CN111502924A

CN111502924A - Power generation equipment maintenance management system and method based on prediction in advance by using data analysis indexes

Info

Publication number: CN111502924A
Application number: CN202010334211.4A
Authority: CN
Inventors: 金桢雨; 韩基軓; 蔡盛基
Original assignee: A2m Co ltd
Current assignee: A2m Co ltd
Priority date: 2019-11-20
Filing date: 2020-04-24
Publication date: 2020-08-07
Also published as: KR102088234B1

Abstract

The invention relates to a maintenance and management system and a method for power generation equipment, in particular to a maintenance and management system and a method for power generation equipment based on pre-prediction by using data analysis indexes. To this end, a construction system is provided, comprising: a control unit (10), a sensing unit (20), an arithmetic unit (30), a storage unit (40), and an information providing unit (50), and a configuration method is provided that includes: a data obtaining step (S10), a normal operation condition range selecting step (S20), a data link table creating step (S30), a representative value updating step (S40), and a maintenance management cost normalizing step (S50) depending on the maintenance history.

Description

Power generation equipment maintenance management system and method based on prediction in advance by using data analysis indexes

Technical Field

The invention relates to a maintenance and management system and a method for power generation equipment, in particular to a maintenance and management system and a method for power generation equipment based on pre-prediction by using data analysis indexes.

Background

Generally, power generation equipment exists in a variety of forms.

The typical power generation facilities include hydraulic power, thermal power and nuclear power generation, but the hydraulic power, thermal power and nuclear power generation have recently been faced with remarkable problems such as carbon dioxide emission and environmental pollution.

Therefore, alternative energy of various forms has become an issue, and wind power generation and solar power generation are typical among them.

Among them, a typical power generation facility is a wind power generation facility, which is greatly affected by meteorological factors including wind as an external factor according to characteristics of wind power generation, and a huge blade is applied for a structure for rotating the blade and large-capacity power generation because the facility is driven by wind according to characteristics of the facility.

In the case of such wind power generation, the amount of impact applied to the equipment is increased as compared with other power generation equipment, and therefore, there is a problem in that it is difficult to predict the critical life of each component or member.

In view of the above problems, many methods have been tried, but in reality, maintenance and management are mostly performed based on the rule of thumb of an administrator or a user, a critical life indication of a manufacturing company, and the like, and it is difficult to predict the life of a component or to cope with it in advance, and a follow-up measure is taken after a failure occurs.

Therefore, the downtime due to replacement of the relevant parts and the like and the downtime due to supply and demand of the parts are continued, and a problem of power generation error often occurs.

To overcome such problems, korean registered patent No. 10-1757117 (wind turbine generator maintenance device and method, hereinafter referred to as "advanced technology") relates to a maintenance device and method for determining maintenance operation timing based on material information, manpower information, and/or weather information when establishing a maintenance plan, the maintenance device managing a plurality of material sites where materials are kept and a plurality of wind parks where wind turbines requiring maintenance are installed. A wind turbine maintenance system for planning maintenance work of a wind turbine dispersed at a remote place is disclosed, comprising: the system comprises a wind power park website server, a plurality of wind power generators and a plurality of server management servers, wherein the wind power park website server is positioned in a wind power park where the plurality of wind power generators are positioned and used for acquiring state information of the plurality of wind power generators; a material storage site web server located at a material storage site for storing materials required for maintenance work of the wind turbine generator, and configured to obtain information on the materials stored in the material storage site; and a wind turbine maintenance device for determining a maintenance operation timing based on the state information of the wind turbine transmitted from the plurality of wind park site servers, the material information transmitted from the plurality of material stocking site servers, and the position information of the plurality of wind park sites and the plurality of material stocking site. According to the present invention, it is possible to minimize the proximity according to the weather state, the influence of supply and demand of parts and equipment, workers, etc., and to make and implement a maintenance plan by sensing the failure of parts at an early stage, thereby preventing a major accident.

In the case of this prior art, since it is only possible to predict the timing of failure of the component and replace the component, there is a problem that it is necessary to separately determine the difficulty level of the component in accordance with the proficiency level of the user or the manager in the portion related to the maintenance and management important cost.

Prior art documents

Patent document

(patent document 1) korean registered patent No. 10-1757117 (wind turbine generator maintenance device and method, registration of 7 months and 5 days in 2017)

Disclosure of Invention

In order to overcome the above-described inconvenience, an object of the present invention is to provide a technique for normalizing and functionalizing maintenance and management costs based on data by using the history of maintenance and repair, thereby efficiently performing maintenance and management of a power generation facility.

To achieve the object of the present invention, there is provided a power plant maintenance management system based on prediction in advance using data analysis indicators, the system including: a control unit 10 for receiving input values obtained and set by the sensing unit 20 and the computing unit 30 to control the system; a sensing unit 20 provided in the accessories and components of the power generation facility, for obtaining information from a plurality of sensors for measuring and sensing information, and for transmitting the obtained information to the arithmetic unit 30 and the storage unit 40 via the control unit 10; a calculation unit 30 that calculates maintenance and management costs for the parts or components using the information obtained from the sensing unit 20; a storage unit 40 for storing information, history, measurement values, and details of each component and part calculated by the calculation unit 30 and converting the information, history, measurement values, and details into a Database (DB); and an information providing unit 50 for providing information to a user or a manager.

The calculation unit 30 sends out a cost function ((GQ × P + GQ × Ptd) ÷ 2) × T × C-MC ═ g q × Ptd) × T × C, and calculates, after mounting, the cost function, where GQ is the amount of electricity generated, P is the initial operating performance, T is the operating time, d is the operating performance attenuation rate, T is the power generation operation stop time, C is the cost for power generation, and MC is the replacement cost of accessories.

Further, the power generation facility maintenance management system based on the prediction in advance using the data analysis index is characterized in that the information provided by the information providing unit 50 can provide information to a user or a manager in the form of a data link table linked to data such as maintenance history, and the link table created by the information providing unit 50 can be stored in the storage unit 40 and made into a database.

According to still another embodiment of the present invention, there is provided a power plant maintenance management method based on a prediction using a data analysis index, including: a data acquisition step S10 of measuring information from the sensing unit 20 that senses each of the accessories and components of the wind turbine generator and acquiring sensed information; a normal operation condition range selection step S20, selecting the accumulated operation value obtained in operation before the set time as the normal operation condition range by using the time of sensing the event condition as the reference when sensing the condition needing to be maintained or the events such as various states in remote areas and SCADA-error codes; a data contact list making step S30, setting data invalid values, maximum values and central values of all sensors before the time when the condition needing maintenance or the SCADA error code is sensed according to specific time and making a data contact list; a representative value updating step S40 of continuously accumulating the contact table during the generation of the maintenance elapsed time, thereby updating the representative value; the normalization step S50 of the maintenance management cost depending on the maintenance history uses the data link table created in the data link table creation step S30 and the representative value calculated based on the representative value and updated in the representative value update step S40 to grasp the critical life of the component or part and to perform the cost normalization based on the representative value.

The normal operation condition range may be adjusted by reflecting the condition of the external factor corresponding to the external factor of the current operation in accordance with the change of the external factor such as weather and season in the normal operation based on the data value of 2 months or more.

Further, the object of the present invention is achieved by providing a power plant maintenance management method based on a prediction in advance using a data analysis index, characterized in that the cost function for normalizing the cost of the step S50 depending on the maintenance management cost experienced in the maintenance is ((GQ × P + GQ × Ptd) ÷ 2) × T × C-MC (GQ × Ptd) × T × C, GQ: power generation amount, P: initial operation performance, T: operation time, d: operation performance attenuation rate, T: power generation operation stop time, C: cost due to power generation, MC: replacement cost of accessories.

The maintenance and management system and method for power generation facilities using data analysis indexes according to the present invention are advantageous in that maintenance and management of power generation facilities can be efficiently performed by normalizing and functionalizing costs for maintenance and management using maintenance and repair history and the like.

Drawings

Fig. 1 is a configuration diagram of a system according to the present invention.

FIG. 2 is a sequence diagram of a power plant maintenance management method based on pre-prediction using data analysis metrics in accordance with the present invention.

Fig. 3 is an exemplary diagram illustrating a data association table according to the present invention.

Description of the reference symbols

10: control unit

20: sensing part

30: arithmetic unit

40: storage unit

50: information providing unit

Detailed Description

The power plant maintenance management system and method using data analysis indexes based on prediction according to the present invention will be described in detail below with reference to the accompanying drawings so that a general technician can easily implement the system.

Referring to fig. 1, the power plant maintenance management system based on prediction in advance using data analysis indexes according to the present invention includes a control unit 10, a sensing unit 20, a calculation unit 30, a storage unit 40, and an information providing unit 50.

The control unit 10 is responsible for overall control of the system, and receives input values obtained and set by the sensing unit 20 and the computing unit 30, thereby controlling the system.

The sensing unit 20 senses each component and part, obtains information from a plurality of sensors provided in each component and part and measuring and sensing a plurality of kinds of information, and transmits the obtained information to the arithmetic unit 30 and the storage unit 40 through the control unit 10.

In this case, various sensors are applicable as the sensor applied to the sensing unit 20, and typically, various sensors such as a vibration sensor for sensing vibration, a rotation sensing sensor for sensing the motion of a driven body such as a generator or a motor, and the like are applicable.

The calculation unit 30 is responsible for calculating maintenance and management costs of the parts or components using the information obtained from the sensing unit 20.

In this case, the calculation unit 30 calls the cost function stored in the storage unit 40 using the cost function stored in the storage unit 40, and carries the call in the calculation unit 30 to perform calculation.

The cost function described above is as follows.

((GQ×P+GQ×Ptd)÷2)×T×C-MC＝(GQ×Ptd)×T×C

GQ: electric energy production

P: initial running performance

t: running time

d: rate of reduction of running performance

T: power generation operation stop time

C: cost of electricity generation

MC: replacement cost of fittings

At this time, the arithmetic unit 30 calls the calculation formula according to the various situations and management conditions stored in the storage unit 40, and applies the calculation formula according to the situation requested by the administrator.

The storage unit 40 stores information, history, measurement values, and details of the calculation performed by the calculation unit 30 for each component and part, and creates a database.

At this time, the user or the manager inputs the critical life of the component or the member and sets a reference value, reflects data repeatedly obtained from the set reference value, and upgrades the reflected data to actual data by deep learning.

The user obtains the set reference value and the actual data according to the manufacturing company or the experience rule, and judges whether the data is similar to the upgraded data, and when the reference value set by the user is different from the data which is actually obtained and calculated, the user is informed so as to guide the setting of the more accurate reference value.

Whether the above-mentioned reference value is similar to the information obtained by the sensing section 20 is judged, and when the reference value is reached, replacement of the component or the like can be judged after guidance thereof.

The information providing unit 50 can provide information to a user or a manager, and provide various forms of information.

The user or manager has a personal communication terminal for displaying or providing information to an external integrated control service.

Such information is not only information on the corresponding power generation facility but also information on other areas, so that the critical physical characteristics of the corresponding component or the corresponding part can be grasped more precisely.

This is because it is necessary to reflect the characteristics of the region where the power generating equipment is installed, the seasonal characteristics, and the weather characteristics of external factors, and therefore, by reflecting such characteristics, the critical physical characteristics of the corresponding parts and components can be grasped more accurately.

In this way, by grasping the critical physical characteristics and predicting the life thereof, it is possible to determine a management method such as replacement or maintenance of an actual component using the relevant information, and in this case, a cost function is used and a cost for maintenance management is applied as a criterion for the determination.

More specifically, the remaining life of the accessories is predicted, and the predicted life is compared with the amount of damage caused by the stop of power generation at the time of replacement in an emergency such as replacement.

For example, if it is said that the new component has 100% of the operating performance, the cost can be calculated by comparing the power generation amount of the operation stop time due to the repair after the operating performance degradation rate at the time of the occurrence of the emergency state is grasped.

The information provided by the information providing unit 50 can be provided to a user or a manager in the form of a data link table (see fig. 3) linked to data such as a maintenance history, and the link table created by the information providing unit 50 is stored in the storage unit 40 and is made into a database.

Hereinafter, a maintenance management method for a power generation facility based on a prediction in advance using a data analysis index according to the present invention will be described in detail with reference to a sequence diagram of a maintenance management method for wind power generation of fig. 2.

The power plant maintenance management method using data analysis index based on prediction in advance according to the present invention includes: a data obtaining step S10, a normal operating condition range selecting step S20, a data link table making step S30, a representative value updating step S40, and a normalizing step S50 depending on maintenance management costs of the maintenance history.

1. Data obtaining step S10

In the data acquisition step S10, a plurality of types of information are measured from the sensing unit 20 that senses each of the accessories and components of the wind turbine generator, and the sensed information is acquired.

In this case, as a device for obtaining information from the sensing unit 20, a plurality of types of measurement sensors are used, data obtained by the measurement sensors are continuously obtained and accumulated, and based on the continuously measured and accumulated data, measured values are calculated and made into a database on the basis of the failure value, the maximum value, and the central value in units of hours, days, weeks, and months.

In the continuous measurement, data values before the last measurement time are accumulated and calculated, and the data values are stored in a database.

2. Normal operation condition Range selection step S20

The normal operation condition range selection step S20 is to collect the condition to be maintained or various states and Information in a Remote area to a Programmable logic Controller (Programmable L organic Controller: P L C) or a Remote Terminal Unit (RTU), transmit a central monitoring and control System and a database to an Information processing System (Information Management System) as a fixed point through a network, analyze and control the evolution of transmitted data values, and thus sense SCADA-error codes for effectively operating various devices.

In this case, the normal operation condition range is usually based on data values two months or more, but is not limited thereto, and may reflect variables according to external factors.

For example, in the normal operation, a change in an external factor such as weather or season is reflected, and a condition corresponding to the external factor in the current operation is reflected, whereby the normal operation condition range can be adjusted.

This is because the value measured by the external factor can be measured differently in order to minimize the error range of the normal operation condition range selected from the data accumulated and calculated by weather, season, etc. of the external factor in the current operation state, and because an error may occur when the abnormality is judged by a general simple comparison.

3. Data contact sheet making step S30

The data association table creation step S30 sets invalid values, maximum values, and central values of data of all sensors before the time when the condition requiring maintenance or the SCADA error code is sensed, at a specific time, and creates a data association table.

Here, the specific time is set in terms of hour, day-hour, week-day-hour, month-week-day-hour.

For example, a day-hour is set to 1 day as a period, the set 1 day is set for a time period (0 to 23), a week-day-hour is set for a period of 1 week, a set time period of 1 week and every day is set, a month-week-day-hour is set for a period of 1 month, and data of the above specific times are set for a week, a day and a time of 1 month and accumulated in total, whereby a case requiring maintenance or an event such as a SCADA-error code can be grasped for each time.

This is to improve the reliability of sensing an event situation, i.e. a need for maintenance or a data value of a SCADA-error code.

If specific data are sensed according to time, day data are compared, week data are compared after day data are compared, and then month data are compared, so that judgment can be carried out according to the situation of the comprehensive events. (refer to FIG. 3)

4. Representative value updating step S40

The representative value updating step S40 continuously accumulates the contact table during the generation of the repair pass, thereby updating the representative value.

Such representative value is a continuously accumulated value, and among the values accumulated and calculated for the longest time, a value at the time of occurrence of an abnormality in the event situation or a value set arbitrarily by the user may be specified.

In this case, the representative value may be automatically updated as data that can recognize and sense the occurrence of an event, or may be manually set and updated by a user or a manager.

Here, the representative value reflecting variable according to the external factor is additionally set for each factor representative value reflecting variable according to the external factor, and is stored in the storage unit 40, and then each factor representative value corresponding to the external factor at the time of occurrence of the event can be transmitted and applied.

For example, among external factors reflecting seasonal factors, air temperature, wind direction and air volume, time factors, and the like, the expired breath reflects the same or similar factors and stores representative values of the respective factors.

More specifically, a representative value is simply calculated and set based on data obtained from a sensor without reflecting an external factor, and the representative value is applied or another representative value depending on the case may be set for a factor reflecting an external factor.

5. Normalization step S50 of maintenance management costs depending on repair experience

The maintenance management cost-dependent normalization step S50 uses the data link table created in the data link table creation step S30 and the representative value calculated and updated in the representative value update step S40 to determine the critical life of the component or part, and uses this as the calculation by the calculation unit 30 to normalize the cost.

At this time, if the information obtained by the sensing part 20 corresponds to a representative value that is a predicted value of the critical life of the component, the information is notified to a user or a manager, so that the direction regarding maintenance management can be determined.

For this reason, the normalization of the maintenance management costs can be achieved using a cost function.

Such a cost function is as follows.

((GQ×P+GQ×Ptd)÷2)×T×C-MC＝(GQ×Ptd)×T×C

GQ: electric energy production

P: initial running performance

t: running time

d: rate of reduction of running performance

T: power generation operation stop time

C: cost of electricity generation

MC: replacement cost of fittings

Here, gq (generation quantity) refers to an amount of electric power generation, P refers to initial operation performance of the corresponding component, T refers to operation time of the corresponding component, d refers to an operation performance attenuation rate, and T refers to a power generation operation stop time for stopping the power generation operation for repair or the like for replacement of accessories in an emergency.

The above-described equation is such that, at the time of initial start of the corresponding component, the start time T at the time of occurrence of an emergency and the start performance decay rate d of the component measured based on data accumulated in the database are applied to the total power generation amount (GQ × P) based on the power generation amount to calculate the average value of the total power generation amount (GQ × Ptd) at the time of occurrence of the emergency, the power generation amount (TGQ) based on the power generation operation stop time T is calculated and the cost based on the power generation amount is calculated, after the replacement cost (MC) of the component is reduced, the estimated power generation amount (TGQ) capable of generating power during the power generation operation stop time T is calculated from the total power generation amount (GQ × Ptd) at the time of occurrence of the event, after the cost based on the power generation amount is calculated, the timing of replacement of the component is determined when the two prices are the same, and maintenance management is performed when the cost generated based on the estimated power generation amount (.

This is more effectively applicable to maintenance management by making the costs as described above functional.

The maintenance management system for power generation and the method thereof of the present invention are provided according to the method as described above.

Claims

1. A pre-prediction based power plant maintenance management system using data analysis metrics, comprising:

a control unit (10) which receives input values obtained and set by the sensing unit (20) and the arithmetic unit (30) to control the system;

a sensing unit (20) which is provided in the accessories and components of the power generation facility, acquires information from a plurality of sensors that measure and sense the information, and transmits the acquired information to the arithmetic unit (30) and the storage unit (40) via the control unit (10);

a calculation unit (30) that uses the information obtained from the sensing unit (20) and is responsible for calculating the maintenance and management costs of the parts or components;

a storage unit (40) for storing information, history, measurement values, and details of the components and parts calculated by the calculation unit (30) and making them into a database;

an information providing unit (50) that can provide information to a user or a manager;

the calculation unit (30) calls a cost function stored in the storage unit (40) and calculates the cost function after mounting, wherein the cost function is ((GQ × P + GQ × Ptd) ÷ 2) × T × C-MC ═ GQ × Ptd) × T × C,

GQ: power generation amount, P: initial running performance, t: operating time, d: running performance attenuation rate, T: power generation operation stop time, C: cost of electricity generation, MC: replacement costs of the fitting.

2. The pre-prediction based power plant maintenance management system using data analysis metrics of claim 1,

the information provided by the information providing unit (50) can provide information to a user or a manager in the form of a data link table linked with data such as maintenance history, and the link table created by the information providing unit (50) can be stored in the storage unit (40) and made into a database.

3. A power plant maintenance management method based on prediction in advance using data analysis indicators, comprising:

a data acquisition step (S10) for measuring information from a sensing unit (20) that senses each of the accessories and components of the wind turbine generator and acquiring sensed information;

a normal operation condition range selection step (S20) for selecting an accumulated operation value obtained during operation before a set time as a normal operation condition range on the basis of the time of sensing the event condition when sensing the condition requiring maintenance or the event such as various states in a remote area and an error code of a data acquisition and monitoring system;

a data contact list making step (S30) for setting data invalid values, maximum values and central values of all sensors before the condition needing to be maintained or the time of error codes of the data acquisition and monitoring system are sensed according to specific time and making a data contact list;

a representative value updating step (S40) of continuously accumulating the contact list during the generation of the maintenance lapse, thereby updating the representative value;

a normalization step (S50) of the maintenance management cost depending on the maintenance history, grasping the critical life of the component or part using the data link table created in the data link table creation step (S30) and the representative value calculated based on the representative value and updated in the representative value update step (S40), and performing the cost normalization based on the grasped life;

the normal operation condition range is based on data values of 2 months or more, and for normal operation, the condition of the external factor corresponding to the external factor of the current operation is reflected according to the change of the external factor such as weather, season, and the like, so that the normal operation condition range can be adjusted.

4. The power plant maintenance management method using data analysis index based on prediction according to claim 3,

the normalization of the cost according to the step of normalizing the maintenance management cost (S50) of the repair experience uses a cost function.

5. The power plant maintenance management method using data analysis indicators based on prediction according to claim 4,

the cost function is ((GQ × P + GQ × Ptd) ÷ 2) × T × C-MC (GQ × Ptd) × T × C,