KR102056701B1 - System and method for diagnosing status - Google Patents

Abstract

Condition diagnosis systems and methods are disclosed. A state diagnosis system according to an embodiment of the present invention is a state diagnosis system that infers a state value according to an input sensing value by using a sensing value collected for a predetermined period and a state value according to the collected sensing value. A data condensation unit for assigning representative values to the values, and condensing the plurality of cases in which the representative values of the sensing values are the same; And a state inference unit for extracting a region belonging to the input sensing value from the virtual space, measuring a similarity between the input sensing value and the cases belonging to the extracted region, and inferring a state value according to the input sensing value.

Description

Condition diagnosis system and method {SYSTEM AND METHOD FOR DIAGNOSING STATUS}

Embodiments of the present invention relate to a condition diagnosis system and method, and more particularly, to a condition diagnosis system and method for diagnosing a current state using a case that has occurred in the past.

The diagnosis of the condition of existing facilities and equipment is mainly performed by a specialist who puts data collected from sensors installed in each facility and equipment into a rule set by an expert. However, according to the conventional method, if the data collected from the sensors are similar to the conditions of a predetermined rule but do not completely match, there is a problem that it is difficult to determine the state of the facility and equipment to be diagnosed.

In addition, according to the existing method, the accuracy and reliability of the diagnosis result may not be reflected due to the fact that the probability of occurrence of failure varies according to the usage history of each facility and equipment, and the problem of failing to continuously reflect the newly generated failure pattern may not be reflected. There is this. In addition, unless highly trained specialists are put in and analyze the current status of each facility and equipment, there is a problem in that the failure state of the facilities and equipment cannot be detected and the immediate response is not provided.

An embodiment of the present invention is to provide a condition diagnosis system and method that can reduce the time required to diagnose the condition.

A state diagnosis system according to an embodiment of the present invention is a state diagnosis system for inferring a state value according to a sensing value input using a sensing value collected for a predetermined period and a state value according to the collected sensing value. A data condensation unit for assigning a representative value to each of the sensing values, and condensing a plurality of cases in which the representative values of the sensing values are the same; A space partitioning unit for disposing each case in a virtual space and dividing the virtual space into a plurality of regions; And a state inference unit extracting an area in which the input sensing value belongs to the virtual space, measuring a similarity between the input sensing value and the cases belonging to the extracted area, and inferring a state value according to the input sensing value. do.

According to an embodiment of the present invention, a method for diagnosing a condition may include: assigning a representative value to each sensing value stored in a database, and reducing the plurality of cases in which representative values of the sensing values are the same; A space partitioner disposing each case in the database into a virtual space, and dividing the virtual space into a plurality of regions; And extracting an area in which the sensing value inputted by the state inference unit belongs to the virtual space, and inferring a state value according to the input sensing value by measuring similarity between the input sensing value and cases belonging to the extracted area. It includes.

According to an exemplary embodiment of the present invention, an absolute data amount can be reduced by assigning a representative value to each sensing value and then reducing a plurality of cases having the same representative value of each sensing value into one case. And, by arranging each case in a virtual space, by dividing the virtual space into a plurality of areas, the amount of calculation can be reduced by measuring the similarity only with the cases of the input sensing value in the area belonging to the virtual space, thereby reducing equipment Alternatively, it is possible to quickly perform a condition diagnosis of the device.

1 is a view showing the configuration of a condition diagnosis system according to an embodiment of the present invention.
2 is a diagram illustrating a sensing value stored in a database according to an embodiment of the present invention.
3 is a diagram illustrating a state in which a data abbreviation unit assigns representative values to respective data according to an embodiment of the present invention.
4 is a diagram illustrating a state in which a data abbreviation unit shortens a plurality of cases in which representative values of respective sensing values are identical to each other, according to an embodiment of the present invention.
5 is a diagram illustrating a state in which a space partition unit divides a virtual space into a plurality of regions by using a tree according to an embodiment of the present invention.
6 is a flowchart illustrating a state inference method according to an embodiment of the present invention.

Hereinafter, a system and method for diagnosing a condition of the present invention will be described in detail with reference to FIGS. 1 to 6. However, this is only an exemplary embodiment and the present invention is not limited thereto.

In describing the present invention, when it is determined that the detailed description of the known technology related to the present invention may unnecessarily obscure the subject matter of the present invention, the detailed description thereof will be omitted. In addition, terms to be described below are terms defined in consideration of functions in the present invention, which may vary according to the intention or custom of a user or an operator. Therefore, the definition should be made based on the contents throughout the specification.

The technical spirit of the present invention is determined by the claims, and the following embodiments are merely means for effectively explaining the technical spirit of the present invention to those skilled in the art to which the present invention pertains.

1 is a view showing the configuration of a condition diagnosis system according to an embodiment of the present invention. In the following description, a building management system is described as an example in which a condition diagnosis system is applied, but a field in which the condition diagnosis system is applied is not limited thereto and may be applied to various other fields.

Referring to FIG. 1, the state diagnosis system 100 includes a database 102, a data collection unit 104, a data condensation unit 106, a space partition unit 108, and a state inference unit 110.

The database 102 may store sensing values collected by the data collector 104 for a predetermined period of time and state values of the corresponding equipment (or apparatus) according to each collected sensing value.

2 is a diagram illustrating a sensing value stored in a database according to an embodiment of the present invention. Referring to FIG. 2, the database 102 stores sensing values s1 and s2 periodically collected from the data collector 104 including a first sensor and a second sensor (not shown) in the past. . Here, the first sensor and the second sensor (not shown) may be, for example, a temperature sensor installed in a cooling device in a building, but is not limited thereto.

In addition, the database 102 stores the state values of the corresponding cooling devices according to the sensing values s1 and s2 collected at each predetermined time point. At this time, when the state value is 0, it indicates that the cooling device is in a normal state, and when the state value is 1, it indicates that the cooling device is in a failure state. That is, each of the cooling devices has a value of one of 0 or 1 as its state value, where 0 and 1 are defined as state setting values of the cooling device. In the embodiment of the present invention, the state setting value refers to a value set in advance to represent the state of each data stored in the database 102. Here, a series of sensing values s1 and s2 collected at a certain time point and a state value thereof form one case. N cases may be stored in the database 102.

Here, although the database 102 has been described as storing the sensing values s1 and s2 collected from the sensors installed in the air conditioner and the state values according to the sensing values collected at a certain point in time, the database 102 is not limited thereto. ) May store sensing values collected from sensors installed in various other devices or devices as well as air conditioners and corresponding state values of the corresponding devices (or devices).

The data collector 104 collects data (ie, sensing values) from each sensor (not shown) installed in a facility or equipment for which a condition is to be diagnosed. For example, the data collection unit 104 senses values from various sensors (eg, temperature sensor, pressure sensor, humidity sensor, motion sensor, etc.) installed in an air conditioner, a cooling device, a heating device, a security device, and the like in a building. Can be collected. In this case, the data collector 104 may periodically collect sensing values from each sensor (not shown), but is not limited thereto and may collect sensing values in real time from each sensor (not shown).

The data abbreviation 106 serves to reduce the number of cases in the database 102. Specifically, the data abbreviation unit 106 assigns a representative value to each sensing value collected during a certain period of time in the database 102. For example, the data reduction unit 106 may divide a range of sensing values stored in the database 102 into n sections, and then assign a representative value to the sensing values belonging to each section.

3 is a diagram illustrating a state in which a data abbreviation unit assigns a representative value to each data according to an embodiment of the present invention. Here, although the range of the sensing value of the corresponding sensor is 24.00 to 29.99, the present invention is not limited thereto, and the sensing value range may be different for each sensor.

Referring to FIG. 3, the data condenser 106 detects a case in which a sensing value is 24.00 to 24.99 as a first section, a case in which a sensing value is 25.00 to 25.99 as a second section, and a sensing value is 26.00 to 26.99. In the third section, a case in which the sensing value is 27.00 to 27.99 may be divided into a fourth section, a case in which the sensing value is 28.00 to 28.99 as a fifth section, and a case in which the sensing value is 29.00 to 29.99 is divided into a sixth section. . For convenience of description, the range of sensing values is divided into six sections, but the present disclosure is not limited thereto and may be divided into various numbers of sections.

The data abbreviation unit 106 assigns a representative value of 24.5 to a sensing value belonging to the first interval among the sensing values stored in the database 102, and gives a representative value of 25.5 to a sensing value belonging to the second interval. , A representative value of 26.5 is assigned to a sensing value belonging to a third section, a representative value of 27.5 is assigned to a sensing value belonging to a fourth section, and a representative value of 28.5 is assigned to a sensing value belonging to a fifth section. In addition, a representative value of 29.5 may be assigned to the sensing value belonging to the sixth section. Here, although the median value of the corresponding section is set as the representative value, the present invention is not limited thereto, and the representative value may be set based on various criteria as necessary.

In this case, the sections of the sensing values collected from each sensor (not shown) may be classified into the same range, but are not limited thereto and may be classified into different ranges. For example, when the first sensing value collected from the first sensor (not shown) and the second sensing value collected from the second sensor (not shown) are stored in the database 102, the range of the first sensing values is Each section may be classified by 1 unit, and the range of the second sensing values may be classified by 0.5 unit.

As such, when a representative value is assigned to each of the sensing values stored in the database 102, a plurality of cases in which the representative values of the sensing values are the same as each other are generated. In this case, the data condensation unit 106 may condense a plurality of cases in which representative values of the sensing values are the same as one case.

4 is a diagram illustrating a state in which a data abbreviation unit shortens a plurality of cases in which representative values of sensing values are identical to each other, according to an exemplary embodiment.

Referring to FIG. 4, it can be seen that representative values of the first sensing value s1 and the second sensing value s2 of the first case, the second case, and the fourth case are the same as 24.5 and 12.2, respectively. This is a phenomenon that occurs as a result of assigning one representative value to sensing values belonging to a predetermined interval, as described above.

In this case, the data abbreviation unit 106 leaves only one case of the first case, the second case, and the fourth case where the representative values of the sensing values s1 and s2 are equal to each other, and the remaining cases are deleted to remove the plurality of cases. Can be abbreviated to one case. For example, the data condenser 106 may delete the second case and the fourth case and leave only the first case. In this case, the remaining state value of the first case may be any one of an average value, a median value, a maximum value, and a minimum value of the state values of the first case, the second case, and the fourth case, but are not limited thereto. no. For example, when the representative value is defined as an average value of each state value, when the state value of the first case is 1, the state value of the second case is 0, and the state value of the fourth case is 1, the remaining first value The status value of the case can be 2/3 (ie 0.666 ...). As such, when a plurality of cases in which representative values of the sensing values are identical to each other is reduced to one case, the total data amount can be reduced.

The space partitioner 108 divides the virtual space into a plurality of regions after arranging each case in an n-dimensional virtual space. In this case, the space partitioner 108 may arrange each case in an n-dimensional virtual space by using a tree structure, and then divide the virtual space into a plurality of regions to structure the case. However, the present invention is not limited to the illustrated embodiment, and the cases may be structured in various forms in addition to the tree structure.

5 is a diagram illustrating a state in which a space partition unit divides a virtual space into a plurality of regions by using a tree according to an embodiment of the present invention.

Referring to FIG. 5, each case is disposed in a two-dimensional virtual space in a root node (RN). For convenience of description, each case includes a first sensing value collected from a first sensor (not shown) and a second sensing value collected from a second sensor (not shown). In this case, each representative value of the first sensing values may be represented on the x-axis, and each representative value of the second sensing values may be represented on the y-axis. However, the present invention is not limited thereto, and if each case includes sensing values collected from n sensors (not shown), each case may be disposed in an n-dimensional virtual space.

Here, the primary child node divides the virtual space of the root node into two areas. The secondary child node divides each virtual space of the primary child node into two areas. The tertiary child node obtains each virtual space of the secondary child node into two regions. The leaf node divides each virtual space of the tertiary child node into two regions. Although the depth of the tree is illustrated as being level 5 herein, the tree is not limited thereto and the tree may be implemented at various levels.

The state inference unit 110 extracts a region belonging to the n-dimensional virtual space from which the sensing value collected by the data collector 104 is located, and then detects the value collected by the data collector 104 and the cases belonging to the region. By measuring the similarity, the state value of the case most similar to the sensing value collected by the data collector 104 may be inferred as a state value according to the sensing value collected by the data collector 104. In this case, it is possible to diagnose the state of the facility (or device) at this point through the inferred state value.

For example, the state inference unit 110 may extract a leaf node to which the sensing values collected by the data collection unit 104 belong by substituting the sensing values collected by the data collection unit 104 into a tree. At this time, the state inference unit 110 may sequentially extract queries from the root node of the tree to each leaf node to extract leaf nodes to which the sensing values collected by the data collector 104 belong.

Here, when the leaf node to which the sensing value collected by the data collector 104 belongs is the fourth leaf node LN4, the state inference unit 110 may detect the sensing value and the fourth leaf collected by the data collector 104. The similarity between the cases belonging to the node LN is measured, and the state value of the case most similar to the sensing value collected by the data collector 104 is inferred as the state value according to the sensing value collected by the data collector 104. can do. In this case, since the similarity needs to be compared only with the cases belonging to the fourth leaf node LN4, the calculation amount can be reduced, and the state value can be inferred quickly. In this case, the similarity may be obtained by measuring the Euclidean distance between the data, but is not limited thereto. Various similarity measurement methods may be used.

Meanwhile, if the state value of the case most similar to the sensing value collected by the data collecting unit 104 is different from the preset state setting value, the state inference unit 110 detects the sensing value collected by the data collecting unit 104 and the corresponding value. The data collector 104 measures the similarity between the actual sensing values (ie, not representative values) of the plurality of cases abbreviated as the cases, and calculates the state values of the cases most similar to the sensing values collected by the data collector 104. It can be inferred as the state value according to the collected sensing value. In this case, as described above, the state setting value refers to a value that is set in advance to express the state of each data stored in the database 102. For example, in the embodiment shown in FIG. 2, since each data has 0 or 1 as the state value, the state set values are 0 and 1 in this case.

Specifically, when the state value of the case most similar to the sensing value collected by the data collection unit 104 among the cases belonging to the fourth leaf node LN is different from the preset state setting value, as shown in FIG. 4. In this case, the representative values of the sensing values are identical to each other so that a plurality of cases are reduced to one case. That is, in the example shown in FIG. 4, when the state value of the first case is 1, the state value of the second case is 0, and the state value of the fourth case is 1, the abbreviated state value is 2/3 (0.666. In this case, the abbreviated state value is different from the state set value 0 or 1. Therefore, the state inference unit 110 measures the similarity between the actual sensing values before the representative values of the plurality of cases abbreviated as the corresponding cases and the sensing values collected by the data collector 104. In this case, it is possible to improve the accuracy of the inferred state value.

For example, referring to FIG. 4, it is assumed that the first case is the case most similar to the sensing value collected by the data collector 104 among the cases belonging to the fourth leaf node LN. Here, the first sensing value s1 of the first case is 24.5, the second sensing value s2 is 12.2, and the state value is 0.666. In this case, the first case is the first case, the second case, and the fourth case is reduced in the state before the case is condensed.

And, in the state before the case is abbreviated, the actual values of the first sensing value s1 and the second sensing value s2 of the first case are 24.9 and 12.1, respectively, and the first sensing value s1 of the second case. And the actual values of the second sensing value s2 are 24.5 and 12.2, respectively, and the actual values of the first sensing value s1 and the second sensing value s2 of the fourth case are 24.1 and 12.3, respectively.

Then, the state inference unit 110 detects the sensing values collected by the data collector 104 and the actual sensing values 24.9 and 12.1 of the first case and the actual sensing values 24.5 of the second case before the case is reduced. , 12.2), and the similarity values between the actual sensing values 24.1 and 12.3 of the fourth case, respectively, and then the sensing values collected by the data collector 104 among the first case, the second case, and the fourth case. The state value of the case most similar to may be inferred as the state value according to the sensing value collected by the data collector 104.

6 is a flowchart illustrating a state inference method according to an embodiment of the present invention.

Referring to FIG. 6, the data condenser 106 divides a range of sensing values stored in the database 102 into n sections, and then assigns a representative value to the sensing values belonging to each section (S 101).

Next, the data condenser 106 extracts a plurality of cases in which representative values of the sensing values are the same, and then condenses the extracted plurality of cases into one case (S 103). For example, the data abbreviation unit 106 may leave only one case among the plurality of cases in which the representative values of the sensing values are identical to each other, and delete the remaining cases, wherein the status value of the remaining case is the state of the extracted plurality of cases. The value can be any one of an average value, a median value, a maximum value, and a minimum value.

Next, the space partition unit 108 arranges each case in the database 102 in an n-dimensional virtual space, and divides the virtual space into a plurality of regions (S 105). For example, the space partitioner 108 may arrange each case in the database 102 in an n-dimensional virtual space using a tree structure, and then divide the virtual space into a plurality of regions to structure the cases.

Next, the state inference unit 110 extracts an area to which the sensing values collected by the data collector 104 belong to the n-dimensional virtual space (S 107). For example, the state inference unit 110 sequentially executes queries from the root node of the tree to each leaf node, so that the leaf node to which the sensing value collected by the data collector 104 belongs (that is, an area in the virtual space). ) Can be extracted.

Next, the state inference unit 110 measures the similarity between the sensing values collected by the data collector 104 and the cases belonging to the extracted region to extract the cases most similar to the sensing values collected by the data collector 104. (S 109).

Next, the state inference unit 110 checks whether the state value of the extracted case is different from the preset state setting value (S 111). As a result of checking in step S 111, when the state value of the extracted case is the same as the preset state setting value, the state inference unit 110 sets the state value of the extracted case according to the sensing value collected by the data collection unit 104. Inferred to (S 113).

When the state value of the extracted case is different from the preset state setting value as a result of checking in step S 111, the state inference unit 110 may determine a sensing value collected by the data collection unit 104 and a plurality of cases abbreviated as the case. The similarity between the actual sensing values (ie, not representative values) is respectively measured (S 115).

Next, the state inference unit 110 infers the state value of the case most similar to the sensing value collected by the data collector 104 as the state value according to the sensing value collected by the data collector 104 (S 117). ).

Meanwhile, an embodiment of the present invention may include a computer readable recording medium including a program for performing the methods described herein on a computer. The computer-readable recording medium may include program instructions, local data files, local data structures, etc. alone or in combination. The media may be those specially designed and constructed for the purposes of the present invention, or they may be of the kind well-known and available to those skilled in the computer software arts. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tape, optical recording media such as CD-ROMs, DVDs, magnetic-optical media such as floppy disks, and ROM, RAM, flash memory, and the like. Hardware devices specifically configured to store and execute program instructions are included. Examples of program instructions may include high-level language code that can be executed by a computer using an interpreter as well as machine code such as produced by a compiler.

Although the present invention has been described in detail with reference to exemplary embodiments above, those skilled in the art to which the present invention pertains can make various modifications to the above-described embodiments without departing from the scope of the present invention. Will understand. Therefore, the scope of the present invention should not be limited to the described embodiments, but should be defined by the claims below and equivalents thereof.

100: diagnosis system 102: database
104: data collection unit 106: data reduction unit
108: space partition unit 110: state inference unit

Claims (10)

A state diagnosis system that infers a state value according to an input sensing value by using a sensing value collected for a predetermined period and a state value according to the collected sensing value.
A data condensation unit for assigning a representative value to each of the sensing values, and condensing a plurality of cases in which the representative values of the sensing values are the same;
A space partitioning unit for disposing each case in a virtual space and dividing the virtual space into a plurality of regions; And
And a state inference unit configured to extract an area in which the input sensing value belongs to the virtual space, measure a similarity between the input sensing value and the cases belonging to the extracted area, and infer a state value according to the input sensing value. , Health diagnostic system.
The method of claim 1,
The data abbreviation unit,
And dividing the range of sensing values collected during the predetermined period into a plurality of sections, and assigning a predetermined representative value to the sensing values belonging to each section.
The method of claim 1,
The data abbreviation unit,
The plurality of cases having the same representative value are abbreviated as one case, and the state value of the abbreviated case is one of an average value, a median value, a maximum value, and a minimum value of the state values of the plurality of cases having the same representative value value. To set up, health diagnostic system.
The method of claim 3,
The state inference unit,
If the state value of the case most similar to the input sensing value among the cases belonging to the extracted area is the same as a preset state setting value, the state value of the case is inferred as a state value according to the input sensing value, Health diagnostic system.
The method of claim 3,
The state inference unit,
If the state value of the case most similar to the input sensing value among the cases belonging to the extracted area is different from the preset state setting value, the actual sensing value of the plurality of cases abbreviated as the case and the input sensing value Measuring a similarity and inferring a state value of a case most similar to the input sensing value to a state value according to the input sensing value.
A data reduction unit assigning a representative value to each of the sensing values stored in the database, and reducing the plurality of cases in which the representative values of the sensing values are the same;
A space partitioner disposing each case in the database into a virtual space, and dividing the virtual space into a plurality of regions; And
Extracting an area in which the sensing value inputted by the state inference unit belongs to the virtual space, and inferring a state value according to the input sensing value by measuring a similarity between the input sensing value and the cases belonging to the extracted area; Including a status diagnosis method.
The method of claim 6,
Condensing the plurality of cases,
Dividing the range of sensing values stored in the database into a plurality of sections;
The data contracting unit assigning preset representative values to sensing values belonging to each section;
Extracting, by the data abbreviation unit, a plurality of cases in which representative values of respective sensing values are the same; And
And contracting the plurality of cases extracted by the data contraction unit into a single case.
The method of claim 6,
Inferring the state value,
Extracting, by the state inference unit, a case most similar to the input sensing value among cases belonging to the extracted region;
Checking whether a state value of the case in which the state inference unit is extracted is different from a preset state setting value; And
And inferring the state value of the extracted case by a state value according to the input sensing value when the state value is equal to a preset state setting value.
The method of claim 8,
After checking whether the state value is different from the preset state setting value,
If the state value is different from a preset state setting value, measuring the similarity between the actual sensing value of the plurality of cases reduced to the extracted case and the input sensing value; And
And inferring, by the state inference unit, a state value of a case most similar to the input sensing value among the reduced plurality of cases to a state value according to the input sensing value.
Combined with hardware,
A computer-readable recording medium having recorded thereon a program for executing the condition diagnosis method according to any one of claims 6 to 9.

Images