JP6517991B1 - Abnormality sign diagnosis system, management device, and abnormality sign diagnosis method - Google Patents

Abstract

[Problem] To provide an abnormality prognostic diagnosis system etc. which appropriately diagnoses the presence or absence of an abnormality prognostic also at the time of reactivation of mechanical equipment. According to a first aspect of the present invention, there is provided, as a normal model used for diagnosing the mechanical equipment, the abnormality predictive diagnosis system 1 as another normal model that is the same model as the mechanical equipment at the time of re-operation after a stop period of the predetermined mechanical equipment. Normal model selecting means 162 for selecting one of the latest normal models, and diagnostic means 163 for diagnosing the presence or absence of an abnormality sign of the mechanical equipment based on the normal model selected by the normal model selecting means 162; Prepare. [Selected figure] Figure 2

Description

  BACKGROUND OF THE INVENTION Field of the Invention The present invention relates to an abnormality prognostic diagnosis system etc. for diagnosing the presence or absence of an abnormality prognostic of mechanical equipment.

  As a technique for diagnosing the presence or absence of abnormality signs of mechanical equipment, for example, the technique described in Patent Document 1 is known. That is, Patent Document 1 teaches a predetermined case model using multidimensional sensor data in normal operation of the machine and equipment, and based on this case model, diagnoses the presence or absence of an abnormality sign of the machine and equipment. It is described about.

Patent No. 4832609

  Usually, mechanical equipment such as power generation equipment is always operated, but in some cases, it may be stopped for a predetermined period. For example, when an abnormality (or abnormality sign) occurs in the mechanical equipment or when periodic maintenance is performed, the mechanical equipment is stopped for a predetermined period.

As described above, when re-operating after stopping the mechanical equipment, if the case model before stopping (also referred to as a normal model) is used as it is for abnormality sign diagnosis, there is a possibility that the diagnostic accuracy is lowered. This is because the case model used before the stop of the mechanical equipment and the actual state of the mechanical equipment at the time of reactivation deviate from each other as the environmental conditions change during the stop period.
Note that if the case model is re-learned from the beginning after restarting the mechanical equipment, abnormality sign diagnosis can not be performed during the learning period (for example, several weeks).

  Then, this invention makes it a subject to provide the abnormal sign diagnostic system etc. which diagnose the presence or absence of an abnormal sign appropriately also at the time of a restart of mechanical equipment.

In order to solve the above problems, according to the present invention, at the time of re-operation after a stop period of a predetermined mechanical equipment, another mechanical equipment which is the same model as the mechanical equipment as a normal model used for diagnosing the mechanical equipment. The normal model selecting means is used for the diagnosis of the mechanical equipment at the time of re-operation after the stop period of the predetermined mechanical equipment, selecting the normal model selecting means for selecting one of the latest normal models in the Among the other machine equipment of the same type as the machine equipment, the one having the normal model most similar to the normal model of the machine equipment just before entering the stop period as the normal model to be The latest normal model of the other identified mechanical equipment is selected as the normal model when the mechanical equipment is restarted .

  According to the present invention, it is possible to provide an abnormality prognostic diagnosis system or the like which appropriately diagnoses the presence or absence of an abnormality prognostic even when the mechanical equipment is restarted.

It is a functional block diagram of an abnormality sign diagnostic system concerning a 1st embodiment of the present invention. It is a functional block diagram of the data mining means with which an abnormal sign diagnostic system concerning a 1st embodiment of the present invention is provided. It is explanatory drawing of the normal model learned by the normal model learning part of the abnormal sign diagnostic system which concerns on 1st Embodiment of this invention. It is a flowchart which shows the process of the learning means with which the abnormal sign diagnostic system which concerns on 1st Embodiment of this invention is equipped. It is explanatory drawing regarding the transition of the normal model of the machinery installation and the abnormality sign flag in the abnormality sign diagnostic system which concerns on 1st Embodiment of this invention. It is a flowchart which shows the process of the normal model selection means with which the abnormal sign diagnostic system which concerns on 1st Embodiment of this invention is equipped. FIG. 7 is an explanatory view schematically showing a temporal change of each normal model of mechanical equipment in the abnormality prognostic diagnosis system according to the first embodiment of the present invention. It is a flowchart which shows the process of the diagnostic means with which the abnormal sign diagnostic system which concerns on 1st Embodiment of this invention is equipped. It is a functional block diagram of an abnormal sign diagnostic system concerning a 2nd embodiment of the present invention. It is a functional block diagram of the data mining means with which the management device of the abnormal sign diagnostic system concerning a 2nd embodiment of the present invention is provided. FIG. 7 is a functional block diagram of an abnormality prognostic diagnosis device connected to mechanical equipment in the abnormality prognostic diagnosis system according to the second embodiment of the present invention.

First Embodiment
FIG. 1 is a functional block diagram of the abnormal sign diagnostic system 1 according to the first embodiment.
The abnormality sign diagnosis system 1 is a system that diagnoses the presence or absence of an abnormality sign of the machine equipment 21 or the like based on a normal model indicating a normal range of operation information of the machine equipment 21 or the like. That is, in order to maintain and improve the operation rate of the mechanical devices 21 to 24 used in factories, commercial facilities, construction sites, hospitals, etc., the abnormality sign diagnosis system 1 Based on the degree of abnormality which is an index indicating the size, the abnormality prediction diagnosis of each of the mechanical equipment 21 to 24 is performed.
The above-mentioned “abnormal symptom diagnosis” is not limited to the diagnosis of whether or not the mechanical equipment 21 to 24 reaches an abnormal state in which it can not operate, and the mechanical equipment 21 can be operated in the normal state range. It is also included to diagnose the degree of performance degradation of H.24.
The abnormality sign diagnosis system 1 may be configured by one device, or may be configured by connecting a plurality of devices via a communication line.

  The four mechanical devices 21 to 24 are, for example, gas engine generators. In addition, the mechanical equipment 21-24 is not limited to this, A medical equipment, a communication installation, etc. other than a chemical plant and a nuclear power plant may be sufficient. Moreover, all the mechanical equipment 21-24 may be installed in the vicinity, and the thing from which the installation location is apart from others may exist.

  Although not shown in the drawings, a plurality of sensors are installed in each of the mechanical equipment 21-24. Then, through the network N, “operation information” including the identification information of the sensor, the identification information of the mechanical equipment (for example, the mechanical equipment 21) in which the sensor is installed, the detection value of the sensor, and the detection date and time It is transmitted to the abnormality sign diagnosis system 1. The identification information of each of the mechanical devices 21 to 24 also includes information indicating the model.

  In the present embodiment, as an example, a configuration in which four machine facilities 21 to 24 of the same model are connected to the abnormality symptom diagnosis system 1 via the network N will be described. The above-mentioned "same model" means that the model (for example, model) of the mechanical equipment 21-24 is the same. It should be noted that machine equipment having different structures but similar in structure and characteristics may be regarded as “same model” although they are different in type.

  A computer 3 illustrated in FIG. 1 is a computer that manages maintenance information (information related to periodic maintenance and inspection) of each of the mechanical facilities 21 to 24. In the computer 3, maintenance information including identification information of the mechanical equipment 21-24, a maintenance period, and a type of maintenance is stored in advance. Then, predetermined maintenance information stored in the computer 3 is transmitted to the abnormality symptom diagnosis system 1 via the network N.

<Configuration of abnormal sign diagnostic system>
As shown in FIG. 1, the abnormality sign diagnosis system 1 includes a communication unit 11, an operation information acquisition unit 12, an operation information storage unit 13, a maintenance information acquisition unit 14, and a maintenance information storage unit 15. There is. In addition to the configuration described above, the abnormality sign diagnosis system 1 further includes a data mining unit 16, a diagnosis result storage unit 17, a display control unit 18, and a display unit 19.

  The communication unit 11 receives predetermined operation information from the mechanical equipment 21 to 24 via the network N, and receives maintenance information on the mechanical equipment 21 to 24 from the computer 3. As such communication means 11, for example, a communication device that receives predetermined information based on a TCP / IP communication protocol can be used.

The operation information acquisition unit 12 acquires operation information (such as detection values of sensors) of the plurality of mechanical devices 21 to 24 via the network N or the like. This operation information is handled as time series data associated with the acquisition time (or the time measured by the sensor). The operation information acquisition unit 12 sequentially stores the acquired latest, that is, current operation information, in the operation information storage unit 13 so as to be accumulated.
The operation information storage unit 13 stores the operation information input from the operation information acquisition unit 12. A magnetic disk device, an optical disk device, a semiconductor storage device or the like can be used as such operation information storage means 13. The operation information stored in the operation information storage unit 13 is appropriately referred to by the data mining unit 16.

The maintenance information acquisition unit 14 acquires maintenance information of the mechanical equipment 21 to 24 via the communication unit 11.
The maintenance information storage unit 15 stores the maintenance information input from the maintenance information acquisition unit 14. As such maintenance information storage means 15, a magnetic disk device, an optical disk device, a semiconductor storage device or the like can be used. The maintenance information stored in the maintenance information storage unit 15 is appropriately referred to by the data mining unit 16.

  Although not shown, the data mining unit 16 is configured to include electronic circuits such as a central processing unit (CPU), a read only memory (ROM), a random access memory (RAM), and various interfaces. Then, the program stored in the ROM is read and expanded in the RAM, and the CPU executes various processing.

The data mining means 16 refers to the current and past operation information stored in the operation information storage means 13, and based on a statistical method (data mining) using this operation information as learning data, A predetermined normal model indicating a normal range of operation information is learned.
The data mining means 16 also has a function to individually diagnose the presence or absence of an abnormality sign of the mechanical equipment 21 to 24 based on the above-mentioned normal model. That is, the data mining means 16 calculates the degree of abnormality indicating the degree of abnormality of the diagnostic object information based on predetermined diagnostic object information (operation information to be diagnosed) and a normal model to which the diagnostic object information belongs. Do. And the data mining means 16 diagnoses individually the presence or absence of the abnormality sign of the mechanical equipment 21-24 based on this abnormality degree.
The details of the data mining means 16 and the normal model will be described later.

  The diagnosis result storage unit 17 stores the diagnosis result by the data mining unit 16. A magnetic disk device, an optical disk device, a semiconductor storage device or the like can be used as such a diagnostic result storage means 17.

  Although not shown, the display control means 18 is configured to include electronic circuits such as a CPU, a ROM, a RAM, and various interfaces. The display control means 18 generates a predetermined control signal for causing the display means 19 to display the diagnosis result of the data mining means 16. Further, the display control means 18 causes the display means 19 to display operation information and the like of the mechanical equipment 21 to 24 in accordance with the operation of the input means (not shown) by the administrator.

  The display unit 19 is, for example, a liquid crystal display, and displays the above-described diagnosis result and the like based on a control signal from the display control unit 18.

FIG. 2 is a functional block diagram of the data mining means 16 provided in the abnormal sign diagnostic system 1.
As shown in FIG. 2, the data mining unit 16 includes a learning unit 161, a normal model selection unit 162, and a diagnosis unit 163. Below, although the matter regarding the mechanical equipment 21 is mainly demonstrated among four machine equipment 21-24 (refer FIG. 1), the same may be said of the other mechanical equipment 22-24.

  The learning unit 161 performs processing for learning a normal model indicating a normal range of operation information based on operation information acquired when the plurality of machine units 21 to 24 are operating normally. To do. As shown in FIG. 3, the learning unit 161 includes a learning target information acquisition unit 161a, a learning target information storage unit 161b, a normal model learning unit 161c, and a normal model storage unit 161d.

  The learning target information acquisition unit 161 a acquires operation information (learning target information) to be learned from the operation information storage unit 13. For example, the learning target information acquisition unit 161a acquires, from the operation information storage unit 13, operation information of a past predetermined period (for example, for two weeks) in which it is known that the mechanical equipment 21 has normally operated.

  Further, although details will be described later, the learning target information acquisition unit 161a reads the operation information of the machine facility 21 that has been diagnosed as "no abnormality sign" by the diagnosis unit 163d from the diagnosis result storage unit 17, and learns this operation information. Add as a new target.

The learning target information storage unit 161 b stores, as a database, learning target information acquired by the learning target information acquisition unit 161 a.
The normal model learning unit 161c learns a normal model related to the mechanical equipment 21 based on the learning target information stored in the learning target information storage unit 161b.

FIG. 3 is an explanatory diagram of a normal model learned by the normal model learning unit 161c.
An axis α shown in FIG. 3 is an axis of a value obtained by normalizing a detection value of a predetermined sensor (for example, a pressure sensor: not shown) installed in the mechanical equipment 21. The axis β and the axis γ are axes of values obtained by normalizing detection values of another sensor (for example, a temperature sensor or a current sensor) installed in the mechanical equipment 21.

  The above-mentioned "normalization" is a process in which the detected values of the respective sensors are divided by a representative value (average value, standard deviation, etc.) or the like to be dimensionless and to be mutually comparable. Such normalized values are also included in the "operation information".

  Although FIG. 3 shows an example in which a normal model is learned in a three-dimensional (axis α, β, γ) vector space, actually, a normal model is learned in a multidimensional vector space of four or more dimensions. There are many. For example, when a normal model is learned based on detection values of 50 sensors installed in the mechanical equipment 21, a predetermined normal model is learned in a 50-dimensional multidimensional vector space.

  Each one of the ● marks shown in FIG. 3 corresponds to the instantaneous detection value of each sensor installed in the mechanical equipment 21. As described above, the state of the mechanical equipment 21 is represented by a position vector (hereinafter simply referred to as “vector”) whose component is the value obtained by normalizing the detection value of each sensor.

  The normal model learning unit 161c (see FIG. 2) learns a normal model related to the mechanical equipment 21 based on, for example, k-means which is one of non-hierarchical clustering. To explain the k-means method, first, the normal model learning unit 161c clusters each vector indicated by a ● mark, randomly assigns a normal model, and calculates coordinate values of the centers of gravity c of a plurality of vectors included in the normal model. .

  Next, the normal model learning unit 161c obtains the distance between the predetermined vector and each center of gravity c, and reassigns the vector to the normal model with the smallest distance. The normal model learning unit 161c executes such processing for all the vectors. Then, when the assignment of the normal model has not changed, the normal model learning unit 161c ends the generation of the normal model, and in the other cases, the normal model learning unit 161c recalculates based on the newly assigned normal model.

  After generating the normal model in this manner, the normal model learning unit 161c calculates the coordinate value of the center of gravity c of the normal model exhibiting a spherical shape and the radius r. The radius r is, for example, an average value of the distances between the vectors included in the normal model and the center of gravity c.

  The method of calculating the radius r is not limited to this. For example, among the vectors included in the normal model, the vector farthest from the center of gravity c may be specified, and the distance between the vector and the center of gravity c may be set as the radius r. Thus, the normal model learning unit 161c learns a normal model related to the mechanical equipment 21. Similarly, the normal model learning unit 161c individually learns normal models related to the other mechanical equipments 22 to 24.

  In the normal model storage unit 161 d shown in FIG. 2, normal models related to the mechanical equipment 21 to 24 are stored as databases. For example, with regard to the normal model regarding the mechanical equipment 21, at least the identification information of the mechanical equipment 21, the identification information of the normal model, the gravity center c and the radius r of the normal model are stored in the normal model storage unit 161d. The same applies to normal models of the other mechanical devices 22-24.

  The diagnostic means 163 shown in FIG. 2 has a function of diagnosing the presence or absence of an abnormality sign of the mechanical equipment 21 or the like based on the normal model selected by the normal model selection means 162. For example, when the operation information used for diagnosing the abnormality precursor of the mechanical equipment 21 is out of the normal model, the diagnosis unit 163 performs processing of diagnosing the mechanical equipment 21 as having the abnormality precursor for each mechanical equipment. As shown in FIG. 2, the diagnosis unit 163 includes a diagnosis target information acquisition unit 163a, a diagnosis target information storage unit 163b, an abnormality degree calculation unit 163c, and a diagnosis unit 163d.

The diagnosis target information acquisition unit 163 a acquires, from the operation information storage unit 13, operation information (diagnosis target information) used for abnormality sign diagnosis. For example, when performing abnormality symptom diagnosis of the mechanical equipment 21, the diagnosis target information acquisition unit 163 a detects the detection value of each sensor installed in the mechanical equipment 21 based on the identification information of the mechanical equipment 21 as the operation information storage unit 13. Get from
In the diagnosis target information storage unit 163b, operation information (diagnosis target information) of the machine facility 21 and the like acquired by the diagnosis target information acquisition unit 163a is stored as a database.

  For example, the abnormality degree calculating unit 163c reads the operation information used for the abnormality symptom diagnosis of the mechanical equipment 21 from the diagnosis target information storage unit 163b, and calculates the abnormality degree u of the operation information. First, the degree-of-abnormality calculation unit 163c normalizes (non-dimensional quantizes) the operation information used for abnormality sign diagnosis and generates a vector representing the state of the mechanical equipment 21 at a predetermined time.

  Then, the abnormality degree calculator 163c calculates the abnormality degree u of the operation information based on the predetermined normal model selected by the normal model selection means 162 described later. For example, the abnormality degree calculating unit 163c calculates a distance d (see FIG. 3) between the center of gravity c of the normal model related to the mechanical equipment 21 and the vector after normalization of the operation information.

  Furthermore, the abnormality degree calculator 163c calculates the abnormality degree u based on the distance d and the radius r of the normal model using the following equation (1). The degree of abnormality u is a numerical value indicating the degree to which the operation information deviates from the normal model.

  u = d / r (1)

  Then, the abnormality degree calculation unit 163c associates the calculated abnormality degree u with the identification information of the mechanical equipment 21, the identification information of the normal model, etc., and outputs the same to the diagnosis unit 163d, and stores it in the diagnosis result storage unit 17. Let

  The diagnosis unit 163d diagnoses, for example, the presence or absence of an abnormality sign of the mechanical equipment 21 based on the abnormality degree u calculated by the abnormality degree calculation unit 163c. That is, when the above-mentioned abnormality degree u ≦ 1, the vector (not shown) having the operation information after normalization as each component exists in the region of the normal model of the mechanical equipment 21. In such a case, the diagnosis unit 163d diagnoses the mechanical equipment 21 as "no abnormality precursor", and sets the abnormality precursor flag (see FIG. 5) of the mechanical equipment 21 at this date and time to "0."

  On the other hand, when the degree of abnormality u> 1, a vector (see FIG. 3) having the operation information after normalization as each component exists outside the region of the normal model of the mechanical equipment 21. That is, the operation information to be subjected to abnormality sign diagnosis is out of the normal model. In such a case, the diagnosis unit 163d diagnoses the mechanical equipment 21 as "prediction of abnormality" and sets the abnormality prediction flag (see FIG. 5) of the mechanical equipment 21 at this date and time to "1 '." Similarly, the diagnostic means 163 individually diagnoses the presence or absence of an abnormality sign of the other mechanical equipment 22-24.

  In this manner, the abnormality symptom diagnosis of the mechanical equipment 21 to 24 is performed, and the diagnosis result is stored in the diagnosis result storage unit 17. The above-mentioned diagnosis results include identification information of the mechanical equipment 21, etc., identification information of a normal model used for diagnosis, abnormality degree u, an abnormality sign flag, operation information (timely detected value of sensor), etc. It may be done. Then, such a diagnosis result is displayed on the display means 19 (see FIG. 1) by the display control means 18 (see FIG. 1).

  The normal model selection means 162 selects a normal model to be used for abnormality prediction diagnosis of the mechanical equipment 21 to 24, respectively, from the normal model storage unit 161d, and outputs the selected normal model to the abnormality degree calculation unit 163c. For example, when the mechanical equipment 21 is in operation and maintenance or the like is not performed, the normal model selecting unit 162 selects the latest normal model of the mechanical equipment 21 from the normal model storage unit 161 d. By this, based on the latest normal model, the abnormality symptom diagnosis regarding the mechanical equipment 21 can be performed appropriately. The details of the normal model selection means 162 will be described later.

  Next, learning of a normal model when the mechanical equipment 21 to 24 (see FIG. 1) is operating as usual will be described using FIG.

<Process of abnormal sign diagnostic system>
FIG. 4 is a flowchart showing the processing of the learning means 161 included in the abnormality sign diagnosis system 1 (see FIG. 2 as needed).
In step S101, for example, the learning unit 161 acquires and stores operation information of the mechanical equipment 21 (operation information acquisition step). That is, the learning unit 161 acquires operation information when it is known that the mechanical equipment 21 is normal by the learning target information acquisition unit 161a, and stores the acquired operation information in the learning target information storage unit 161b.

  In step S102, the learning unit 161 learns and updates a normal model (learning step). That is, the learning unit 161 learns a predetermined normal model (see FIG. 3) represented by the center of gravity c and the radius r by the normal model learning unit 161c based on the operation information of the mechanical equipment 21 acquired in step S101.

  In addition, when it is diagnosed by the diagnostic means 163 that “the abnormality does not occur in the mechanical equipment 21”, the learning means 161 uses the operational information used for the diagnosis as the operational information when the mechanical equipment 21 is operating normally. Add as new. Then, the learning unit 161 updates the normal model and sequentially learns based on the operation information after the addition (S102). As a result, changes in the environmental conditions of the mechanical equipment 21 and changes over time can be reflected in the normal model, which in turn can improve the accuracy of the abnormality sign diagnosis.

  In addition, when updating the normal model by adding the operation information diagnosed as “abnormal precursor” as a new learning target, the oldest operation information included in the normal model is deleted from the members of the normal model. You may As a result, even if the state of the mechanical equipment 21 gradually changes due to changes in environmental conditions or aging, it can be appropriately updated to a normal model of the mechanical equipment 21 by actively performing so-called metabolism.

  Next, in step S103, the learning means 161 stores the learning result. That is, the learning unit 161 causes the normal model storage unit 161 d to store the normal model regarding the mechanical equipment 21 together with the identification information of the mechanical equipment 21 and the identification information of the normal model. Similarly, the learning means 161 learns the normal models of the other mechanical equipments 22 to 24 individually.

  In addition, a series of processes shown in FIG. 4 are performed regularly (for example, once per day) for the mechanical equipment which continues to operate normally. On the other hand, with regard to mechanical equipment stopped due to maintenance or the like, new learning of normal models is often not performed because new operation information can not be obtained.

FIG. 5: is explanatory drawing regarding the transition of the normal model of the mechanical equipment 21-24, and an abnormality sign flag.
The first line of FIG. 5 shows the date and time when the abnormality precursor diagnosis was performed. T1, t2, t3, t4, ..., tN are described in order from the earliest date and time. For example, assuming that 9:00 am of October 1, 2017 is date and time t1, it is assumed that an abnormality symptom diagnosis is performed using operation information acquired once a day at 9:00 am. In this case, the time t2 is 9 am on October 2, 2017, and the time t3 is 9 am on October 3, 2017.

  In the example shown in FIG. 5, the center of gravity of the normal model at the time t1 of the mechanical equipment 21 is CC11 and the radius is RR11. The center of gravity "CC11" and the radius "RR11" are actually specific numerical values.

  After that, the normal model of the mechanical equipment 21 is sequentially updated at time t2, t3, t4, ..., tN. For example, the mechanical equipment 21 is updated to a normal model with a center of gravity CC12 and a radius RR12 at time t2. Further, the mechanical equipment 21 continues to operate normally at least until date and time t1 to tN, and it is assumed that the abnormality sign flags are all '0' at each date and time. The same applies to the mechanical equipments 22 and 24.

  In the example shown in FIG. 5, the normal model used for the N number of abnormal symptom diagnosis is stored in the normal model storage unit 161d (see FIG. 2). For example, although not shown in FIG. 5, when a normal model with date and time t (N + 1) is newly generated, the normal model with the oldest date and time t1 is deleted from the normal model storage unit 161d. As described above, in the present embodiment, N histories of normal models are left.

  Further, with regard to the mechanical equipment 23, the abnormality sign flag is '0' at the time t1, but the abnormality sign flag is '1' at the times t2 and t3. That is, at the time t2 and t3, the diagnosis unit 163d (see FIG. 2) diagnoses that “the mechanical equipment 23 has an abnormality sign”. Thereafter, based on the judgment of the administrator who has confirmed the result of such an abnormal sign, the mechanical equipment 23 is stopped until just before the date and time t4 to tN. A planned stop based on the result of such abnormal sign diagnosis is called "planned stop of mechanical equipment".

  As shown in FIG. 5, during the stop period of the mechanical equipment 23 (until immediately before the date and time t4 to tN), the abnormality symptom diagnosis of the mechanical equipment 23 is not performed. This is because it is not necessary to perform abnormality sign diagnosis during the stop period of the mechanical equipment 23. Therefore, for example, at the time t4 included in the stop period of the mechanical equipment 23, the abnormality prediction flag is neither '1' nor '0'. Thereafter, it is assumed that the mechanical equipment 23 is reactivated immediately before the time tN, and the abnormality symptom diagnosis of the mechanical equipment 23 is performed based on the operation information of the time tN.

  In such a case, the normal model selection means 162 (see FIG. 2) is another mechanical equipment 21 which is the same model as the mechanical equipment 23 as a normal model used for diagnosing abnormality signs after reactivation of the mechanical equipment 23. Select one of the 22 or 24 most recent normal models. This is one of the main features of the present embodiment. The selection of such a normal model will be described in detail using FIG.

FIG. 6 is a flowchart showing the processing of the normal model selection means 162 (see FIG. 2 as appropriate).
In addition, a series of processes shown in FIG. 6 are performed for every mechanical equipment as pre-processing of the abnormality sign diagnosis of the mechanical equipment 21-24. Although the process which selects the normal model of the mechanical equipment 23 is demonstrated below as an example, the same may be said of the other mechanical equipment 21,22,24.

  In step S201, the normal model selection means 162 determines whether or not the mechanical equipment 23 is reactivated after a predetermined stop period. In the case where the mechanical equipment 23 is not reactivated after the stop period (S201: No), the processing of the normal model selection means 162 proceeds to step S202.

In step S202, the normal model selection unit 162 selects the latest normal model as the normal model used for the abnormality symptom diagnosis of the mechanical equipment 23. As a result, it is possible to perform highly accurate abnormality symptom diagnosis on the basis of the normal model which is successively updated so as to follow the change of the environmental condition of the mechanical equipment 23.
On the other hand, when it is the re-operation after a stop period in Step S201 (S201: Yes), processing of normal model selection means 162 progresses to Step S203.

  In step S203, the normal model selection means 162 determines whether or not the mechanical equipment 23 is reactivated after the planned stop. That is, the normal model selection means 162 determines whether or not the mechanical equipment 23 is reactivated after being systematically stopped based on the diagnosis result of "prediction of abnormality".

  When it is re-operation after the planned stop in step S203 (S203: Yes), the processing of the normal model selection means 162 proceeds to step S204. In addition, it is determined based on the predetermined | prescribed signal accompanying the management's operation via an input means (not shown), for example, when carrying out a plan stop, whether it is re-operation after plan stop.

  In step S204, the normal model selection unit 162 selects the latest normal models of other mechanical equipment closest to the latest normal model whose abnormality prediction flag of the mechanical equipment 23 is '0' (normal model selection step). The step S204 will be described in more detail. When the mechanical equipment 23 is restarted after being stopped during the stop period based on the diagnosis result with abnormality sign by the diagnosis means 163 (S203: Yes), the normal model selecting means 162 performs the following processing.

  That is, the normal model selection means 162 refers to the diagnosis result in the past of the stop period of the mechanical equipment 23, and specifies a predetermined normal model at the latest date and time when the diagnosis means 163 has diagnosed no abnormality precursor. The normal model selecting means 162 is the closest model to the predetermined normal model of the mechanical equipment 23 among the other mechanical equipment 21, 22, 24 of the same type as the mechanical equipment 23, at the above-mentioned date and time ( Identify those that were similar. Furthermore, the normal model selection means 162 selects the latest normal model of the specified other mechanical equipment as a normal model of the mechanical equipment 23 at the time of reactivation. A specific example of the process of step S204 will be described in more detail with reference to FIG.

FIG. 7: is explanatory drawing which shows typically the time-dependent change of each normal model of the mechanical equipment 21-24.
The horizontal axis in FIG. 7 is date and time t1, t2,..., T (N-1), tN, t (N + 1),. The vertical axis in FIG. 7 indicates the barycentric position of the normal model. That is, in FIG. 7, the barycentric position of each normal model is represented only by the position in the vertical axis direction in a simplified manner.
Each circle shown in FIG. 7 represents a normal model. The normal model actually changes the center of gravity and the radius each time it is updated, but in FIG. 7, the radius of the normal model of the mechanical equipment 21 to 24 is made constant by simplification.

For example, at time t1, the normal model (see FIG. 5) of the mechanical equipment 21 having the center of gravity CC11 and the normal model of the mechanical equipment 24 having the center of gravity CC41 (see FIG. 5) are not similar. This is because the distance between the centers of gravity of normal models is relatively long.
On the other hand, at time t1, the normal model of the mechanical equipment 22 having the center of gravity CC21 (see FIG. 5) and the normal model of the mechanical equipment 23 having the center of gravity CC31 (see FIG. 5) are similar. This is because the distance between the centers of gravity of normal models is relatively short.

  At the time t2 shown in FIG. 7, since the operation information Q2 deviates from the normal model of the mechanical equipment 23 having the center of gravity CC32, the mechanical equipment 23 is diagnosed with "prediction of abnormality" and "1" is specified as an abnormality precursor flag at that time. It has been recorded (see FIG. 5). The same applies to time t3.

Further, since the operation information Q2 and Q3 of "presence of abnormality" is not used for updating the normal model of the mechanical installation 23, the coordinate value of the center of gravity of the mechanical installation 23 at the time t2 to t4 remains the center of gravity CC32. It has not changed.
Further, during the stop period of the mechanical equipment 23 at the time t4 to t (N-1), since the abnormality symptom diagnosis is not performed, a circular normal model is illustrated by a broken line.

  It is assumed that the mechanical equipment 23 is reactivated immediately before the time tN after a planned stop based on the diagnosis result with abnormality sign. In such a case, as described above, the normal model selecting unit 162 specifies the normal model (center of gravity CC31, radius RR31) at the latest time t1 at which the mechanical equipment 23 is diagnosed as "no abnormality precursor". As shown in FIG. 5, at the time t1, it is understood that the abnormality prediction flag of the mechanical equipment 23 is '0', and it has been diagnosed that “no abnormality prediction is present”.

  In the example shown in FIG. 7, at the time t1, the one closest to the center of gravity CC31 of the normal model of the mechanical equipment 23 is the center of gravity CC21 of the normal model of the mechanical equipment 22. Therefore, it can be said that the mechanical equipment 22 is closest to the state of the mechanical equipment 23 at the time t1 at which the mechanical equipment 23 was operating normally. Therefore, the normal model selecting unit 162 uses the latest normal model of the mechanical equipment 22 having the gravity center CC2N as the normal model at the time of reactivation of the mechanical equipment 23 at the time tN.

  That is, as indicated by the outlined arrows in FIG. 5, the normal model selection means 162 temporarily diverts the normal model of the mechanical equipment 22 as a normal model when the mechanical equipment 23 is reactivated. As described above, since the distance between the centers of gravity of the respective normal models is relatively short at the time t1 when both of the mechanical facilities 22 and 23 are operating normally, the mechanical facilities 22 and 23 are not It is likely that the states are similar.

  In the example shown in FIG. 7, at time tN, normal models of the mechanical equipments 22 and 23 having the center of gravity CC2N overlap each other. And, based on whether the operation information (after normalization) of the mechanical equipment 23 at the time tN is deviated from the normal model (refer to FIG. 5) of the gravity center CC2N and the radius RR2N, the presence or absence of abnormality sign of this mechanical equipment 23 Are diagnosed. Further, with regard to the mechanical equipment 22 as well, the abnormality sign diagnosis is performed based on the normal model of the center of gravity CC2N and the radius RR2N.

  As described above, in the present embodiment, when the mechanical equipment 23 is operating normally (time t1), a normal model of the mechanical equipment 22 having a normal model similar to itself is temporarily diverted, We are trying to make an abnormal sign diagnosis. By this, even immediately after the reactivation, the abnormality symptom diagnosis of the mechanical equipment 23 can be appropriately performed. In addition, even if the environmental conditions change significantly during the stop period, there is an advantage that it is not necessary to re-learn a normal model of the mechanical equipment 23 from the beginning.

  After the re-operation, the learning means 161 newly fetches the “abnormal precursor” diagnosis result of the mechanical equipment 23 and sequentially updates the normal model. As a result, after the time t (N + 1), the normal model of the mechanical equipment 23 gradually deviates from the normal model of the mechanical equipment 22. Thus, the actual state of the mechanical equipment 23 is gradually reflected in the normal model by sequentially updating the normal model. Therefore, the abnormality sign diagnosis of the mechanical equipment 23 can be performed with higher accuracy after the restart.

Again, returning to FIG. 6, the explanation will be continued.
When it is not the reoperation after the planned stop in step S203 (S203: No), the processing of the normal model selection means 162 proceeds to step S205.
In step S205, the normal model selection means 162 determines whether or not the mechanical equipment 23 is reactivated after the occurrence of an abnormality. That is, after the abnormality occurs in the mechanical equipment 23, the normal model selection means 162 determines whether or not it corresponds to the case of being reactivated after being stopped during a predetermined stop period. Although rare, there is also a possibility that an abnormality caused by some trouble may occur in the mechanical equipment 23.

  In addition, it is determined based on the signal accompanying the predetermined operation by the administrator via the input means (not shown), for example, when stopping the mechanical equipment 23 whether it is reactivation after occurrence of abnormality. . When it is the re-operation after the abnormality occurrence in step S205 (S205: Yes), the processing of the normal model selection means 162 proceeds to step S204. The process of this step S204 is the same as that at the time of reoperation after the planned stop (S203: Yes). That is, the normal model selection means 162 selects the latest normal models of other mechanical equipment closest to the nearest normal model whose abnormality prediction flag of the mechanical equipment 23 is '0'.

In addition, when the re-operation after the occurrence of an abnormality is not performed in step S205 (S205: No), the processing of the normal model selection unit 162 proceeds to step S206.
In step S206, the normal model selection means 162 determines that the mechanical equipment 23 is to be reactivated after maintenance. Note that whether maintenance of the mechanical equipment 23 has been performed is determined based on the maintenance information stored in the maintenance information storage unit 15 (see FIG. 2).

  In step S207, the normal model selection means 162 selects the latest normal models of other mechanical equipment closest to the normal model just before entering the stop period of the mechanical equipment 23 (normal model selection step).

  The normal model selection means 162 is another mechanical equipment 21 of the same type as the mechanical equipment 23 at the time of re-operation after maintenance performed during the stop period of the mechanical equipment 23. , 22 and 24, the one having the normal model that is closest to (similar to) the normal model of the mechanical equipment 23 immediately before entering the shutdown period is identified. Then, the normal model selection means 162 selects the latest normal model of the specified other mechanical equipment (for example, the mechanical equipment 22) as a normal model when the mechanical equipment 23 is reactivated.

  The process of step S204 will be described by taking a specific example. After the mechanical equipment 23 which has been operating until the predetermined date and time t3 is maintained during the subsequent stop period of the date and time t4 to t (N-1), the date and time tN Be reactivated.

  Here, the normal model selection means 162 reads the normal model of the mechanical equipment 23 at the time t3 immediately before entering the stop period. The normal model selecting means 162 is one in which the normal model at the time t3 is the most similar to the normal model of the mechanical equipment 23 (the distance between the centers of gravity is the shortest) among the other mechanical equipments 21, 22 and 24. Identify In the following, it is assumed that the machine equipment 22 is specified as such. In this case, the machine equipments 22 and 23 of the same model (type) are considered to be similar in their external and internal states, since the most recent normal models when they were operating normally are similar. Be

  Therefore, during maintenance of the mechanical equipment 23, even if environmental conditions change with time, by using the normal model of the mechanical equipment 22 successively updated so as to follow the change, even after reactivation, the machine An abnormality sign diagnosis of the equipment 23 can be appropriately performed.

  Thus, the normal model selection means 162 identifies one of the other mechanical facilities 22 having identified a model having a normal model closest (similar to) the normal model of the mechanical facility 23 before the stop period. The latest normal model is selected as a normal model when the mechanical equipment 23 is reactivated.

  Next, abnormality symptom diagnosis based on the normal model selected in step S202, S204 or S207 of FIG. 6 will be described using FIG.

FIG. 8 is a flowchart showing the processing of the diagnosis means 163 (see FIG. 2 as appropriate).
In step S301, the diagnosis unit 163 acquires and stores operation information of a predetermined mechanical equipment (for example, the mechanical equipment 23) (operation information acquisition step). That is, the diagnosis unit 163 causes the diagnosis target information acquisition unit 163a to acquire operation information (diagnosis target information) used for abnormality symptom diagnosis, and stores the acquired operation information in the diagnosis target information storage unit 163b.

  In step S302, the diagnosis unit 163 reads a predetermined normal model selected by the normal model selection unit 162. That is, the diagnosis unit 163 reads the predetermined normal model selected in step S202, S204 or S207 of FIG.

  In step S303, the diagnosis unit 163 calculates the degree of abnormality u. That is, the diagnosis unit 163 uses the normal model read in step S302 to calculate the degree of abnormality u of the operation information acquired in step S301. As described above, the degree of abnormality u is a numerical value indicating the degree to which the operation information deviates from the normal model.

  In step S304, the diagnosis unit 163 executes a diagnosis (diagnosis step). That is, based on the comparison between the degree of abnormality u calculated in step S303 and the predetermined threshold (for example, predetermined threshold = 1), the diagnosis unit 163 determines whether or not there is an abnormality sign of the predetermined mechanical equipment by the diagnosis unit 163d. To diagnose

  In step S305, the diagnosis unit 163 stores the diagnosis result. That is, the diagnosis unit 163 stores the diagnosis result of step S304 in the diagnosis result storage unit 17. This diagnosis result includes the detection value of each sensor and the degree of abnormality u in addition to the identification information of the predetermined mechanical equipment (for example, the mechanical equipment 23) to be diagnosed and the identification information of the normal model used for the diagnosis. It may be In addition, the process of step S301-S305 is performed separately about each of the mechanical equipment 21-24.

  In step S306, the diagnosis unit 163 displays the diagnosis result. That is, the diagnostic means 163 uses the display control means 18 (see FIG. 1) to identify the identification information of the mechanical equipment 21-24, the presence or absence of an abnormality sign, the degree of abnormality u, detected values of each sensor (time series graph), etc. It is displayed on the display means 19. After performing the process of step S306, the diagnosis unit 163 ends the series of processes related to the abnormality symptom diagnosis (END).

<Effect>
According to the first embodiment, for example, the normal model of one of the other types of mechanical equipment 21, 22, 24 is diverted even immediately after the mechanical equipment 23 is stopped for a predetermined period and then reactivated. By doing this, the abnormality symptom diagnosis of the mechanical equipment 23 can be properly performed.

  In addition, if the normal model immediately before stopping the mechanical equipment 23 is used as it is at the time of re-operation, there is a possibility that the normal model may deviate from the actual state of the re-operating mechanical equipment 23. This is because environmental conditions such as temperature and humidity change significantly while the mechanical equipment 23 is stopped. In addition, if the normal model is re-learned from the beginning after the mechanical equipment 23 is reactivated, the abnormal sign diagnosis based on the normal model can not be performed during the learning period (for example, several weeks).

On the other hand, according to the present embodiment, by using the latest normal model of the other mechanical equipment (for example, the mechanical equipment 22) having the normal model closest to the normal model before stopping the mechanical equipment 23. The abnormality sign diagnosis of the mechanical equipment 23 after the re-operation can be appropriately performed.
Furthermore, by sequentially updating the normal model, the original normal model of the reactivated mechanical equipment 23 can be gradually brought close to it, and an abnormality sign diagnosis with high accuracy can be performed.

Second Embodiment
In the second embodiment, a management device 10 for learning a normal model and the like, and a failure prediction diagnosis device 41 to 44 separately installed in the mechanical equipment 21 to 24, respectively, the failure prediction diagnosis system 1A (see FIG. 9). Is different from the first embodiment in that The other configuration is the same as that of the first embodiment. Therefore, only the parts different from the first embodiment will be described, and the descriptions of the overlapping parts will be omitted.

FIG. 9 is a functional block diagram of the abnormal sign diagnostic system 1A according to the second embodiment.
As shown in FIG. 9, the abnormality prognostic diagnosis system 1A includes a management device 10 and four abnormality prognostic diagnosis devices 41 to 44.
The management device 10 has functions of storing operation information of each of the machine facilities 21 to 24, learning of a normal model, selection of a normal model, and the like. As shown in FIG. 9, the management apparatus 10 includes a communication unit 11A, an operation information acquisition unit 12, an operation information storage unit 13, a maintenance information acquisition unit 14, and a maintenance information storage unit 15. In addition to the above-described configuration, the management apparatus 10 further includes a data mining unit 16, a diagnosis result storage unit 17, a display control unit 18, and a display unit 19.

The components other than the communication unit 11A and the data mining unit 16A are the same as those in the first embodiment, and thus the description thereof is omitted.
In addition to the function of the communication means 11 (refer to FIG. 1) described in the first embodiment, the communication means 11A determines the presence or absence of an abnormality sign of the predetermined mechanical equipment 21 from the normal model selected by the normal model selection means 162. It has a function to transmit to the abnormal symptom diagnostic device 41 to diagnose. The communication unit 11A also has a function of receiving the diagnosis result of the abnormality symptom diagnosis device 41 and storing the result in the diagnosis result storage unit 17.

FIG. 10 is a functional block diagram of data mining means 16A included in the management apparatus 10.
As shown in FIG. 10, the data mining unit 16A of the management device 10 includes a learning unit 161 and a normal model selection unit 162. The management apparatus 10 has a configuration in which the diagnostic means 163 is omitted from the data mining means 16 (see FIG. 2) described in the first embodiment. The configuration of the learning unit 161 is the same as that of the first embodiment, and thus will not be described.

  The normal model selection means 162 has a function of selecting a normal model to be used for diagnosing an abnormality sign of a predetermined mechanical equipment (for example, the mechanical equipment 21) and outputting the selected normal model to the communication means 11. The identification information of the target mechanical equipment 21 is added to the information on the normal model described above. And the communication means 11 is transmitted to the abnormality sign diagnostic apparatus 41 (refer FIG. 9) installed in the mechanical equipment 21 via the network N (refer FIG. 9). The process of the normal model selection unit 162 is the same as that of the first embodiment, and thus the detailed description is omitted.

FIG. 11 is a functional block diagram of the abnormality symptom diagnostic device 41 connected to the mechanical equipment 21. As shown in FIG.
The abnormality symptom diagnostic device 41 shown in FIG. 11 is connected to the mechanical equipment 21 via a communication line. As shown in FIG. 11, the abnormality predictor diagnostic apparatus 41 includes a communication unit 411, a normal model acquisition unit 412, and a normal model storage unit 413. In addition to the configuration described above, the abnormality symptom diagnosis device 41 further includes a diagnosis unit 163 and a diagnosis result storage unit 414.

  The communication unit 411 has a function of receiving, from the management apparatus 10 via the network N, a predetermined normal model used for diagnosing a failure sign of the mechanical equipment 21. The communication unit 411 also has a function of transmitting the diagnosis result of the mechanical equipment 21 stored in the diagnosis result storage unit 414 to the management apparatus 10 via the network N.

The normal model acquisition unit 412 acquires, through the communication unit 411, a normal model to be used for diagnosis of an abnormality sign of the mechanical equipment 21.
The normal model storage unit 413 stores a normal model acquired by the normal model acquisition unit 412 as a database.

The diagnosis unit 163 diagnoses the presence or absence of an abnormality sign of the mechanical equipment 21 based on the normal model stored in the normal model storage unit 413. The above-mentioned normal model is selected by the normal model selection means 162 (refer to FIG. 10) of the management device 10 for diagnosis of abnormality sign of the mechanical equipment 21. The configuration of the diagnosis unit 163 is the same as that described in the first embodiment (see FIG. 2), and thus the description thereof is omitted.
In the diagnosis result storage means 414, the results of abnormality sign diagnosis of the mechanical equipment 21 by the diagnosis means 163 are stored as a database. This diagnosis result is, as described above, communication means 411
And the network N sequentially to the management device 10. Then, it is displayed on the display means 19 (see FIG. 9) of the management device 10.

  In addition to the abnormality prognostic diagnosis device 42 installed in the mechanical equipment 22 (see FIG. 9) and the abnormality prognostic diagnosis device 43 installed in the mechanical equipment 23, the abnormality prognostic diagnosis device 44 installed in the mechanical equipment 24 is Since the configuration is the same as that described above, the description is omitted.

<Effect>
According to the second embodiment, in addition to the management device 10, the abnormality sign diagnosis devices 41 to 44 are individually installed in the respective mechanical facilities 21 to 24. For example, a large capacity operation information storage means 13 for storing operation information of the mechanical equipment 21 to 24 is provided in the management device 10, and a learning means 161 having a relatively large calculation load is also provided in the management device 10. On the other hand, by providing the diagnosis means 163 or the like having a relatively small calculation load in one-to-one correspondence with the mechanical equipment 21 to 24, the calculation load and storage capacity required for abnormality sign diagnosis can be distributed to each device.

«Modification»
As mentioned above, although abnormal sign diagnostic system 1 and 1A concerning the present invention was explained by each embodiment, the present invention is not limited to these descriptions, and various change can be made.
For example, in each embodiment, although the case where one normal model is learned regarding one mechanical equipment (for example, mechanical equipment 21) was explained, it is not limited to this. That is, a plurality of normal models having different center of gravity and the like may be learned with respect to one mechanical equipment.

  Therefore, when a plurality of normal models coexist with respect to the mechanical equipment 23 when the predetermined mechanical equipment 23 stop period is entered, the normal model selection means 162 may perform the following processing. . That is, at the time of the re-operation after the stop period of the mechanical equipment 23, the normal model selection means 162 is the closest (similar) normal to one normal model among the plurality of normal models regarding the mechanical equipment 23 before the stop period. The other mechanical equipment (for example, mechanical equipment 22) that had the model is identified. Then, the normal model selection means 162 selects the latest normal model of the specified other mechanical equipment. The normal model selection means 162 executes such processing for each of the plurality of normal models.

  For example, when there are three normal models of the mechanical equipment 23 at the time t3 immediately before the maintenance, one of them may be closest to the normal model Ma (not shown) of the mechanical equipment 21 at the time t3. Further, one of the remaining two may be closest to the normal model Mb (not shown) of the mechanical equipment 22 at the time t3 and the other may be closest to the normal model Mc (not shown) of the mechanical equipment 24. In such a case, the normal model selection means 162 selects, as a normal model to be used when the mechanical equipment 23 is reactivated, the latest normal model obtained by updating the above-described normal models Ma, Mb, and Mc. By this, even immediately after the re-operation of the mechanical equipment 23, its abnormality sign can be diagnosed with high accuracy. The same applies to the case where the mechanical equipment 23 is restarted after the planned stop, or the case where it is restarted after an abnormality occurs.

  Further, in the first embodiment, the abnormality symptom diagnosis of the four mechanical devices 21 to 24 (see FIG. 1) of the same type has been described, but the present invention is not limited to this. That is, in the case where different machine types are mixed among a plurality of machine equipments, the operation information acquired by the operation information acquiring means 12 (see FIG. 1) is an identification indicating the machine equipment types. Information shall be included. Then, the normal model selection means 162 refers to the above-mentioned identification information and, as a normal model to be used for diagnosis of the predetermined mechanical equipment to be reactivated after the stop period, other mechanical equipment of the same type as the mechanical equipment. Select one of the most recent normal models.

  In addition, the normal model selection means 162 selects any one of normal models of a plurality of other mechanical equipment of the same model without considering the closeness between the normal models (the distance between the centers of gravity). You may do it. Also, the normal model selection means 162 may select a normal model for which the values such as temperature and humidity around each mechanical equipment are closest. The same applies to the second embodiment.

  In addition, this invention is not limited to what has all the structures demonstrated by each embodiment. Also, part of the configuration of one embodiment can be replaced with the configuration of another embodiment, and the configuration of one embodiment can be added with the configuration of another embodiment. Moreover, it is also possible to add, delete, and replace other configurations for part of the configurations of the respective embodiments.

  Further, each configuration shown in FIG. 1, FIG. 2, and FIG. 9 to FIG. 11 may be realized by hardware by designing a part or all of them with an integrated circuit, for example. Further, each configuration described above may be realized by software by a processor interpreting and executing a program that realizes each function. Information to realize each function, such as a program, a tape, and a file can be stored in a memory, a hard disk, a recording device such as an SSD (Solid State Drive), or a recording medium such as an IC card, an SD card, or a DVD. .

  Further, control lines and information lines indicate what is considered to be necessary for the description, and not all control lines and information lines in the product are necessarily shown. In practice, almost all configurations may be considered to be mutually connected.

1, 1 A Abnormality Diagnosis System 3 Computer 10 Management Device 11, 11A Communication Means 12 Operation Information Acquisition Means 13 Operation Information Storage Means 14 Maintenance Information Acquisition Means 15 Maintenance Information Storage Means 16, 16A Data Mining Means 17 Diagnosis Result Storage Means 18 Display Control means 19 Display means 21, 22, 23, 24 Mechanical equipment 41, 42, 43, 44 Abnormal sign diagnostic device 161 Learning means 162 Normal model selection means 163 Diagnostic means N Network

Claims (9)

An abnormality sign diagnosis system for diagnosing the presence or absence of an abnormality sign of the machine facility based on a normal model indicating a normal range of operation information of the machine facility,
At the time of re-operation after a predetermined period of stoppage of the mechanical equipment, as the normal model used for diagnosing an abnormality sign of the mechanical equipment, the latest one of the other mechanical equipment that is the same model as the mechanical equipment. Normal model selection means for selecting the normal model of
And diagnostic means for diagnosing the presence or absence of an abnormality sign of the mechanical equipment based on the normal model selected by the normal model selection means ,
The normal model selecting means is, when the predetermined operation of the mechanical equipment is restarted after the stop period, another mechanical equipment which is the same model as the mechanical equipment as the normal model used for the diagnosis of the mechanical equipment. Among them, the one having the normal model most similar to the normal model of the mechanical equipment just before entering the stop period is identified, and the latest normal model of the other mechanical equipment identified is the one concerned. An abnormal sign diagnostic system characterized by selecting as said normal model at the time of mechanical equipment restart . The normal model selection means enters the stop period among the other machine equipment of the same type as the machine equipment at the time of re-operation after maintenance performed during the stop period of the predetermined machine equipment. The most recent normal model of the other mechanical equipment identified is identified by identifying the one that had the normal model most similar to the normal model of the mechanical equipment immediately before, and the identified one when the mechanical equipment is reactivated abnormal sign diagnosis system according to claim 1, characterized in that selected as the normal model. The normal model selecting means is for the machine equipment when the predetermined machine equipment is reactivated after being in a stopped state during the stop period based on the diagnosis result of the abnormality sign by the diagnosis means. Identify a predetermined normal model at the latest date and time when it is diagnosed that there is no abnormality precursor by the diagnostic means as the normal model of the mechanical equipment just before entering the stop period with reference to diagnosis results in the past before the stop period. Among the other machine equipment of the same type as the machine equipment, the normal machine at the date and time identified the one most similar to the predetermined normal model of the machine equipment, and the other machine identified The abnormal sign diagnostic system according to claim 1 , wherein the latest normal model of equipment is selected as the normal model of the mechanical equipment at the time of reactivation. When an abnormality occurs in the predetermined mechanical equipment, and it is reactivated after being in the stop state during the stop period, the normal model selection means determines the diagnosis result of the mechanical equipment in the past before the stop period. As the normal model of the machine facility just before entering the stop period, a predetermined normal model at the latest date and time when the diagnosis means is diagnosed as having no abnormality sign is specified, and the model is the same model as the machine facility. Among the other mechanical equipments, a normal model at the date and time identifies the one that is most similar to the predetermined normal model of the mechanical equipment, and the latest normal model of the other mechanical equipment that has been identified, The abnormality sign diagnostic system according to claim 1 , wherein the normal model is selected as the normal model of the mechanical equipment at the time of reactivation. When a plurality of normal models coexist with respect to the mechanical equipment when the predetermined period of time for the mechanical equipment enters the stop period, the normal model selection means is configured to restart the mechanical equipment after the stop period, Among the plurality of normal models related to the machine facility just before entering the stop period , the other machine facility having identified the normal model most similar to the one normal model is identified and identified The abnormal sign diagnostic system according to claim 1 , wherein the process of selecting the latest normal model of the mechanical equipment is executed for each of a plurality of normal models. The system further comprises learning means for learning the normal model based on the operation information acquired when the mechanical equipment is operating normally.
When it is diagnosed by the diagnostic means that there is no abnormality sign, the learning means newly adds the operation information used for the diagnosis as the operation information when the mechanical equipment is operating normally. The abnormal sign diagnostic system according to claim 1, wherein the normal model is updated based on the operation information after addition. In the case where different machine types are mixed among a plurality of the machine equipment, the operation information includes identification information indicating the machine type of the machine equipment,
The normal model selecting means refers to the identification information and, as the normal model used for the diagnosis of the predetermined mechanical equipment to be reactivated after a stop period, another mechanical equipment having the same model as the mechanical equipment. The abnormal sign diagnostic system according to claim 1, wherein one of the latest normal models is selected. A management apparatus that manages processing for diagnosing the presence or absence of an abnormality sign of the mechanical equipment based on a normal model indicating a normal range of operation information of the mechanical equipment,
At the time of re-operation after a predetermined period of stoppage of the mechanical equipment, as the normal model used for diagnosing an abnormality sign of the mechanical equipment, the latest one of the other mechanical equipment that is the same model as the mechanical equipment. Normal model selection means for selecting the normal model of
Communication means for transmitting the normal model selected by the normal model selection means to an abnormality indication diagnostic device for diagnosing the presence or absence of an abnormality indication of the mechanical equipment ;
The normal model selecting means is, when the predetermined operation of the mechanical equipment is restarted after the stop period, another mechanical equipment which is the same model as the mechanical equipment as the normal model used for the diagnosis of the mechanical equipment. Among them, the one having the normal model most similar to the normal model of the mechanical equipment just before entering the stop period is identified, and the latest normal model of the other mechanical equipment identified is the one concerned. A management apparatus characterized in that it is selected as the normal model at the time of restarting the mechanical equipment . An abnormality sign diagnosis method for diagnosing the presence or absence of an abnormality sign of the machine facility based on a normal model indicating a normal range of operation information of the machine facility,
At the time of re-operation after a predetermined period of stoppage of the mechanical equipment, as the normal model used for diagnosing an abnormality sign of the mechanical equipment, the latest one of the other mechanical equipment that is the same model as the mechanical equipment. A normal model selection step of selecting the normal model of
Wherein based on the normal model selected normal model selection step, seen containing a diagnostic step of diagnosing the presence or absence of an abnormality sign of the machinery, and
In the normal model selection step, at the time of restarting the predetermined mechanical equipment after the stop period, another mechanical equipment that is the same model as the mechanical equipment as the normal model used for the diagnosis of the mechanical equipment Among them, the one having the normal model most similar to the normal model of the mechanical equipment just before entering the stop period is identified, and the latest normal model of the other mechanical equipment identified is the one concerned. A method for diagnosing abnormality sign that is selected as the normal model at the time of reactivation of mechanical equipment .

Images