rest system fault data in Table 1; : No. fault occurrence time point, and this fault is a relevant fault; : the same subsystem of the same machine tool The previous failure time point immediately before the failure time point; : the same subsystem of the same machine tool the next time between failures; The same subsystem on the same machine tool The next failure time point immediately after the failure time point.

Table 1 Fault data processing of tool holder system

The data processing of the servo system and the hydraulic system is like the data processing process of the tool holder system, and the value of the fault interval time after processing is shown in Table 2.

Table 2 Subsystem failure interval time

The servo system in Figure 7 is the starting point of the relevant fault, and is not affected by other systems. In the fault data of the hydraulic system in Table 2, the data with "*" indicates that the fault interval time is caused by the relevant fault of the servo system; the data with "※" in the fault data of the tool post system indicates that the fault interval time is caused by the relevant fault of the hydraulic system ;The data with "#" indicates that the fault interval time is caused by the fault related to the servo system.

Step 2: Identify the type of fault chain according to the interaction form between subsystems:

According to the complex correlation relationship of the three subsystems shown in Figure 7, this fault chain belongs to the fifth type of fault chain, which is the simplest complex correlation form, and the correlation coefficient value is determined by modeling according to formulas (7) to (10).

Step 3: For the types of related fault chains determined in step 2, use the corresponding correlation coefficient model to analyze and solve.

1. The independent failure rate model, the comprehensive failure rate model and the related failure rate model of the three subsystems are obtained from the production failure data.

Each subsystem of the machine tool eliminates the interference of related failure factors, and the failure rate is its own inherent independent failure rate under normal standard production status.

Table 3 shows the independent failure rate functions of the above three subsystems, and the reliability functions conform to the Weibull distribution.

Table 3 Subsystem independent failure rate function

The comprehensive failure rate of the three subsystems is obtained by processing the failure data collected at the production site. The Weibull distribution is a hypothetical distribution model, and the D test method is used to test the distribution fit. Such as formula (18)~(20) in Table 4.

Table 4 Subsystem comprehensive failure rate function

This case takes the correlation coefficient modeling calculation deduction process of the tool holder system and the other two subsystems obtained from the correlation relationship of the three subsystems in Figure 7 as an example. set up is the relative failure rate of the tool post system affected by the servo system, which is calculated by removing the fault data of the hydraulic system on the tool post system from the fault data of the tool post system, as shown in formula (21); is the relative failure rate of the tool post system affected by the hydraulic system, which is calculated by removing the fault data of the servo system on the tool post system from the tool post fault data, as shown in formula (22).

Table 5 Tool holder system-related failure rate function

2. Correlation analysis of tool holder system:

Determine the relevant subsystems of the tool holder system according to the correlation diagram 7. According to the formula (7), the comprehensive failure rate formula of the tool holder system is as follows: (twenty three)

in is the relative failure rate of the servo system to the tool post system; is the relative failure rate of the hydraulic system to the tool holder system. According to the formula (8), the formula (23) is decomposed into the relevant failure rate model of the tool holder system and the servo system and the hydraulic system, such as formula (24).

(twenty four)

The fault correlation coefficient of the servo system to the tool post system is:

(25)

Since the servo system is the starting point of the fault, so

(26)

Substitute failure rate models (15), (17), (21) into equation (25), if expression substitution of point estimate , then there is , .

The model of the relevant failure coefficient of the hydraulic system to the tool holder system is:

Obtained by formula (23)

(27)

in , is the known failure rate, To be determined. Further analysis, the related failure of the hydraulic system to the tool post system is related to the failure chain of the servo system to the hydraulic system, that is, the correlation coefficient of the servo system to the hydraulic system and Related, the failure rate of the hydraulic system affected by the servo system will affect the change of the failure rate of the tool holder system. According to the formula (10):

(28)

Substituting failure rate models (17), (19), (22) into equation (27), if substituting of point estimate , then there is , .

3. According to the correlation coefficient obtained above, according to the formula (23), the comprehensive failure rate of the tool post system is:

Step 4: Analysis of maintenance strategies considering relevant failures.

According to Equation (11), the predictive maintenance of the tool holder system analyzed by independent faults is as follows: , when the given reliability threshold is When , the predicted maintenance time is Hours; the predicted repair time for the tool holder system considering related failures is as follows: , when the given reliability threshold is When , the predicted maintenance time is Hours; from the analysis of the maintenance strategy considering the related faults, it can be seen that the delay of the maintenance time node is ignored and the independent reliability is used as the design basis for the related faults. As a result, the subsystems or components of the machine tool cannot be repaired in a timely manner within the threshold of reliability decline, so the failure rate of the machine increases and the reliability decreases. At the same time, it also affects the formulation of order quantities and order dates for spare parts.

Claims (1)

1. A method for analyzing dynamic changes of numerical control machine tool fault correlation is characterized by comprising the following steps:

step 1: processing fault data by using an FMECA analysis method, performing statistical analysis on the fault data of each subsystem of the machine tool, and sorting the data with related faults among the subsystems;

step 2: analyzing related fault data, summarizing and summarizing the action form of the mutual interference fault I.F fault types among related subsystems, and defining the types and elements of fault chains;

and step 3: aiming at different related fault chains, respectively deducing correlation coefficient calculation models of all fault chains by utilizing the independent fault rate, the dependent relation of the related fault rate and the comprehensive fault rate and utilizing the comprehensive fault rate model of a related fault subsystem, wherein the correlation coefficient models of all fault chains form a correlation coefficient model system of the related fault;

and 4, step 4: analysis of a maintenance strategy that takes into account the associated fault;

the types of fault chains include five:

the first method comprises the following steps: the fault chain only has two subsystems, has a unidirectional function, and is a subsystem in the running process of the machine toolCan affect the subsystemUp to the subsystemIn case of failure, the subsystemSubsystem pairNo influence is generated;

and the second method comprises the following steps: the fault chain is only provided with two subsystems, the two subsystems interact with each other, the bad motion states of the two subsystems are mutually influenced to form a vicious circle until one system stops running due to the fault;

and the third is that: the fault chain is composed of a plurality of subsystems, and the subsystemsThe motion state of (2) simultaneously affects a plurality of subsystemsIs not influenced by other related subsystems;

and fourthly: the fault chain is composed of a plurality of subsystems, and the subsystemsAffected by multiple subsystems without any relevant effect on any other subsystem;

and a fifth mode: the fault chain is composed of more than three subsystems, is a combination form of part or all of the four basic fault chains, at least one subsystem is subjected to the correlation action of more than two subsystems, and has a fault intermediate point subsystem, and complicated correlation relations are formed among the related subsystems and belong to the type of the complicated related fault chain;

the definition of the fault chain elements refers to the definition of the fault subsystem according to the position and the action of the fault subsystem in the fault chain in the fault occurrence process; the fault chain elements include:

(1) starting point of related fault:

in a fault chain with correlation, a subsystem which only affects other subsystems but is not affected by other subsystems is called a correlation fault starting point;

(2) end of related failure:

in a fault chain with a correlation relationship, a subsystem which is only influenced by other subsystems and does not influence the other subsystems is called a correlation fault terminal;

(3) the fault intermediate point:

in a fault chain with a correlation relationship, a subsystem with the influence and influenced relationship is called as a fault intermediate point;

for different related fault chains, the correlation coefficient calculation models of all fault chains are respectively deduced by using the independent fault rate, the dependent relation of the related fault rate and the comprehensive fault rate and using the comprehensive fault rate model of the related fault subsystem, specifically according to the comprehensive fault rate calculation models:

（1）

: for the subsystem in the case of a relevant faultThe comprehensive fault rate is obtained by calculating fault data in production;

: as subsystems ofThe independent failure rate is determined by inherent reliability, and the product is obtained through test or production data before leaving factory and is in a subsystemWithout being affected by the associated fault, theoretically；

: as subsystems ofReceptor systemThe correlation coefficient of the effect is such that,when is coming into contact withTime, without correlation, i.e. subsystemsWill not cause failureIn case of failure whenTime, fully correlated, i.e. sub-systemThe occurrence of a fault necessarily causesA failure occurs;

: to a subsystemSubsystems for producing a correlationThe associated failure rate of (c);

according to a formula (1), respectively establishing correlation coefficient calculation models of all fault chains aiming at the characteristics of the fault chains, wherein the correlation coefficient calculation models of all fault chains form a correlation coefficient model system of the relevant fault;

aiming at the first and the third types of the fault chains, the correlation relationship of the fault chains is one-way correlation, and the correlation fault terminal subsystemSubject to only one subsystemIs affected by the fault, then the subsystem in question is faultyThe overall failure rate of (2) is obtained from equation (1):

(2)

then there are:

(3)

due to the relevant subsystemsIs the starting point of the related fault and is not related to other subsystems, so(ii) a Wherein:: as subsystems ofThe independent failure rate of;

: for the subsystem in the case of a relevant faultThe overall failure rate of;

: as subsystems ofReceptor systemCorrelation coefficient of action;

aiming at the fourth type of the fault chain, the correlation of the fault chain is multi-system one-way correlation, and the correlation fault end pointAre simultaneously affected by the correlation of a plurality of subsystems, andfor the end of a related fault, related fault endThe running state does not affect other related subsystems; sub-systemIntegrated failure rate ofByDetermination of correlation as a calculatorSystem for controlling a power supplyReceptor systemCorrelation coefficient of actionValues, make the following assumptions:

(1) sub-systemAndsub-system dependent correlation coefficientThere is no linear correlation between them,(ii) a Order toThen, the following equations (2) and (3) hold, and the following equation (4) is derived from the equation (2):

（4）

as subsystems ofIs subjected toSub-failure rates of subsystem-related actions; by subsystemFault data removal ofModeling the relevant fault data of other subsystems except the subsystems;

(2) sub-systemTo receiveSub-failure rate of related actions of subsystemsAnd subsystemComprehensive failure rateIn a functional relationship:

（5）

sub-systemAll are related fault starting points and are used for the subsystemIs equal to the subsystemThe independent failure rate of; in formula (4),,WherebyAccording to the formulas (1) and (4), the formula (5) is derived as the formula (6):

（6）

: as subsystems ofReceptor systemCorrelation coefficient of action;

aiming at the second type and the fifth type of the fault chain, the correlation relationship of the fault chain is multi-system correlation, and when the comprehensive fault rate of each subsystem is calculated, the correlation fault rates and correlation coefficients of the subsystem and other related subsystems are required to be respectively solved; the names of the system are respectively: starting point of related faultEnd of related failureIntermediate point of correlation failureThen the related fault end pointThe comprehensive failure rate model of (2) is as follows:

（7）

obtaining the subsystem from equation (4)The failure rate model of (1) is as follows:

（8）

: is a receptor systemInfluencing subsystemThe associated failure rate of (c);

: is a receptor systemInfluencing subsystemThe associated failure rate of (c);

: to a subsystemSubsystems for producing a correlationThe associated failure rate of (c);

: to a subsystemSubsystems for producing a correlationThe associated failure rate of (c);

as subsystems ofReceptor systemCorrelation coefficient of action;

as subsystems ofReceptor systemCorrelation coefficient of action;

、calculating the failure rate of the sub-system according to the formula (4); in formula (8)Is the starting point of the related fault, thereforeThen, thenCan be solved;

the pair subsystemSubsystems for producing a correlationOf related failure rateThe value analysis process is as follows, and the subsystem is usedFor the object of study, the subsystemAs fault intermediate point, subsystemWhile being influenced by the subsystemThe influence of (a):

（9）

: as subsystems ofThe overall failure rate of;

: as subsystems ofThe independent failure rate of;

: as subsystems ofReceptor systemCorrelation coefficient of action;

: to a subsystemSubsystems for producing a correlationThe associated failure rate of (c);

due to the fact thatIs a starting point of the related fault, so，ByThe production failure data of (a) is obtained,as subsystems ofIs known to be capable of independent failure rates,as determined by equation (9); according toAndwhether the information is related or not is judged,as subsystems ofReceptor systemThe correlation coefficient of the effect is such that,the value is taken in two cases,

the first method comprises the following steps:，andis not related to

And the second method comprises the following steps:，andcorrelation (10)

The value is determined,can be solved according to the formula (8);

analyzing the maintenance strategy considering the related faults, specifically revising the reliability of the subsystems by utilizing an established correlation coefficient model system among the subsystems, and calculating and predicting maintenance nodes;

the maintenance strategy of the system is formulated and optimized according to the reliability, and when the reliability of the system or the subsystem is less than a scheduled threshold value, preventive maintenance is required to be carried out on the system or the subsystem; if subsystemThe reliability is as follows if there is no related fault:

（11）

: as subsystems ofOnly the reliability model under the condition of independent failure rate;

consider a subsystem with an associated faultThe reliability model of (2) is:

（12）

comprises the following steps:

（13）

: to the sub-systemEstablishing a reliability model of the comprehensive failure rate;

the reliability calculated by the independent failure rate has a large possibility when the subsystemWhen the reliability of (2) reaches a prescribed threshold value, the subsystemWhen the time point calculated by the reliability model of (1) is used as the maintenance node, since the time node calculated by the independent failure rate is larger than the time node calculated by the related failure rate, the subsystem calculated by the equation (11)The maintenance time node of (a) is taken as the basis of the maintenance plan, and the backward delay of the preventive maintenance plan is inevitably caused.