KR20220150648A - Predictive maintenance operating system based on failure prediction model - Google Patents

Abstract

The present invention relates to a predictive maintenance operating system based on a failure prediction model. According to one embodiment of the present invention, the predictive maintenance operating system based on a failure prediction model includes: a failure prediction module which predicts an occurrence probability of a failure mode and a period of one or more cost breakdown structure (CBS) models constructed in one or more plants and stores the predicted occurrence probability in a database; a failure probability calculation module which determines a failure probability and the type of failure mode to be generated for each CBS model based on the failure probability and the type of failure mode stored in the failure prediction module or the database and determines a failure probability and a failure mode for each CBS of the lowest level, middle level, upper level, and the highest level, respectively; a failure cost management module which manages loss costs caused by a failure of each CBS and input costs to solve failures based on the failure probability and the failure mode of each CBS of the lowest level, middle level, upper level, and the highest level determined from the failure probability calculation module, respectively; a predictive maintenance establishment module which calculates a predictive maintenance probability considering an amount of money for each CBS based on the input costs and the loss costs managed by the failure cost management module and then determines whether to execute based on the same; and a predictive maintenance management module which executes the failure prediction module for each period pre-set in the database and generates a predictive maintenance report based on whether to execute and the predictive maintenance probability calculated from the predictive maintenance establishment module. The present invention may improve reliability and accuracy of a predictive maintenance plan.

Description

Predictive maintenance operating system based on failure prediction model}

The present invention relates to a predictive maintenance operating system based on a failure prediction model. More specifically, a failure prediction model is built, a value for the final module failure probability according to the failure mode of each module is calculated, and predictive maintenance is planned and established based on the calculated value to create a failure prediction model for power plant operation. It is about a predictive maintenance operating system based on

In general, a plurality of power generation facilities are complexly concentrated in a power plant. These multiple power generation facilities should be able to be monitored in real time at all times, and should be able to be repaired within the fastest possible time when a failure or abnormal state of the facility occurs.

Conventional power plants perform regular preventive maintenance for each of a plurality of power generation facilities according to a predetermined schedule, and predict failure of power generation facilities by collecting and analyzing sensor data that can determine the state of power generation facilities, We are taking countermeasures accordingly. However, in a conventional power plant, it simply notifies the occurrence of an alarm using sensor data collected from power generation facilities, but cannot analyze the cause of the alarms that actually occurred and notify the action, so that the operators of the power plant can There was a difficulty in finding a facility with a failure among power generation facilities.

Recently, research on a power plant failure prediction system based on a learning model has been actively conducted. This failure prediction system learns a learning model using a plurality of sensor data collected from power generation facilities, thereby predicting and diagnosing failures of power generation facilities from sensor data generated in actual power generation facilities.

However, most of the sensor data collected in a typical power plant is data according to the normal operation of power generation facilities. Therefore, even if a power plant failure prediction system is built based on the learning model, learning data provided to the learning model, that is, sensor data, cannot reflect various operating environments. Therefore, in the conventional power plant failure prediction system, the learning efficiency of the learning model based on the sensor data collected from the power generation facility is lowered, so that the accuracy and reliability of the failure prediction of the power plant is lowered.

Republic of Korea Patent Publication No. 10-2020-0001910

The purpose of the present invention is to construct a failure prediction model, calculate a value for the final module failure probability according to the failure mode for each module, and plan and establish predictive maintenance based on the calculated value to operate a power plant. To provide a predictive maintenance operating system based on

In order to solve the above problem,

A predictive maintenance operating system based on a failure prediction model according to an embodiment of the present invention predicts the period and occurrence probability of failure modes of one or more CBS models built in one or more power plants, a failure prediction module for storing the probability of occurrence in a database;

Based on the failure mode type and failure probability stored in the failure prediction module or the database, the failure mode type and failure probability to occur for each cost division structure model are determined, and cost division of the lowest level, middle level, upper level, and top level is performed. a failure probability calculation module that determines failure modes and failure probabilities for each structure;

Based on the failure mode and failure probability for each cost division structure of the lowest level, middle level, upper level, and top level determined by the failure probability calculation module, the loss cost due to failure for each cost division structure and the input cost for failure resolution Failure cost management module to manage;

A predictive maintenance establishment module that calculates a predictive maintenance probability considering the amount of each cost division structure based on the loss cost and input cost managed by the failure cost management module, and determines whether to implement based on this calculation;

a predictive maintenance management module that executes the failure prediction module for each period set in the database and generates a predictive maintenance report based on the predictive maintenance probability and whether or not the predictive maintenance is performed, calculated by the predictive maintenance establishment module; can include

In addition, in the predictive maintenance operating system based on the failure prediction model according to an embodiment of the present invention, the failure prediction module sets the period for the probability of occurrence of the failure mode to 1 day, 7 days, 30 days, and 90 days. It can be calculated by setting at least one of days, 180 days, and 365 days later.

In addition, in the predictive maintenance operating system based on the failure prediction model according to an embodiment of the present invention, the failure mode and failure probability generated for each cost division structure in the failure probability calculation module are,

(p:발전소 전력 / i:비용분할구조 모듈,1,2,3,n / j:고장모드,1,2,3,n)일 수 있다.

(p: power plant power / i: cost division structure module, 1,2,3, n / j: failure mode, 1,2,3, n).

In addition, in the predictive maintenance operating system based on the failure prediction model according to an embodiment of the present invention, the failure mode type and probability in the cost division structure model of the lowest level in the failure probability calculation module are,

including,

The failure mode J with the maximum failure probability can be determined as the failure mode in the cost partitioning structure module i.

In addition, in the predictive maintenance operating system based on the failure prediction model according to an embodiment of the present invention, the failure probability calculation module determines a weight for each failure mode type, calculates the lowest level adjusted failure probability, and Based on the lowest-level adjusted failure probability, failure modes and failure probabilities for each middle-level cost split structure model can be calculated.

In addition, in the predictive maintenance operating system based on the failure prediction model according to an embodiment of the present invention, the weight determination is performed by assigning weights differently for each inspection, test, partial repair, and failure replacement, and inspection=test< Differential weighting can be given in the order of partial repair < replacement of failure.

)은 In addition, in the predictive maintenance operating system based on the failure prediction model according to an embodiment of the present invention, the lowest level adjusted failure probability (

)silver

(j=점검, 테스트 k=1 / j=부분수리 k=2 / j=고장교체 k=3)일 수 있다.

(j=check, test k=1 / j=partial repair k=2 / j=breakdown replacement k=3).

In addition, in the predictive maintenance operating system based on the failure prediction model according to an embodiment of the present invention, the failure mode type and probability for each intermediate level cost division structure model are,

이며,

is,

)이 최대인 하위 비용분할 모듈 I에서 고장 모드 JI가 중간 레벨 비용분할구조 모델에서의 고장모드로 결정될 수 있다.The failure mode type is a fixed failure probability (

) can be determined as the failure mode in the middle-level cost splitting structure model.

In addition, in the predictive maintenance operating system based on the failure prediction model according to an embodiment of the present invention, the method for calculating the type and probability of the failure mode for each of the upper level or the highest level cost division structure model is the weight calculation It can be finally calculated through the result of calculating the value, the adjusted failure probability value for each level, and the failure mode type and probability for each cost division structure model.

In addition, in the predictive maintenance operating system based on a failure prediction model according to an embodiment of the present invention, the predictive maintenance establishment module performs predictive maintenance when the loss cost is higher than the input cost, and the cost division structure failure mode And whether the predictive maintenance is performed may be stored in the database.

In addition, in the predictive maintenance operating system based on a failure prediction model according to an embodiment of the present invention, the predictive maintenance management module calculates a system level failure probability based on the value calculated by the failure probability calculation module, , a predictive maintenance plan may be established based on the system level failure probability, and the established plan may be stored in a database.

These solutions will become more apparent from the following detailed description of the invention based on the accompanying drawings.

Prior to this, the terms or words used in this specification and claims should not be interpreted in a conventional and dictionary sense, and the inventor should properly define the concept of the term in order to explain his or her invention in the best way. It should be interpreted as a meaning and concept consistent with the technical idea of the present invention based on the principle that it can be.

According to an embodiment of the present invention, a failure prediction model may be constructed based on a cost division structure (CBS) model, and failure mode types and failure probabilities from the lowest level module to the highest level module of the cost division structure model can be calculated respectively, and a predictive maintenance plan can be established based on the calculated value.

In addition, according to an embodiment of the present invention, the type and probability of failure modes occurring for each cost division structure model for each level can be calculated, and a predictive maintenance plan can be established based on the calculated reliability and accuracy of the predictive maintenance plan. can improve

1 is a block diagram showing components of a predictive maintenance operating system based on a failure prediction model according to an embodiment of the present invention.

Specific aspects and specific technical features of the present invention will become more apparent from the following specific details and embodiments in conjunction with the accompanying drawings. In adding reference numerals to components of each drawing in this specification, it should be noted that the same components have the same numerals as much as possible, even if they are displayed on different drawings. In addition, in describing an embodiment of the present invention, if it is determined that a detailed description of a related known configuration or function may obscure the gist of the present invention, the detailed description will be omitted.

In addition, in describing the components of the present invention, terms such as first, second, A, B, (a), and (b) may be used. These terms are only used to distinguish the component from other components, and the nature, sequence, or order of the corresponding component is not limited by the term. When an element is described as being “connected,” “coupled to,” or “connected” to another element, that element is directly connected or connectable to the other element, but there is another element between the elements. It will be understood that elements may be “connected”, “coupled” or “connected”.

Hereinafter, an embodiment of the present invention will be described in detail based on the accompanying drawings.

As shown in FIG. 1, the predictive maintenance operating system based on the failure prediction model according to an embodiment of the present invention generates the period and failure mode of one or more Cost Breakdown Structure (CBS) models built in one or more power plants. The failure prediction module 10 predicts the probability, the failure probability calculation module 20 determines the failure mode and failure probability of the cost division structure model for each level, and each level determined by the failure probability calculation module 20 Based on the failure cost management module 30 that manages the loss cost by failure mode and failure probability and the input cost, and the loss cost and input cost managed by the failure cost management module 30, the amount for each cost division structure is considered. The predictive maintenance establishment module 40 that calculates the predictive maintenance probability and determines whether to implement it based on this, the function of executing the failure prediction module 10, the predictive maintenance probability calculated by the predictive maintenance establishment module 40, and It may be configured to include a predictive maintenance management module 50 that generates a predictive maintenance report based on whether or not it has been implemented.

The failure prediction module 10 can predict the probability of occurrence for each period of the cost splitting structure model or for each failure mode of the cost splitting structure model. The probability of occurrence may be calculated by setting at least one period of 30 days, 90 days, 180 days, and 365 days. After the failure prediction module 10 calculates the probability of occurrence of the failure mode of the cost division structure model for each date, it may be stored in the database 60 .

Here, the cost break down structure in the cost break down structure model is an abbreviation of Cost Break down Structure, and is a structure of cost elements for setting criteria for the life cycle of photovoltaic power generation facilities and cost evaluation for each year. In particular, it can be replaced by the operation and maintenance cost of a solar power plant, and the acquisition cost and operation and maintenance cost can be composed of each detailed cost element.

Since this cost splitting structure model makes it easy to collect and manage failure information including failure history, failure type, and presence/absence of failures for each power plant, a cost splitting structure model is created and managed for each power plant. It can be done.

The failure probability calculation module 20 determines the type of failure mode and failure probability to occur for each cost division structure model based on the failure mode type and failure probability previously stored in the failure prediction module 10 or the database 60, and determines the lowest Probability values may be calculated to determine failure modes and failure probabilities for each cost division structure of the level, middle level, upper level, and top level, respectively.

Specifically, the failure probability calculation module 20 may first calculate the probability of one or more failure modes for each cost division structure in order to determine the failure mode and failure probability to occur for each cost division structure model, and the calculation formula is as follows. .

(p: power plant power / i: cost division structure module, 1,2,3, n / j: failure mode, 1,2,3, n)

Next, the failure mode types and failure probabilities can be calculated for each level of the cost splitting structure model. First, the type and probability of failure modes for each cost splitting structure model at the lowest level can be calculated. The calculation formula is as follows.

Here, as the type of failure mode, failure mode J having the maximum failure probability may be determined as a failure mode in i of the cost division structure module.

After calculating the types and probabilities of failure modes for each cost division model at the lowest level, the types and probability of failure modes for each cost division structure model at the middle level can be calculated.

Specifically, when calculating the types and probabilities of failure modes for each intermediate-level cost splitting structure model, first, weights are determined for each type of failure mode, the lowest level adjusted fixed probability is calculated, and based on the calculated lowest level adjusted fixed probability, Failure modes and failure probabilities for each intermediate level cost splitting structure model can be calculated.

First, in the method of determining the weight for each type of failure mode, assuming that the type of failure mode, for example, the type of failure mode consists of inspection, test, partial repair, and replacement, the weight of inspection and test is given as 1, , the weight of partial repair is given as 10, and the weight of replacement for failure can be given as 100. That is, the type of failure mode can be considered with a weighted value in consideration of the severity of the failure.

)은 다음과 같은 계산식을 통해 산출할 수 있다.After that, the lowest level adjustment fixed probability (

) can be calculated through the following formula.

(j=check, test k=1 / j=partial repair k=2 / j=replacement k=3)

For example, the lowest level adjustment fixed probability can be calculated differentially depending on the case where the failure mode is inspection or test, partial repair, or failure replacement, and there are various types of failure modes in addition to the above types. can be included

After calculating the weight and the lowest level adjustment fixed probability, the failure mode type and probability for each intermediate level cost splitting structure model can be calculated, and can be calculated through the following formula.

)이 최대인 하위 비용분할 모듈 I에서 고장 모드 JI가 중간 레벨 비용분할구조 모델에서의 고장모드로 결정될 수 있다.The type of failure mode is the fixed failure probability (

) can be determined as the failure mode in the middle-level cost splitting structure model.

In other words, the types and probabilities of failure modes for each cost splitting structure model of the middle level are determined by 1) calculating the weight, 2) calculating the fixed probability adjusted to the lowest level, and 3) calculating the type and probability of failure modes for each cost splitting structure model. and probability can be calculated.

In addition, the method of calculating the type and probability of failure modes for each cost division structure model of the upper level and the top level can be performed in the same way as the method of calculating the type and probability of failure mode for each cost division structure model of the middle level described above. .

First, the methods for calculating the types and probabilities of failure modes for each high-level cost splitting model are as follows: 1) Weight calculation 2) Lowest level adjusted fixed probability calculation 3) Mid-level adjusted fixed probability calculation 4) Failure mode by cost splitting structure model It can be finally calculated through type and probability calculation.

In addition, the methods for calculating the types and probabilities of failure modes for each top-level cost splitting structure model are: 1) Calculation of weights 2) Calculation of fixed probabilities adjusted for the lowest level 3) Calculation of fixed probabilities adjusted for the lowest level 4) Calculation of fixed probabilities adjusted for the upper level 5 ) It can be finally calculated through the calculation of failure mode types and probabilities for each cost division structure model.

That is, the method for calculating the failure mode type and probability for each higher level or top-level cost splitting structure model is the same as the method for calculating the failure mode type and probability for each weight and cost splitting structure model, and each level adjustment fixed probability It may be made of a structure in which a method of calculating is added.

The failure cost management module 30 performs a function of managing a loss cost due to failure for each cost division structure model and an input cost for solving failures for each cost division structure model. Specifically, the cost calculated by the failure probability calculation module Loss cost and input cost may be determined based on the type and probability value of the failure mode for each level of each divided structure model, and the determined value may be transmitted to a manager or separately stored in the database 60 .

The predictive maintenance establishment module 40 performs a function of calculating whether to perform predictive maintenance considering the cost of each cost division structure model. It is possible to finally calculate whether or not to implement the predictive maintenance by calculating the input cost for each and calculating the probability of performing the predictive maintenance based on the above calculated value.

For example, whether to implement predictive maintenance can be finally calculated as predictive maintenance if the adjusted loss cost is calculated to be higher than the adjusted input cost, and if the adjusted loss cost is calculated to be lower than the adjusted input cost, the final calculation is to not implement the predictive maintenance. It can be. In addition, when the adjusted loss cost and the input cost are calculated to be the same, it can be implemented with a preset value in the database. Here, if it has been previously set to perform predictive maintenance when the same loss cost and input cost are calculated in the database, the above information may be configured to perform predictive maintenance.

The predictive maintenance management module 50 performs predictive maintenance based on the function of executing the failure prediction module 10 for each period preset in the database 60 and the predictive maintenance probability calculated by the predictive maintenance establishment module 40 and whether or not it is implemented. A report can be generated, and the predictive maintenance report can be transmitted to a management server (not shown) operated by an administrator or stored separately in the database 60 .

In addition, the predictive maintenance management module 50 may calculate a system level failure probability based on the value calculated by the failure probability calculation module, and establish a predictive maintenance plan based on the system level failure probability. It can be transmitted to the management server or stored separately in the database 60.

That is, according to an embodiment of the present invention, a failure prediction model can be built based on a cost division structure (CBS) model, and failure mode types from the lowest level module to the highest level module of the cost division structure model and Each failure probability can be calculated, and a predictive maintenance plan can be established based on the calculated value.

In addition, according to an embodiment of the present invention, the type and probability of failure modes occurring for each cost division structure model for each level can be calculated, and a predictive maintenance plan can be established based on the calculated reliability and accuracy of the predictive maintenance plan. can improve

Although the present invention has been described in detail through an embodiment, this is for explaining the present invention in detail, and the predictive maintenance operating system based on the failure prediction model according to the present invention is not limited thereto. In addition, terms such as "comprise", "comprise", or "have" described above mean that the corresponding component may be inherent unless otherwise stated, so excluding other components is not recommended. All terms, including technical or scientific terms, are generally understood by those skilled in the art to which the present invention belongs, unless otherwise defined. has the same meaning as being

In addition, the above description is merely an example of the technical idea of the present invention, and various modifications and variations can be made to those skilled in the art without departing from the essential characteristics of the present invention. Therefore, one embodiment disclosed in the present invention is not intended to limit the technical idea of the present invention, but to explain, and the scope of the technical idea of the present invention is not limited by this embodiment. The protection scope of the present invention should be construed according to the claims below, and all technical ideas within the equivalent range should be construed as being included in the scope of the present invention.

10 - Failure prediction module
20 - Failure probability calculation module
30 - Failure cost management module
40 - Predictive Maintenance Establishment Module
50 - predictive maintenance management module
60 - database

Claims (11)

A failure prediction module for predicting the duration and occurrence probability of failure modes of one or more CBS models built in one or more power plants, and storing the predicted occurrence probability in a database;
Based on the failure mode type and failure probability stored in the failure prediction module or the database, the failure mode type and failure probability to occur for each cost division structure model are determined, and cost division of the lowest level, middle level, upper level, and top level is performed. a failure probability calculation module that determines failure modes and failure probabilities for each structure;
Based on the failure mode and failure probability for each cost division structure of the lowest level, middle level, upper level, and top level determined by the failure probability calculation module, the loss cost due to failure for each cost division structure and the input cost for failure resolution Failure cost management module to manage;
A predictive maintenance establishment module that calculates a predictive maintenance probability considering the amount of each cost division structure based on the loss cost and input cost managed by the failure cost management module, and determines whether to implement based on this calculation;
a predictive maintenance management module that executes the failure prediction module for each period preset in the database and generates a predictive maintenance report based on the predictive maintenance probability and whether or not the predictive maintenance is performed, calculated by the predictive maintenance establishment module; A predictive maintenance operating system based on a failure prediction model that includes a.
The method of claim 1,
The failure prediction module sets a failure prediction model for calculating the probability of occurrence of the failure mode by setting at least one of 1 day, 7 days, 30 days, 90 days, 180 days, and 365 days later. based predictive maintenance operating system.
The method of claim 1,
In the failure probability calculation module, the failure mode and failure probability that will occur for each cost division structure are,

(p: power plant power / i: cost division structure module, 1,2,3, n / j: failure mode, 1,2,3, n), a predictive maintenance operating system based on a failure prediction model.
The method of claim 3,
The failure mode type and probability in the cost division structure model of the lowest level in the failure probability calculation module are,

including,
A predictive maintenance operating system based on a failure prediction model in which the failure mode J with the maximum failure probability is determined as the failure mode in the cost partitioning structure module i.
The method of claim 1,
The failure probability calculation module determines a weight for each failure mode type, calculates the lowest level adjusted failure probability, and calculates the failure mode and failure probability for each intermediate level cost division structure model based on the lowest level adjusted failure probability. A predictive maintenance operating system based on a failure prediction model.
The method of claim 5,
The weight determination is a predictive maintenance operating system based on a failure prediction model that assigns weights differently for each inspection, test, partial repair, and replacement, and assigns different weights in the order of inspection=test<partial repair<failure replacement. .
The method of claim 6,
The lowest level adjustment failure probability (

)silver

(j=check, test k=1 / j=partial repair k=2 / j=replacement k=3), a predictive maintenance operating system based on a failure prediction model.
The method of claim 7,
The failure mode types and probabilities for each of the intermediate level cost splitting structure models are,

is,
The failure mode type is a fixed failure probability (

) is maximum, a predictive maintenance operating system based on a failure prediction model in which the failure mode JI is determined as the failure mode in the middle-level cost splitting structure model in the lower cost splitting module I.
The method of claim 8,
The method for calculating the type and probability of failure modes for each higher-level or top-level cost division structure model is calculated by calculating the weight calculation value, the adjusted failure probability value for each level, and the type and probability of failure mode for each cost division structure model. A predictive maintenance operating system based on a failure prediction model that is finally calculated through the result value.
The method of claim 1,
The predictive maintenance establishment module performs predictive maintenance when the loss cost is higher than the input cost, and stores the cost division structure failure mode and predictive maintenance in the database. A predictive maintenance operating system based on a failure prediction model.
The method of claim 1,
The predictive maintenance management module calculates a system level failure probability based on the value calculated by the failure probability calculation module, establishes a predictive maintenance plan based on the system level failure probability, and stores the established plan in a database , a predictive maintenance operating system based on a failure prediction model.

Images