CN116186459A

CN116186459A - Equipment health degree assessment method, system, medium, device and hierarchical assessment method based on combination weight and generalized power mean

Info

Publication number: CN116186459A
Application number: CN202211678034.7A
Authority: CN
Inventors: 杨立俭; 刘凯; 请求不公布姓名; 訾飞跃; 杨启硕; 林启慧; 叶文柱; 潘凯; 李曙光; 谢元浩; 王亚中
Original assignee: Chinese People's Liberation Army 95616 Unit Support Department; Chengdu Shuzhilian Technology Co Ltd
Current assignee: Chinese People's Liberation Army 95616 Unit Support Department; Chengdu Shuzhilian Technology Co Ltd
Priority date: 2022-12-26
Filing date: 2022-12-26
Publication date: 2023-05-30

Abstract

The invention belongs to the technical field of intelligent maintenance and big data analysis, and provides a device health degree evaluation method, a system, a medium, a device and a hierarchical evaluation method based on combination weight and generalized power mean value, wherein the method comprises the following steps: step S1: calculating the combined weight of each lower-level health degree evaluation index of the upper-level health degree evaluation indexes; step S2: calculating the health degree of the upper health degree evaluation index based on the generalized power mean value according to the combination weight and the health degree of each lower health degree evaluation index; step S3: optimizing parameters p and theta, and determining the optimal value of the parameters p and theta; step S4: and (3) recalculating the health degree of each level of health degree evaluation index according to the optimal values of the parameters p and theta. The overall health of the equipment obtained based on the method well reflects the importance difference among different systems and components in the equipment and the influence of the serial-parallel coupling degree of the equipment on the overall health in a structure to a certain extent.

Description

Equipment health degree assessment method, system, medium, device and hierarchical assessment method based on combination weight and generalized power mean

Technical Field

The invention belongs to the technical field of intelligent maintenance and big data analysis, and particularly relates to a device health degree assessment method, a system, a medium, a device and a hierarchical assessment method based on combination weight and generalized power mean value.

Background

The health evaluation of the equipment is an essential basis for carrying out full life cycle health management on the equipment, and aims to quantitatively measure the availability, reliability and the like of the equipment so as to provide relevant decision support for the use, maintenance, scrapping and the like of the equipment.

The structural composition of the device is very complex, one device usually comprises a plurality of systems, each system comprises a plurality of subsystems, the components in the whole device are numerous and complex, and different components are mutually influenced, if the device is directly regarded as a whole, the health degree is given by using methods such as fault prediction and the like, which part has hidden danger cannot be well explained, however, the existing health degree assessment method usually treats the device as a whole for health assessment, and the disadvantage of the assessment method is that when the whole health degree is too low, the specific components influencing the whole health degree cannot be traced after the root, and the specific components are difficult to find.

The part of research adopts a generalized power mean method and a combined weighting method for evaluation, but the selection of parameters depends on experience and expert opinion, and the objectivity is lacking; there are also some evaluation methods based on serial-parallel models, however, in practical application, the relationship between the components is not completely serial or parallel, and there are some complex structures which do not completely belong to the serial-parallel relationship.

Disclosure of Invention

The invention aims to provide a device health degree evaluation method based on combination weight and generalized power mean value, which reflects the importance difference among different systems and components in the device and the influence of the serial-parallel coupling degree of the different systems and components on the structure on the whole health degree, so as to solve the technical problems that the system structure considered in the prior art is too simple, faults cannot be traced, objectivity is lacked and the like.

In order to achieve the above purpose, the technical scheme adopted by the invention is as follows:

a device health evaluation method based on combination weight and generalized power mean comprises the following steps:

step S1: calculating the combination weight of each lower-level health evaluation index of the upper-level health evaluation indexes:

w _i ＝θα _i +(1-θ)β _i (1)

Wherein θ ε [0,1 ]]A, which represents the combination coefficient of subjective and objective weights _i Representing subjectivityWeight, beta _i The method comprises the steps of expressing objective weight, wherein the upper-level health degree evaluation index comprises a plurality of lower-level health degree evaluation indexes;

step S2: calculating the health degree of the upper health degree evaluation index based on the generalized power mean value according to the combination weight and the health degree of each lower health degree evaluation index;

step S3: optimizing parameters in the step S1 and the step S2, and determining the optimal value of the parameters;

step S4: and (5) recalculating the health degree of each level of health degree evaluation index according to the optimal value of the parameter.

Further, the calculation formula in the step S2 is as follows:

wherein x is _i (i=1, 2, …, n) is the health degree of the lower level health degree evaluation index, w _i (i=1, 2, …, n) is the combining weight and p is the power exponent of the generalized power mean.

Further, the parameters optimized in the step S3 are p and θ, and the specific method is as follows:

s3.1: based on the historical data of the equipment, obtaining health degree change curves of the equipment at different stages in the life cycle according to the step S1 and the step S2;

s3.2: constructing a loss function in the parameter optimization process based on the experience curve of the equipment health degree change and the health degree change curve in the step S3.1, wherein the loss function is expressed by the formula 3:

Loss＝W(F ₀ ||F) (3)

Wherein F represents a health degree change curve, F ₀ An empirical curve representing the change in the health of the device, W being a function of the degree of difference between the two curves;

s3.3: and taking the loss function as an optimization target, and solving an optimal solution of the parameters p and theta by using an optimization algorithm.

Further, the specific method of step S3.1 is as follows: and randomly extracting historical state parameter data of the same number of equipment with the same type in different life cycle stages, calculating the health degree of each equipment according to the step S1 and the step S2, and solving the average value of the health degree of the equipment under the same life cycle, thereby obtaining the health degree change curve of the equipment with the same type on different life cycles.

Further, in the step S3.2, a deviation between the health degree change curve and the experience curve is measured by using the wasperstein distance.

Further, the subjective weight alpha _i Is calculated by the following steps: based on expert experience, subjective weight alpha of each lower-level health evaluation index is obtained by using analytic hierarchy process ₁ ,α ₂ ,α ₃ …α _n ；

The objective weight beta _i Is calculated by the following steps: based on entropy weight method, objective weight beta of each subordinate health degree evaluation index is calculated ₁ ,β ₂ ,β ₃ …β _n 。

In order to achieve the above object, the present invention further provides a hierarchical evaluation method, including the following steps:

(1) Constructing a hierarchical health evaluation index system according to the equipment structure composition;

(2) Collecting equipment state parameter data, and obtaining the health degree of a final-stage health degree evaluation index according to the equipment state parameter data;

(3) Based on the health degree of the final health degree evaluation index, the equipment health degree evaluation method based on the combination weight and the generalized power mean value as claimed in any one of claims 1 to 6 is applied to calculate the health degree of each level of health degree evaluation index in an upward step-by-step layer-by-layer iteration mode.

Further, the specific method of the step (1) is as follows:

(11) Taking the whole equipment as a first-level health evaluation index;

(12) Constructing secondary health evaluation indexes of all secondary system structures according to the secondary system structure composition of the equipment structure;

(13) Constructing three-level health degree evaluation indexes of all three-level system structures in each two-level system structure according to three-level system structures forming the two-level system structure, and analogizing in sequence until constructing a final-level health degree evaluation index aiming at a specific component of the equipment;

(14) And combining all the health degree evaluation indexes to form the hierarchical health degree evaluation index system.

In order to achieve the above object, the present invention further provides a system for implementing the method for evaluating the health degree of a device based on the combination weight and the generalized power mean, which comprises:

A combination weight calculation module: calculating the combination weight of each lower-level health evaluation index of the upper-level health evaluation indexes:

w _i ＝θα _i +(1-θ)β _i (1)

wherein θ ε [0,1 ]]A, which represents the combination coefficient of subjective and objective weights _i Representing subjective weight, beta _i The method comprises the steps of expressing objective weight, wherein the upper-level health degree evaluation index comprises a plurality of lower-level health degree evaluation indexes;

the health degree calculating module is used for: calculating the health degree of the upper health degree evaluation index according to the combination weight and the health degree of each lower health degree evaluation index based on the generalized power mean value and the formula (2);

wherein x is _i (i=1, 2, …, n) is the health degree of the lower level health degree evaluation index, w _i (i=1, 2, …, n) is a combining weight, p is a power exponent of the generalized power mean;

parameter optimization module: optimizing parameters p and theta, and determining the optimal value of the parameters p and theta;

and an optimization calculation module: and (3) recalculating the health degree of each level of health degree evaluation index based on the formula (1) and the formula (2) according to the optimal values of the parameters p and theta.

To achieve the above object, the present invention also provides a computer-readable storage medium having stored thereon a computer program to be executed by a processor to implement the combination weight and generalized power average based device health assessment method.

In order to achieve the above object, the present invention further provides a device health evaluation apparatus based on a combination weight and a generalized power mean, including: a processor and a memory;

the memory is used for storing a computer program;

the processor is connected with the memory and is used for executing the computer program stored by the memory, so that the device health evaluation device based on the combination weight and the generalized power mean executes the device health evaluation method based on the combination weight and the generalized power mean.

Compared with the prior art, the invention has the following beneficial effects:

(1) The invention combines the structural composition of the equipment to be evaluated, uses the generalized power average value to approach the min-max function, thereby being capable of measuring the series-parallel coupling relation inside the system more flexibly, and further calculating and obtaining the health evaluation result of the whole system by the mode; according to the invention, the influence of the complex system structure in the equipment on the equipment health degree is considered, and the constructed health degree index system is a hierarchical system, so that the key subsystem or key component affecting the overall health degree of the equipment can be positioned according to the overall health degree value of the equipment and the health degree value of each hierarchical index, thereby realizing the tracing of the root cause of the fault.

(2) According to the invention, the generalized power exponent p and the weight coefficient theta are optimized to obtain the optimal p value, so that the serial-parallel coupling degree of different systems (components) in the equipment is depicted, and the weight of each subsystem obtained based on the weight coefficient theta is depicted, so that the importance of each subsystem in the hierarchy is depicted; the overall health of the equipment obtained based on the method well reflects the importance difference among different systems and components in the equipment and the influence of the serial-parallel coupling degree of the equipment on the overall health in a structure to a certain extent.

(3) In the traditional index system adopting the power average value, the power exponent (P value) is often directly specified through expert opinion or practical experience, the subjectivity is strong, and the power exponent (P value) in the invention is calculated through an optimization algorithm by combining with statistical data, so that the method is more objective.

(4) When the results of the two different weighting methods are combined (namely, the combined weighting method), the weight (theta value) of each weighting method is usually directly designated by the traditional method, and the proportion of the weight results of the different weighting methods is calculated through an optimization algorithm, so that the method is more objective and reasonable.

Drawings

FIG. 1 is an equal-weighted power mean

A function curve for the power exponent p.

Fig. 2 shows a classical fault signature curve, the "bathtub curve" (the time-dependent fault rate).

Fig. 3 is a health-in-use time curve.

Fig. 4 is a schematic (partial) structural view of an aircraft of a certain model.

FIG. 5 is a schematic flow chart of the present invention-example 1.

FIG. 6 is a schematic flow chart of the present invention-example 2.

Fig. 7 is a schematic block diagram of embodiment 3 of the present invention.

Fig. 8 is a schematic block diagram of the present invention-example 5.

Detailed Description

The present invention will be further described in detail with reference to examples so as to enable those skilled in the art to more clearly understand and understand the present invention. It should be understood that the following specific embodiments are only for explaining the present invention, and it is convenient to understand that the technical solutions provided by the present invention are not limited to the technical solutions provided by the following embodiments, and the technical solutions provided by the embodiments should not limit the protection scope of the present invention.

It should be noted that the illustrations provided in the following embodiments merely illustrate the basic concept of the present invention by way of illustration, so that only the components related to the present invention are shown in the drawings and are not drawn according to the number, shape and size of the components in actual implementation, the form, number and proportion of each component in actual implementation may be arbitrarily changed, and the layout of the components may be more complicated.

Example 1

As shown in fig. 5, the present embodiment provides a method for evaluating the health degree of a device based on a combination weight and a generalized power average value, where the design principle of the method is to use the generalized power average value to approach a min-max function in combination with the structural composition of the device to be evaluated, so that the serial-parallel coupling relationship inside the system can be measured more flexibly, and the health degree evaluation result of the whole system can be obtained by calculation in this way.

The method for evaluating the health degree of the equipment provided by the embodiment specifically comprises the following steps:

1. step S1: calculating the combination weight of each lower-level health evaluation index of the upper-level health evaluation indexes:

w _i ＝θα _i +(1-θ)β _i (1)

wherein θ ε [0,1 ]]A, which represents the combination coefficient of subjective and objective weights _i Representing subjective weight, beta _i Representing objective weights; the upper level health degree evaluation index includes several lower level health degree evaluation indexes, for example: the aircraft fuel system includes three subsystems, namely an aircraft fuel system is an upper-level health evaluation index, a fuel indicating system, a fuel supply system and a fuel storage system are lower-level health evaluation indexes, and for example: the fuel oil indicating system is composed of a fuel oil measuring indicator and a sensor, namely the fuel oil indicating system is an upper-level health degree evaluation index, and the fuel oil measuring indicator and the sensor are lower-level health degree evaluation indexes. It should be noted that, the θ value needs to be optimized in the subsequent steps, the initial value may be 0.5, and the initial value is only used for the first calculation, and the θ values adopted by different levels or components are different, for example: in the calculation of the fuel measurement indicator and the sensor, the θ value is different, and for example: the θ values of the fuel indication system and the fuel supply system are different.

Subjective weight alpha _i Is calculated by the following steps: based on expert experience, each lower-level health degree evaluation finger is obtained by using analytic hierarchy processSubjective weighting of target alpha ₁ ,α ₂ ,α ₃ …α _n The method comprises the steps of carrying out a first treatment on the surface of the The objective weight beta _i Is calculated by the following steps: based on entropy weight method, objective weight beta of each subordinate health degree evaluation index is calculated ₁ ,β ₂ ,β ₃ …β _n The method comprises the steps of carrying out a first treatment on the surface of the The analytic hierarchy process adopted in the embodiment is not dependent on the scores of the index items, and is obtained by directly carrying out ranking scoring calculation on the importance of different index items in the hierarchical system through an expert, while the entropy weight method is a common objective weight solving method, the weight of the entropy weight solving method is obtained by calculating the health degree scores of all index items, and the two methods are all of the prior mature technologies, so that the description is omitted.

2. Step S2: calculating the health degree of the upper health degree evaluation index according to the combination weight and the health degree of each lower health degree evaluation index based on the generalized power mean value and the formula (2);

it should be noted that, the p value needs to be optimized in the subsequent steps, and p adopted by each level or component is different, for example: in the calculation of the fuel indication system and the sensor, the p value is different, for example: the p-value of the fuel indication system and the fuel supply system are different. In this embodiment, the initial value may be 1, and the initial value is only used for the first calculation; for example: when p=2 is the square average, p=1 is the arithmetic average, p=0 is the geometric average, and p= -1 is the harmonic average.

When x and w are completely determined, then S _p (x, w) is a unitary function with respect to p and is a continuous single increasing function, defining a domain as the entire real number axis. The limits at positive and negative infinity are max (x ₁ ,x ₂ ,...,x _n ) With min (x) ₁ ,x ₂ ,...,x _n )。For example: the average value of the equal weight powers of 10 natural numbers from 1 to 10 can be recorded as

The power mean is plotted against the power exponent p as shown in figure 1. Different p indices embody different calculated average ideas, for example: p takes 1 (i.e. arithmetic mean), the thought is "encourage balanced development", p takes 2 (i.e. square mean), then "encourage development advantage", which is characterized in that the mean is more susceptible to larger numbers, and when p takes 1 (i.e. harmonic mean), then "encourage replenishment short-plates", which is characterized in that the mean is more susceptible to smaller numbers.

In this embodiment, when the health degree of each level of index is evaluated, the power exponent calculation method is not only used to obtain the comprehensive score by weighting and summing each index item, but also indirectly reflects the influence of the serial-parallel connection degree of the system on the whole health degree. For example: for a level of subsystems, when there is a series relationship between each of the sub-items in the level (i.e., the lower level system or the secondary sub-system) (e.g., the aircraft fuel supply system and the engine system), or when the overall "degree of series" between each of the sub-items in the level is greater, it is indicated that any system failure will be highly likely to result in an overall failure of the level of subsystems. Thus, in connection with the power mean versus power exponent functional relationship of FIG. 1, the smaller the selected p value, the more "trending" the health of the hierarchical subsystem will be to select the lowest health secondary subsystem in the hierarchical subsystem. Conversely, when the subsystems are in parallel relationship (e.g., the main power system and the standby power system of the aircraft), or the overall "parallel degree" between the sub-items (i.e., the lower level system or the secondary subsystem) in the hierarchy is greater, it indicates that even if one of the sub-items fails, the hierarchy subsystem is not likely to fail, so that a larger p should be selected, i.e., if the p value is larger, the health value of the hierarchy subsystem will be "prone" to the secondary subsystem with the highest health in the hierarchy.

3. Step S3: optimizing the parameters p and theta, and determining the optimal value of the parameters p and theta.

The specific method comprises the following steps: s3.1: based on the historical data of the equipment, obtaining health degree change curves of the equipment in different stages of life cycle according to the step S1 and the step S2, wherein the specific method is as follows: randomly extracting historical state parameter data of the same number of equipment with the same type and different life cycle stages, calculating the health degree of each equipment according to the step S1 and the step S2, and calculating the average value of the health degree of the equipment under the same life cycle, thereby obtaining the health degree change curve of the equipment with the same type on different life cycles;

Loss＝W(F ₀ ||F) (3)

wherein F represents a health degree change curve, F ₀ An experience curve representing the change of the health degree of the equipment, wherein W is a function for measuring the difference degree between the two curves, and preferably, the Wasserstein distance is selected to measure the deviation between the health degree change curve and the experience curve; according to the fault rule of the equipment, the characteristics of the structure of the equipment and the like, experience, theory and the like, an experience curve of the change of the health degree of the equipment can be obtained; the health degree evaluation has a certain subjectivity, and if the equipment health degree change curve calculated by the health degree evaluation method provided by the embodiment is very close to the experience curve of the equipment health degree change, the health degree evaluation method provided by the invention is relatively reliable, the equipment health degree change condition can be better reflected, and the method has a certain guiding significance; therefore, the degree of difference between the health degree change curve and the experience curve (namely, the Loss function-formula (3)) calculated by measuring the invention is used as an optimization target, and the optimal values of the parameters p and theta can be obtained by optimizing the Loss solution.

S3.3: taking the loss function as an optimization target, and utilizing an optimization algorithm to obtain an optimal solution of parameters p and theta; the optimization principle of the step is as follows: the distribution of the calculated health degree values is continuously approximated to the theoretical health degree distribution of the equipment, so that the optimal values of the parameters p and theta are obtained, namely, the constructed loss function is taken as an optimization target, the optimal solution of the parameters p and theta is obtained by utilizing an optimization algorithm, preferably, a classical particle swarm algorithm is selected as the optimization algorithm, the particle swarm algorithm is a common optimization algorithm, and the specific calculation process is not the key point of the invention and is not repeated. When the parameter data of the device is updated, the values of the parameter p and the parameter θ are different when the health degree is calculated.

4. Step S4: and (3) recalculating the health degree of each level of health degree evaluation index based on the formula (1) and the formula (2) according to the optimal values of the parameters p and theta. It should be noted that, the calculation logic provided in this embodiment is iterative calculation step by step, after determining the optimal values of the parameters p and θ, the health degree of the previous stage should be calculated according to the health degree of the last stage health degree evaluation index (obtained according to the equipment history data) and the optimal values of p and θ, then the health degree of the previous stage is calculated according to the health degree of the previous stage and the optimal values of p and θ, and so on.

Example 2

As shown in fig. 1 to 6, the present embodiment provides a hierarchical evaluation method, which includes the following steps: (1) Constructing a hierarchical health evaluation index system according to the equipment structure composition; (2) Collecting equipment state parameter data, and obtaining the health degree of a final-stage health degree evaluation index according to the equipment state parameter data; (3) Based on the health degree of the final health degree evaluation index, the equipment health degree evaluation method based on the combination weight and the generalized power mean value described in the embodiment 1 is applied to calculate the health degree of each level of health degree evaluation index in an upward step-by-step layer iteration mode.

The steps are refined, and the following steps are obtained:

step S1: constructing a hierarchical health evaluation index system according to the equipment structure composition; step S2: collecting equipment state parameter data, and obtaining the health degree of a final-stage health degree evaluation index according to the equipment state parameter data; step S3: based on the health degree of the final-stage health degree evaluation index, iteratively calculating the health degree of each stage of health degree evaluation index step by step upwards; step S4: optimizing parameters p and theta, and determining the optimal value of the parameters p and theta; step S5: and (3) recalculating the health degree of each level of health degree evaluation index based on the formula (1) and the formula (2) according to the optimal values of the parameters p and theta.

Each step is described in detail below:

1. step S1: according to the equipment structure composition, a hierarchical health degree evaluation index system comprising a first-level health degree evaluation index and a second-level health degree evaluation index … … N-level health degree evaluation index is constructed, wherein the N-level health degree evaluation index is a last-level health degree evaluation index aiming at specific parts of equipment.

The specific method for constructing the hierarchical health evaluation index system comprises the following steps: (1) taking the whole equipment as a first-level health evaluation index; (2) Constructing secondary health evaluation indexes of all secondary system structures according to the secondary system structure composition of the equipment structure; (3) Constructing three-level health degree evaluation indexes of all three-level system structures in each two-level system structure according to three-level system structures forming the two-level system structure, and analogizing in sequence until constructing a final-level health degree evaluation index aiming at a specific component of the equipment; (4) And combining all the health degree evaluation indexes to form a hierarchical health degree evaluation index system.

The hierarchical health degree evaluation index system constructed in the mode is characterized in that the secondary health degree evaluation index is a lower level index of the primary health degree evaluation index, the tertiary health degree evaluation index is a lower level index of the secondary health degree evaluation index, and the final health degree evaluation index is the lowest level index, which also represents a specific part of the equipment. It should be noted that, the above specific method is a logic expression for constructing the hierarchical health degree evaluation index system, which is not limited to the construction level of the device, and the hierarchical health degree evaluation index system of different devices may only be constructed to two levels, i.e. the hierarchical health degree evaluation index system of a certain device includes a first level health degree evaluation index and a second level health degree evaluation index, where the second level health degree evaluation index is the last level health degree evaluation index, and for example, the hierarchical health degree evaluation index system of another device may only be constructed to three levels, i.e. the hierarchical health degree evaluation index system includes a first level health degree evaluation index, a second level health degree evaluation index and a third level health degree evaluation index, where the third level health degree evaluation index is the last level health degree evaluation index.

The following takes the fuel system of a certain model aircraft as the equipment main body to describe the construction logic of a hierarchical health evaluation index system, and the health evaluation of the aircraft fuel system has two outstanding characteristics: (1) The system is in an obvious hierarchical structure, and the component support subsystem operates and supports the whole fuel system to operate; (2) Different subsystems and components in the system have stronger series-parallel coupling relation and mutually influence. For example: the fuel system of the aircraft of a certain model consists of a fuel indication system, a fuel supply system and a fuel storage system; wherein, the fuel oil indicating system (abbreviated as "A system") is composed of a fuel oil measuring indicator and a sensor, and is named as "A1 and A2"; the fuel supply system (abbreviated as "B system") is composed of a low-pressure cut-off valve, a suction fuel supply valve, an alternate delivery fuel supply valve and a fuel supply pump, and is denoted as "B1, B2, B3 and B4"; the fuel oil storage system (abbreviated as "C system") consists of a fuel oil tank, an auxiliary fuel tank and an anti-overflow fuel tank, and is named as "C1, C2 and C3"; three levels total from the last health score indicator to the fuel system level; the fuel system is a first-level fuel system, the health degree evaluation index of the fuel system corresponds to the first-level health degree evaluation index, the fuel system is composed of a fuel indication system, a fuel supply system and a fuel storage system, namely, the fuel indication system, the fuel supply system and the fuel storage system are two-level system structures of the fuel system, the health degree evaluation index of the fuel indication system, the health degree evaluation index of the fuel supply system and the health degree evaluation index of the fuel storage system are two-level health degree evaluation indexes, the fuel indication system, the fuel supply system and the fuel storage system are respectively composed of specific parts, and the health degree evaluation index of each part is a three-level health degree evaluation index and also is a final-level health degree evaluation index, and the fuel system comprises: the fuel oil measuring indicator health degree evaluation index and the sensor health degree evaluation index corresponding to the fuel oil indicating system, the low-pressure cut-off valve health degree evaluation index and the suction fuel supply valve health degree evaluation index corresponding to the fuel oil supplying system, the cross fuel supply valve health degree evaluation index and the fuel supply pump health degree evaluation index corresponding to the fuel oil storing system, and the fuel oil tank health degree evaluation index, the auxiliary fuel tank health degree evaluation index and the anti-overflow fuel tank health degree evaluation index corresponding to the fuel oil storing system.

2. Step S2: and collecting equipment state parameter data, and obtaining the health degree of the final-stage health degree evaluation index according to the equipment state parameter data.

The health degree score of the hierarchical health degree evaluation index system constructed in the step S1 depends on the health degree score of the final health degree evaluation index, namely the health degree of the specific parts of the equipment, and the score of the final grade index, namely the health degree score of the specific parts, can be calculated by utilizing the acquired equipment state parameter data.

For example: x is an important parameter of the component A1 during operation, and the value of the parameter is x during normal operation of the component A1 ₀ The health of the component A1 can be determined by

Given that x deviates from x ₀ The more, the lower the health. In this embodiment, the user may flexibly set the score interval of the health degree (e.g., 0-100 minutes or 1-5 minutes, etc.). It should be noted that, for the calculation of the health degree of the specific parts of the device, the user may perform specific processing in combination with a specific service method (rule), and the specific calculation method is not limited in this embodiment.

The description will be given by way of example with the fuel system of a certain type of aircraft as the main body of the apparatus: according to the fault rate prediction model of the fuel system, aircraft historical flight data are collected, and numerical values of final-stage health evaluation indexes of the fuel system, such as health scores of nine parts including a sensor health evaluation index, an indicator health evaluation index and a low-pressure cut-off valve health evaluation index, are calculated and obtained, wherein the numerical values correspond to numerical values of A1, A2, B1, B2, B3, B4, C1, C2 and C3 in the table 1 respectively. 10 pieces of data which are uniformly distributed in different life stages of the aircraft are randomly sampled from historical flight data of the aircraft, so that the extracted flight data cover the whole life cycle of the aircraft, and the accuracy of a change curve of the aircraft health along with the use time based on statistical calculation in the subsequent steps is ensured. To facilitate scoring, the calculated final indicator health values are scaled to the interval [0,100], as shown in Table 1:

TABLE 1 random decimated final health index values for 10 different life stage aircraft

3. Step S3: and according to the health degree of the final-stage health degree evaluation index, iteratively calculating the health degree of each stage of health degree evaluation index layer by layer to obtain the overall health degree of the equipment.

In the step, the specific method for calculating the health degree evaluation index of each level is as follows:

step S3.1: the weight of each lower index in the grade health evaluation index is calculated, and the significance of the step is that the contribution degree of each subsystem to the whole is different, and different weights are required to be configured. For example: the aircraft power supply system can be divided into a main power supply system and an emergency power supply system, wherein the emergency power supply system can only supply power to the most important part of electric equipment on the aircraft, and some secondary equipment is tended to be abandoned, so that the two devices are placed in the same place, and different weights are required to be allocated to the power supply system when the health degree of the power supply system is calculated. The embodiment fully utilizes expert priori knowledge based on the combined weighting ideas of the subjective weighting method and the objective weighting method on the basis of ensuring certain objectivity, and the specific weighting calculation method comprises the following steps:

step 3.1.1: based on expert experience, subjective weight alpha of each lower index item is obtained by using analytic hierarchy process ₁ ,α ₂ ,α ₃ …α _n ；

Step 3.1.2: based on entropy weight method, objective weight beta of each lower index item is calculated ₁ ,β ₂ ,β ₃ …β _n ；

Step 3.1.3: based on the subjective weight and the objective weight, a combination weight of the subjective weight and the objective weight is calculated according to formula (1)w _i As final weights:

w _i ＝θα ₁ +(1-θ)β _i (1)

wherein, θ ε [0,1] represents the combination coefficient of subjective and objective weights.

It should be noted that, the θ value needs to be optimized in the subsequent steps, the initial value may be 0.5, and the initial value is only used for the first calculation, and the θ values adopted by different levels or components are different, for example: in the calculation of the second-level index item and the third-level index item, the θ value is different, for example: the theta value of one part in the three-level index item is different from the theta value of the other part.

The analytic hierarchy process adopted in the embodiment is not dependent on the scores of the index items, and is obtained by directly carrying out ranking scoring calculation on the importance of different index items in the hierarchical system through an expert, while the entropy weight method is a common objective weight solving method, the weight of the entropy weight solving method is obtained by calculating the health degree scores of all index items, and the two methods are all of the prior mature technologies, so that the description is omitted.

Step S3.2: calculating the health degree of each lower level index according to the weight in the step S3.1 and the generalized power mean value and the formula (2);

Wherein x is _i (i=1, 2, …, n) is the health degree of the lower level index, w _i (i=1, 2, …, n) is the weight calculated in the step S3.1, and p is the power exponent of the generalized power mean.

It should be noted that, the p value needs to be optimized in the subsequent steps, and p adopted by each level or component is different, for example: in the calculation of the second-level index item and the third-level index item, the p value is different, for example: the p value of one part in the three-level index item is different from that of the other part. In this embodiment, the initial value may be 1, and the initial value is only used for the first calculation; for example: when p=2 is the square average, p=1 is the arithmetic average, p=0 is the geometric average, and p= -1 is the harmonic average.

When x and w are completely determined, then S _p (x, w) is a unitary function with respect to p and is a continuous single increasing function, defining a domain as the entire real number axis. The limits at positive and negative infinity are max (x ₁ ,x ₂ ,...,x _n ) With min (x) ₁ ,x ₂ ,...,x _n ). For example: the average value of the equal weight powers of 10 natural numbers from 1 to 10 can be recorded as

Based on the calculation logic of the steps, the calculation logic of the overall health degree of the equipment is as follows: starting from the final-stage health index value (i.e. the health of the final-stage specific component), based on the initial p and θ values (p=1, θ=0.5), the health of the last-stage subsystem of the final-stage specific component is calculated based on the formula (2), then the health of the last-stage subsystem of the final-stage specific component is calculated according to the health of the last-stage subsystem of the final-stage specific component, and so on until the overall health of the device is obtained. Taking a three-level hierarchical health degree evaluation index system as an example, the three-level hierarchical health degree evaluation index system comprises a first-level health degree evaluation index, a second-level health degree evaluation index and a third-level health degree evaluation index (the level is a final-level health degree index value), firstly, calculating the second-level health degree evaluation index according to the final-level health degree index value, and then, calculating the first-level health degree evaluation index according to the second-level health degree evaluation index, thereby obtaining the overall health degree (namely, the first-level health degree evaluation index) of the equipment. It should be noted that, because of the differences among the subsystems of different levels, the subjective weight α, objective weight β, and the optimization parameters p and θ of each index item adopted in calculating the subsystems of different levels are all different; the subjective weight alpha and the objective weight beta are obtained through expert experience and data calculation respectively, and the numerical values of the parameters p and theta are optimized, and are obtained through continuous optimization according to historical data.

Taking a fuel system of a certain model of aircraft as an equipment main body for example, taking the calculation of the health degree of an A subsystem (namely a fuel indicating system) as an example: the health score of the last level indicator "sensor (i.e., A1)" and the last level indicator "measurement indicator (i.e., A2)" in the 10 recordings is obtained according to step S2:

first, the weights of the measurement indicators (i.e., A2) and the sensors (i.e., A1) in the a system (i.e., the "fuel indication system") are calculated: the subjective weights are scored and obtained by an expert, and because the measurement indicator (namely A2) and the sensor (namely A1) can lead to the situation of misjudging the oil quantity by a driver, no obvious importance difference exists, the subjective weights can be regarded as being equally important, the subjective weights are 0.5 and 0.5, and meanwhile, the objective weights of the subjective weights and the subjective weights are calculated by an entropy weight method based on the data of the table 1, and the results are 0.3424021,0.6575979 respectively;

then, the initial values of the undetermined parameters p and θ in the hierarchical health degree expression are set to 1 and 0.5, respectively, and the health degree of the a system is calculated. For example: aircraft number 1A System health degree is

Similarly, the health of the A system of the rest 9 airplanes is calculated, and the results are shown in Table 2:

TABLE 2 health of A systems

Similarly, the initial values of the undetermined parameters p and theta in the level health degree expression are respectively set to be 1 and 0.5, the subjective weight and the objective weight of the system B and the subjective weight of the system C are sequentially calculated according to the method for calculating the health degree of the system A, the initial values of the undetermined parameters p and theta are kept unchanged, the initial values are still 1 and 0.5, and the health degree results of the system A, the system B and the system C of 10 aircraft are calculated, as shown in table 3:

TABLE 3 health of subsystems

Finally, according to the health degrees of the A system, the B system and the C system, calculating to obtain the overall health degree of the equipment of the 10 aircraft fuel systems, wherein subjective weights are obtained through scoring by an expert, the indicating system, the supplying system and the storing system are considered to be equally important, the weights are 0.33,0.33,0.33, objective weights are calculated according to the data in the table 3 through an entropy weight method, the result is [0.31454068,0.26674935,0.41870998], the undetermined parameter p and the initial value of theta in the level health degree expression are set to be 1 and 0.5, and the overall health degree result of the 10 aircraft fuel systems is calculated, as shown in the table 4:

TABLE 4 sample overall health of 10 aircraft fuel systems (parameters p and θ initial values)

4. Step S4: optimizing the parameters p and theta, and determining the optimal value of the parameters p and theta.

Step S4.1, based on the historical data of the equipment, obtaining health degree change curves of the equipment at different stages in the life cycle according to the step S2 and the step S3; the specific method comprises the following steps: randomly extracting historical state parameter data of the same number of equipment with the same type and different life cycle stages, calculating the overall health degree of each equipment according to the step S2 and the step S3, and calculating the average value of the overall health degree of the equipment under the same life cycle, so as to obtain the health degree change curve of the equipment with the same type on different life cycles;

step S4.2, constructing a loss function in the parameter optimization process based on an experience curve and a health degree change curve of the equipment health degree change, wherein the loss function is expressed by a formula 3:

Loss＝W(F ₀ ||F) (3)

wherein F represents a health degree change curve, F ₀ The empirical curve representing the change of the health degree of the equipment, W is a function for measuring the difference degree between the two curves, preferably, the embodiment selects the deviation between the health degree change curve calculated by the method for calculating the health degree by Wasserstein distance (also called EM distance);

according to the fault rule of the equipment, the characteristics of the structure of the equipment and the like, experience, theory and the like, an experience curve of the change of the health degree of the equipment can be obtained; the health degree evaluation has a certain subjectivity, and if the equipment health degree change curve calculated by the health degree evaluation method provided by the embodiment is very close to the experience curve of the equipment health degree change, the health degree evaluation method provided by the invention is relatively reliable, the equipment health degree change condition can be better reflected, and the method has a certain guiding significance; therefore, the degree of difference between the health degree change curve and the experience curve (namely, the Loss function-formula (3)) calculated by measuring the invention is used as an optimization target, and the optimal values of the parameters p and theta can be obtained by optimizing the Loss solution.

In particular, the choice of theoretical distribution varies according to the specific situation; in addition, because the different design principles, production and manufacturing processes, quality inspection and maintenance measures and the fault rate change characteristics of different types of equipment are different, the used target distribution is not the same, and the corresponding theoretical distribution of the equipment is determined by combining the conditions of the type of equipment. In the field of equipment reliability, a typical failure curve "bathtub curve" is chosen as an example for illustration in this embodiment, for example: assuming that the failure characteristic curve of a certain device conforms to a classical "bathtub curve", as shown in fig. 2, the failure rate of most devices is a function of time, in the field of device reliability, failure factors of the devices are generally considered, including defects in design and manufacture of the devices, failures caused by improper assembly and installation, random occasional failures, and wear failures caused by long-term operation, which can be generally classified into three stages (the failure characteristic curves are shown in the bathtub curve in fig. 2) of early handling, installation, configuration, improper operation, occasional failure period (or "occasional failure period") caused by improper operation, and wear failure period (or "wear failure period") caused by wear failure, and the proportion and performance of these three stages will be different in different devices.

Based on historical experience and historical fault data, it can be assumed that the device health satisfies:

where t represents the used time period, n represents the design service life, and n=2400 is set, then according to formula (5):

a health degree curve (as shown in fig. 3) can be obtained, and obviously, in the process of changing the health degree of the equipment obtained according to the typical fault model 'bathtub curve', the health degree of the equipment is lower in the initial 'early failure period', and after the period lasts for a short time, the equipment enters into the 'sporadic failure period', and the health degree of the equipment is higher; as time increases, the device enters a "wear out period" where the health begins to decline, and when the period of use reaches a certain time, as the device's performance deteriorates dramatically, the health will decline dramatically until a serious failure occurs and becomes unusable.

It should be noted that, in practice, the entire failure rate curve of the device does not always show the complete bathtub curve in the above example during the specified service life; moreover, even with the same equipment, operating in different environments and service strengths, failure rate curves are different, for example: some devices are designed reliably, and are subjected to very strict factory detection before factory delivery, and are subjected to strict debugging (such as automobiles and the like) during use, so that the early failure period of the devices basically does not occur, but directly enters a relatively stable sporadic failure period; the failure rate of some devices increases sharply in the occasional failure period after the early failure period due to poor design and manufacture; and due to improper replacement maintenance of some equipment, the replacement overfrequency causes the mechanical equipment to break down due to repeated re-running, and the failure rate frequently fluctuates in early failure period, accidental failure period and the like. Therefore, because of the differences of the actual equipment, the use environments and the intensities of the equipment, the target health degree curves are different, the specific situation should be combined, the specific selection should be performed based on the actual failure rate statistical data, and the obtaining mode and the specific form are not limited in this embodiment.

Step S4.3: taking the loss function as an optimization target, and utilizing an optimization algorithm to obtain an optimal solution of parameters p and theta; the optimization principle of the step is as follows: the distribution of the calculated health degree values is continuously approximated to the theoretical health degree distribution of the equipment, so that the optimal values of the parameters p and theta are obtained, namely, the constructed loss function is taken as an optimization target, the optimal solution of the parameters p and theta is obtained by utilizing an optimization algorithm, preferably, a classical particle swarm algorithm is selected as the optimization algorithm, the particle swarm algorithm is a common optimization algorithm, and the specific calculation process is not the key point of the invention and is not repeated. When the parameter data of the device is updated, the values of the parameter p and the parameter θ are different when the health degree is calculated.

5. Step S5: and (3) recalculating the health degree of each level of health degree evaluation index based on the step (S3) according to the optimal values of the parameters (p) and (theta), thereby obtaining the final health degree of the whole equipment. Since the parameters p and θ are initially set to 1 and 0.5 and are not optimal values when the overall health of the device is calculated in step S3 (i.e., first calculation), when the optimal values of the parameters p and θ are determined by optimizing the loss function, the parameters p and θ should be substituted into the calculation in step S3 for recalculation, i.e., step S3 is repeated again based on the determined optimal values of the parameters p and θ, thereby recalculating the final overall health of the device. It should be noted that, the calculation logic provided in this embodiment is iterative calculation step by step, after determining the optimal values of the parameters p and θ, the health degree of the previous stage should be calculated according to the health degree of the last stage health degree evaluation index (obtained according to the historical data of the device) and the optimal values of p and θ, then the health degree of the previous stage is calculated according to the health degree of the previous stage and the optimal values of p and θ, and then the health degree of the previous stage is calculated, and so on, and finally the overall health degree of the device is calculated.

The description will be given by way of example with the fuel system of a certain type of aircraft as the main body of the apparatus:

firstly, uniformly distributing the fuel oil systems of the 10 aircraft which are randomly extracted in different stages of the whole life cycle of the aircraft, and calculating the average value of the aircraft in the same life stage in the whole life cycle based on the health degree calculated in the step S3 so as to obtain a health degree change curve F of the fuel oil system of the aircraft in the model at different life stages;

secondly, based on the model airplanePrinciple construction and structural characteristics of the fuel system, and 'bathtub curve' is taken as an experience curve F of the health degree of the fuel system of the aircraft of the model ₀ The empirical curve describes the health of the system over different life cycles for any one aircraft, F ₀ The empirical distribution for determination is not affected by other parameters. Based on the health degree change curve F (or health degree numerical distribution) obtained by statistical calculation, the health degree change curve F is only influenced by each undetermined parameter p and θ in the calculation process;

if the health degree value distribution F obtained based on statistical calculation and the health degree experience distribution F obtained based on equipment principle ₀ The closer the health degree is, the health degree evaluation calculation method of the structure can better reflect the health degree condition of the equipment, is consistent with the related principle and experience, and has certain guiding significance; thus, two distributions F and F can be calculated ₀ The degree of difference between the two parameters is minimized as an optimization target, so that the optimal value of the parameters p and theta is obtained by optimizing the degree of difference between the two distributions by using an optimization algorithm;

secondly, carrying out parameter optimization on Loss (p, theta) by using an optimization algorithm, wherein the optimization range of each theta is uniformly set to 0 to 1 in the optimization algorithm, and the optimization range of p can be selected by a system coupling relation in priori knowledge; for example: in the fuel system, the indication, supply and storage functions are not indispensable, so that a series model is formed, and the p value range can be set to be-2 to 0; in the storage system, the parallel connection (alternative) exists between the oil supply tank, the auxiliary tank and the anti-overflow tank, and the p value can be set to be 2 to 4. After the optimization iteration reaches the termination condition, each p and theta value is determined; wherein, the optimal values of p and theta are (0, -2, 4) and (1, 1) when the health degree of the three subsystems of A, B, C is calculated, and the optimal values of p and theta are 0 and 1 when the overall health degree of the equipment is calculated by the three subsystems of A, B, C.

Finally, substituting the optimal values of the parameters p and theta into an aircraft fuel system health degree calculation expression, repeating the step S3, iterating layer by layer to obtain a final health degree evaluation result of the model aircraft fuel system, and sampling to obtain the overall health degree of 10 aircraft, wherein the overall health degree is shown in Table 5:

TABLE 5 overall health of the Fuel System (optimal values for parameters p and θ)

The health index system constructed in this embodiment is a hierarchical system, so that a key subsystem or a key component affecting the overall health of the device can be located according to the overall health value of the device and the health values of the indexes of each hierarchy.

Example 3

As shown in fig. 7, the present embodiment provides a device health evaluation system based on a combination weight and a generalized power mean, which is configured to implement the device health evaluation method based on a combination weight and a generalized power mean provided in embodiment 1, and specifically includes:

w _i ＝θα _i +(1-θ)β _i (1)

/>

It should be noted that, the structure and/or principle of each module corresponds to the steps in the method for evaluating the health degree of the device based on the combination weight and the generalized power mean described in embodiment 1, so that the description is omitted here.

It should be noted that, it should be understood that the division of each module of the above system is merely a division of a logic function, and may be fully or partially integrated into a physical entity in actual implementation, or may be physically separated, and the modules may be fully implemented in a form of software called by a processing element; or can be realized in hardware; the method can also be realized in a form of calling software by a processing element, and the method can be realized in a form of hardware by a part of modules. For example, a module may be a processing element that is set up separately, may be implemented in a chip of an apparatus, may be stored in a memory of the apparatus in the form of program codes, may be called by a processing element of the apparatus and perform functions of a module, and may be implemented similarly. In addition, all or part of the modules can be integrated together or can be independently implemented. The processing element described herein may be an integrated circuit having signal processing capabilities. In implementation, each step of the above method or each module above may be implemented by an integrated logic circuit of hardware in a processor element or an instruction in a software form.

For example, the modules above may be one or more integrated circuits configured to implement the methods above, such as: one or more specific integrated circuits, or one or more microprocessors, or one or more field programmable gate arrays, etc. For another example, when a module above is implemented in the form of a processing element scheduler code, the processing element may be a general purpose processor, such as a central processing unit or other processor that may invoke the program code. For another example, the modules may be integrated together and implemented in a system-on-chip form.

Example 4

The present embodiment provides a computer-readable storage medium having stored thereon a computer program that is executed by a processor to implement the apparatus health assessment method based on the combination weights and generalized power means provided in embodiment 1. Those of ordinary skill in the art will appreciate that: all or part of the steps of implementing the method provided in embodiment 1 may be implemented by hardware associated with a computer program, where the computer program may be stored in a computer readable storage medium, and when executed, the program performs steps including the method provided in embodiment 1; and the storage medium includes: various media that can store program code, such as ROM, RAM, magnetic or optical disks.

Example 5

As shown in fig. 8, the present embodiment provides a device health evaluation apparatus based on a combination weight and a generalized power mean, including: a processor and a memory; the memory is used for storing a computer program; the processor is connected to the memory, and is configured to execute a computer program stored in the memory, so that the device health evaluation apparatus based on the combination weight and the generalized power average performs the device health evaluation method based on the combination weight and the generalized power average provided in embodiment 1.

Specifically, the memory includes: various media capable of storing program codes, such as ROM, RAM, magnetic disk, U-disk, memory card, or optical disk.

Preferably, the processor may be a general-purpose processor, including a central processor, a network processor, etc.; but also digital signal processors, application specific integrated circuits, field programmable gate arrays or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

The above embodiments are merely illustrative of the principles of the present invention and its effectiveness, and are not intended to limit the invention. Modifications and variations may be made to the above-described embodiments by those skilled in the art without departing from the spirit and scope of the invention. Accordingly, it is intended that all equivalent modifications and variations of the invention be covered by the claims, which are within the ordinary skill of the art, be within the spirit and scope of the present disclosure.

Claims

1. The equipment health degree evaluation method based on the combination weight and the generalized power mean is characterized by comprising the following steps of:

w _i ＝θα _i +1-θβ _i (1)

2. The method for evaluating the health of a device based on the combination weights and the generalized power means according to claim 1, wherein the calculation formula in the step S2 is:

3. The method for evaluating the health of a device based on the combination weight and the generalized power mean according to claim 2, wherein the parameters optimized in the step S3 are p and θ, and the specific method is as follows:

Loss＝W(F ₀ ||F)(3)

4. The method for evaluating the health of a device based on the combination weights and the generalized power mean according to claim 3, wherein the specific method of step S3.1 is as follows: and randomly extracting historical state parameter data of the same number of equipment with the same type in different life cycle stages, calculating the health degree of each equipment according to the step S1 and the step S2, and solving the average value of the health degree of the equipment under the same life cycle, thereby obtaining the health degree change curve of the equipment with the same type on different life cycles.

5. The method for evaluating the health of a device based on combination weights and generalized power means according to claim 4, wherein the deviation between the Wasserstein distance metric health variation curve and the empirical curve is selected in the step S3.2.

6. The method for estimating the health of a device based on a combination weight and a generalized power mean according to claim 5, wherein the subjective weight α _i Is calculated by the following steps: based on expert experience, subjective weight alpha of each lower-level health evaluation index is obtained by using analytic hierarchy process ₁ ,α ₂ ,α ₃ …α _n ；

7. A hierarchical evaluation method, comprising the steps of:

8. The hierarchical device health assessment method of claim 7, wherein the specific method of step (1) is as follows:

(11) Taking the whole equipment as a first-level health evaluation index;

9. A system for implementing the method for evaluating the health of a device based on combining weights and generalized power means according to any one of claims 1 to 6, comprising:

w _i ＝θα _i +1-θβ _i (1)

10. A computer-readable storage medium, on which a computer program is stored, characterized in that the computer program is executed by a processor to implement the combination weight and generalized power mean-based device health assessment method according to any one of claims 1 to 5.

11. A device health assessment apparatus based on a combination weight and a generalized power mean, comprising: a processor and a memory;

the memory is used for storing a computer program;

the processor is connected to the memory, and is configured to execute a computer program stored in the memory, so that the device health evaluation device based on the combination weight and the generalized power average performs the device health evaluation method based on the combination weight and the generalized power average according to any one of claims 1 to 6.