JP2018036916A - Facility state monitoring method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a facility state monitoring method which can improve the accuracy of abnormality detection.SOLUTION: In a state monitor 20, a frequency spectrum calculation part 22 calculates a spectrum at the time of operation by Fourier-transforming an electric signal in an operation state of a machine 11 acquired by an acquisition part 21. A vector calculation part 23 calculates a state feature vector on the basis of the spectrum at the time of operation calculated by the frequency spectrum calculation part 22. An angle calculation part 24 calculates an angle formed by a reference vector and the state feature vector calculated by the vector calculation part 23. The determination part 25 monitors the state of a piece of equipment 10 on the basis of the angle calculated by the angle calculation part 24 and a determination threshold.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a facility state monitoring method.

  Conventionally, a technique related to an abnormality detection / diagnosis method for facilities (including plants and equipment) has been disclosed (for example, Patent Document 1). In this equipment state monitoring method, sensor data obtained from a plurality of sensors attached to the equipment or operation data such as operation time and operation time is obtained, and almost normal data among the obtained sensor data and / or operation data. Is modeled as learning data. Then, an abnormality measure of the sensor data and operation data acquired using the modeled learning data is calculated as a vector, and an abnormality of the facility is detected based on the magnitude or angle of the calculated vector.

JP 2013-041448 A

  However, in the conventional equipment state monitoring method, emphasis is placed on the construction of a database for judging the state of equipment and the optimization of judgment thresholds, and it is put into statistical analysis by devising the database and judgment method. In addition, the acquired data is simplified. For this reason, there are problems that slight signs and changes of abnormality that are originally included in the electric signal are lost by filtering, and the accuracy of detecting the abnormality of the facility is lowered.

  The present invention has been made to solve such a problem, and an object of the present invention is to provide an equipment state monitoring method capable of improving the accuracy of abnormality detection.

  An equipment state monitoring method according to an aspect of the present invention is a state monitoring method for equipment including equipment, and obtains an electrical signal that changes with operation of the equipment from a sensor attached to the equipment, A first frequency spectrum is calculated by Fourier-transforming the electrical signal acquired in the operating state, and a peak value corresponding to each frequency in a predetermined frequency group in the calculated first frequency spectrum is calculated for each element. The first vector is calculated, and the state of the facility is monitored based on the second vector corresponding to the reference state of the device, the angle formed by the calculated first vector, and the threshold value.

  ADVANTAGE OF THE INVENTION According to this invention, the equipment state monitoring method which can improve the precision of abnormality detection can be provided.

It is a block diagram which shows an example of the state monitoring system which concerns on 1st Embodiment. It is a figure where it uses for description of the operation example of the state monitoring system of 1st Embodiment. It is a figure which shows the structure of the state monitoring system of an Example. It is a figure which shows an example of the waveform of the current signal obtained with the sensor connected to the power cable in the state monitoring system of an Example, and the frequency spectrum obtained from the said waveform. It is a figure which shows an example of transition of the monitoring parameter regarding each U phase, V phase, and W phase of No. 1 machine and No. 2 machine. It is a figure which shows an example of a display form.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

<First Embodiment>
<Outline of status monitoring system>
FIG. 1 is a block diagram showing an example of a state monitoring system according to the first embodiment of the present invention. In FIG. 1, the state monitoring system 1 includes a facility 10 and a state monitoring device 20. The facility 10 includes a device 11 to be monitored and a sensor 12 attached to the facility 10.

  The device 11 operates by receiving power supply from a power source (not shown). The device 11 may be, for example, an electric motor (motor).

  The sensor 12 senses (that is, acquires) an electrical signal that changes as the device 11 operates, and sends the sensed electrical signal to the state monitoring device 20. Here, the sensor 12 and the state monitoring device 20 are connected via an information transmission network regardless of, for example, wireless or wired. Then, the electrical signal sensed by the sensor 12 may be transmitted to the state monitoring device 20 via this information transmission network.

  The state monitoring device 20 monitors the state of the facility 10 using the electrical signal sensed by the sensor 12. Details will be described later.

<Configuration example of status monitoring device>
As illustrated in FIG. 1, the state monitoring device 20 includes an acquisition unit 21, a frequency spectrum calculation unit 22, a vector calculation unit 23, an angle calculation unit 24, a determination unit 25, and a storage unit 26.

  The acquisition unit 21 acquires the electrical signal sensed by the sensor 12 and outputs the acquired electrical signal to the frequency spectrum calculation unit 22.

  The frequency spectrum calculation unit 22 calculates a frequency spectrum by performing a Fourier transform on the electrical signal acquired by the acquisition unit 21. Here, hereinafter, the frequency spectrum calculated based on the electrical signal sensed in the operating state of the device 11 may be referred to as “operating spectrum (or first frequency spectrum)”.

  The vector calculation unit 23 calculates a “state feature vector” based on the frequency spectrum calculated by the frequency spectrum calculation unit 22. The “state feature vector” is a vector having each element as a peak value corresponding to each frequency in the “predetermined frequency group” in the operating spectrum calculated by the frequency spectrum calculation unit 22. In addition, the “predetermined frequency group” is a frequency group that is preferentially included in the electrical signal sensed in the operating state of the device 11 compared to other frequencies (that is, the electrical signal includes components of other frequencies). Compared to other frequency groups in the frequency spectrum, the frequency group has a peak value larger than that of other frequencies), and a frequency group in which a difference between normal and abnormal appears remarkably. Hereinafter, the “state feature vector” may be referred to as a first vector.

ａは、基準ベクトルであり、ｂは、状態特徴ベクトルである。 The angle calculation unit 24 calculates an angle formed by the “reference vector” and the “state feature vector” calculated by the vector calculation unit 23. The “reference vector” is stored in the storage unit 26. For example, the angle calculation unit 24 calculates the angle formed by the “reference vector” and the “state feature vector” by using the following equation (1).

a is a reference vector, and b is a state feature vector.

  Here, the “reference vector” may be, for example, a state feature vector calculated in advance based on an electrical signal sensed by the sensor 12 when the device 11 is operating normally, or the device 11 11 may be a state feature vector estimated to be obtained in a state where it is expected to operate normally. The above formula (1) is an example of a calculation method for calculating the angle formed by the “reference vector” and the “state feature vector”, and the calculation for calculating the angle formed by the “reference vector” and the “state feature vector”. The method is not limited to this.

  The determination unit 25 monitors the state of the equipment 10 based on the angle calculated by the angle calculation unit 24 and the “determination threshold value”. That is, here, an angle formed between the “reference vector” and the “state feature vector” calculated by the vector calculation unit 23 is used as the “state monitoring parameter”.

  For example, the determination unit 25 determines whether the device 11 is in a normal state or an abnormal state based on the magnitude of the angle calculated by the angle calculation unit 24 and the “abnormality determination threshold value”. Specifically, when the angle calculated by the angle calculation unit 24 is equal to or greater than the “abnormality determination threshold value”, the determination unit 25 greatly deviates from the “reference vector”. 11 or that an abnormality has occurred around the device 11. On the other hand, when the angle calculated by the angle calculation unit 24 is less than the “abnormality determination threshold value”, the deviation of the “state feature vector” from the “reference vector” is within an allowable range. Is determined to be normal.

  Here, based on the “caution state determination threshold value” smaller than the “abnormality determination threshold value”, the determination unit 25 determines whether or not the device 11 and the situation around the device 11 are normal but are in the “caution state”. May be determined. In this case, if the angle calculated by the angle calculation unit 24 is less than the “abnormality determination threshold” but is greater than or equal to the “caution state determination threshold”, the determination unit 25 is abnormal in the situation of the device 11 and the surroundings of the device 11 It is determined that the state is an “attention state” where there is a sign of “ The “abnormality determination threshold value” and the “abnormality determination threshold value” are stored in the storage unit 26.

  And the determination part 25 outputs a determination result to a display part (not shown). An example of the display form on the display unit will be described in detail later.

<Operation example of status monitoring system>
An operation example of the state monitoring system having the above configuration will be described. FIG. 2 is a diagram for explaining an operation example of the state monitoring system according to the first embodiment. Here, as an example, a case where the sensor 12 is connected to a power cable (not shown) that supplies current to the device 11 will be described.

  The sensor 12 senses a current signal (time waveform) flowing through a power cable (not shown) as shown in a graph 101 (horizontal axis: time (s), vertical axis: amplitude (V)) in FIG. The current signal sensed by the sensor 12 is acquired by the acquisition unit 21 of the state monitoring device 20.

Then, the frequency spectrum calculation unit 22 calculates a frequency spectrum using the current signal at time A acquired by the acquisition unit 21. Here, the time A is a time immediately after the device 11 starts to move, that is, a time when the device 11 is likely to be operating normally. Then, the vector calculation unit 23 uses, as each element, a peak value corresponding to each frequency in the “predetermined frequency group” in the frequency spectrum (graph 102 in FIG. 2) calculated based on the current signal at time A. A vector to be performed (that is, a state feature vector) is calculated. In the graph 102 of FIG. 2, the element frequencies of the “predetermined frequency element group” are “F, G, H, I, J”. The peak values of the element frequencies “F, G, H, I, J” are respectively “F ₀ (V), G ₀ (V), H ₀ (V), I ₀ (V), J ₀ ( V) ", which is a vector element of the state feature vector a at time A. That is, the state feature vector a at time A = (F ₀ , G ₀ , H ₀ , I ₀ , J ₀ ).
Here, the state feature vector a = (F ₀ , G ₀ , H ₀ , I ₀ , J ₀ ) at time A is used as the “reference vector”.

In addition, the frequency spectrum calculation unit 22 calculates a frequency spectrum using the current signal at time B later than time A acquired by the acquisition unit 21. Then, the vector calculation unit 23 uses the peak value corresponding to each frequency in the “predetermined frequency group” in the frequency spectrum (graph 103 in FIG. 2) calculated based on the current signal at time B as each element. A vector to be performed (that is, a state feature vector) is calculated. Also in the graph 103 of FIG. 2, the element frequencies of the “predetermined frequency group” are “F, G, H, I, J”. The peak values of the element frequencies “F, G, H, I, J” are respectively “F ₁ (V), G ₁ (V), H ₁ (V), I ₁ (V), J ₁ ( V) ", which is a vector element of the state feature vector b at time B. That is, the state feature vector b at time B = (F ₁ , G ₁ , H ₁ , I ₁ , J ₁ ).

  Then, the angle calculation unit 24 calculates an angle θ formed by the reference vector a and the state feature vector b as shown in the schematic diagram 104 of FIG. As shown in the schematic diagram 105 of FIG. 2, the angle θ is obtained by sequentially shifting the time B (for example, B = B + predetermined value), and the state feature vector b obtained based on the current signal at each time B. And the reference vector a. In the schematic diagram 105 of FIG. 2, the state feature vector b is a vector indicated by a dotted line.

  Then, the determination unit 25 determines whether the device 11 is in a normal state or an abnormal state based on the magnitude of the angle θ obtained based on the current signal at each time B and the “abnormality determination threshold”. . In this way, the state monitoring method is realized.

<Example>
Next, as an example, a case where the device 11 is a turbo compressor and the sensor 12 is connected to a power cable connecting the device 11 and the control panel will be specifically described.

  FIG. 3 is a diagram illustrating a configuration of the state monitoring system according to the embodiment. As shown in FIG. 3, the state monitoring system 1 according to the embodiment includes, for example, equipments 11-1 to 4, sensors 12-1 to 4, which are turbo compressors, and a control panel 15, and a state monitoring device. 20. Each of the devices 11-1 to 11-4 and the control panel 15 are connected by power cables L1, L2, L3, and L4, respectively. Sensors 12-1, 12-2, 12-3, and 12-4 are connected to the power cables L1, L2, L3, and L4, respectively. The sensors 12-1, 12-2, 12-3, 12-4 and the acquisition unit (data logger) 21 of the state monitoring device 20 are connected by a signal transmission line.

  For example, the current measurement was performed by the sensor 12-1 connected to the power cable L1 connecting the device 11-1 and the control panel 15, and the current signal shown in the left diagram of FIG. 4 was obtained. The left diagram of FIG. 4 shows current signal waveforms L111, L112, and L113 respectively corresponding to the U-phase, V-phase, and W-phase of the device 11-1 that is a turbo compressor. Here, although the current signal waveforms L111, L112, and L113 have different waveforms, if Fourier transform is performed to convert them into a frequency spectrum as shown in the right diagram of FIG. That is, the peak frequency (which corresponds to the frequency included in the above frequency group) is substantially the same. Therefore, the difference in the waveforms of the current signal waveforms L111, L112, and L113 is due to the difference in the balance (ratio) of the peak intensity of the peak frequency. These peak frequencies can be called “characteristic frequencies” because they represent the characteristics of the device 11-1 prominently.

  For this reason, a method of monitoring the balance (ratio) of the signal intensity at the characteristic frequency is appropriate as a method of monitoring the operating state of the device 11-1. This is because when the operating state of the device 11-1 changes, the current signal waveform is considered to change according to the change, and this is because the balance (ratio) of the peak values of the characteristic frequencies of the frequency spectrum has changed. It is because it corresponds. The change in the current signal waveform means that the operation of the device 11-1 or the ratio of the amount of current supplied to each part has changed.

  In this embodiment, in order to monitor this balance, as described above, a “state feature vector” is defined, and a state monitoring method using this “state feature vector” (ie, “vector monitoring method”) is employed. ing.

  Here, 16 frequencies having a relatively high peak intensity in the peak group in the frequency spectrum shown in the right diagram of FIG. 4, “50 Hz, 100 Hz, 150 Hz, 300 Hz, 400 Hz, 650 Hz, 1050 Hz, 1300 Hz, 1700 Hz, 1800 Hz, 2100 Hz. , 2400Hz, 2450Hz, 2800Hz, 3100Hz, 3400Hz "was selected as a" frequency group "that is a collection of characteristic frequencies. A vector whose element is the peak intensity of each feature frequency of the “frequency group” is a “state feature vector”. Here, the dimension of the “state feature vector” is 16 dimensions.

  As described above, when the operation state of the device 11-1 changes (for example, when the state of the supply current accompanying the operation state of the device 11-1 changes), the balance of the element group of the state feature vector changes. Since the direction of the state feature vector is different before and after the balance is changed, for example, the angle formed by the two vectors can be obtained by using the above equation (1). The larger the angle, the greater the change.

  Therefore, for example, an angle formed between the “reference vector” corresponding to the initial state of the device 11-1 (eg, the normal state immediately after overhaul) and the “state feature vector” corresponding to the current operation state of the device 11-1. , “Monitoring parameters”.

  Next, a method of calculating an angle that is a “monitoring parameter” will be specifically described. The calculation of the angle as the “monitoring parameter” is performed by the angle calculation unit 24 as described above.

  When the current signal obtained by measuring the U phase of the device 11-1 for a certain period (a certain 10 seconds) is Fourier transformed, the frequency groups “50Hz, 100Hz, 150Hz, 300Hz, 400Hz, 650Hz, 1050Hz, 1300Hz, 1700Hz, 1800Hz , 2100Hz, 2400Hz, 2450Hz, 2800Hz, 3100Hz, 3400Hz ", the peak height (that is, signal amplitude) is" 0.0603435 (V), 0.00298553 (V), 0.00447738 (V), 0.132506 (V), 0.0679784 (V ), 0.0375036 (V), 0.0228004 (V), 0.0150712 (V), 0.003894 (V), 0.00475024 (V), 0.00516008 (V), 0.0348993 (V), 0.00281858 (V), 0.00214451 (V), 0.00177439 (V) ), 0.000133124 (V) ”. Note that only the signal amplitude of 50 Hz is a value divided by 20 for adjusting the ratio. This vector is set as the reference vector a.

  State feature vectors other than the reference vector a obtained from the U-phase measurement signal of the device 11-1 are respectively state feature vectors b. For example, assume that one of the state feature vectors b is the next vector. b = (0.0424524, 0.000958385, 0.0030038, 0.0991667, 0.0549956, 0.0354294, 0.0272109, 0.0161954, 0.00641181, 0.00360079, 0.0036587, 0.0378731, 0.00398024, 0.00252317, 0.00212475, 0.0000968)

  If the reference vector a and the state feature vector b are used, the inner product value of the vector a and the vector b is “0.023085”. The absolute value of the vector a is “0.171361”, and the absolute value of the vector b is “0.135857”.

  When these are substituted into the above equation (1) and calculated, θ = 7.43 ° is obtained.

  In the same manner as the angle calculation method described above, the angle at each sample timing (the U phase, the V phase, and the W phase of each of the first machine (device 11-1) and the second machine (device 11-2) ( Monitoring parameters). FIG. 5 is a diagram illustrating an example of the transition of the monitoring parameters related to the U phase, V phase, and W phase of each of Unit 1 and Unit 2.

  In FIG. 5, starting from the left side, the transition of U phase (trend) of Unit 1 (device 11-1), transition of V phase, transition of W phase, transition of U phase of Unit 2 (device 11-2), V The transition of the phase and the transition of the W phase are shown.

  In FIG. 5, the threshold value TH11 is an “abnormality determination threshold value” related to the first device (device 11-1), and the threshold value TH12 is a “caution state determination threshold value” related to the first device (device 11-1). The threshold value TH21 is an “abnormality determination threshold value” related to the second machine (device 11-2), and the threshold value TH22 is a “caution state determination threshold value” related to the second machine (device 11-2).

  In FIG. 5, the transition of any monitoring parameter does not reach the “abnormality determination threshold value”, and the U-phase, V-phase, W-phase of Unit 1 (device 11-1), and Unit 2 (device 11-2) All of the U-phase, V-phase, and W-phase are in a normal state and are not in a caution state because they have not reached the “caution state determination threshold”. This determination is performed by the determination unit 25 as described above.

  Here, the reliable “abnormality determination threshold” and “caution state determination threshold” can be set by accumulating monitoring parameters at the time of abnormality. For example, the “abnormality determination threshold” may be set to 6 times the standard deviation (6σ) and the “attention state determination threshold” may be set to 3 times the standard deviation (3σ) until the monitoring parameters at the time of abnormality are sufficiently accumulated. . In addition, here, different values of “abnormality determination threshold” and different values of “caution state determination threshold” are set for each of the first machine (device 11-1) and the second machine (device 11-2). However, the present invention is not limited to this, and a common “abnormality determination threshold value” and “attention state determination threshold value” may be used.

  FIG. 6 is a diagram illustrating an example of a display form. As shown in FIG. 6, on the display unit connected to the state monitoring device 20, first, identification information of the first to fourth machines (that is, the devices 11-1 to 11-4) to be monitored is displayed. . For each of Units 1 to 4, the current angle value (monitor parameter value), the transition of the angle value (monitor parameter) over a certain period (trend graph), and status information (normal status, caution status, abnormal status) Status) is displayed. For example, the current angle value (monitoring parameter value), the transition of the angle value (monitoring parameter) over a certain period (trend graph), and the state information (normal state, caution state, abnormal state) are in a predetermined cycle. (For example, every 10 seconds).

  As described above, according to the first embodiment, in the state monitoring device 20, the frequency spectrum calculation unit 22 performs Fourier transform on the electrical signal in the operation state of the device 11 acquired by the acquisition unit 21. A spectrum (first frequency spectrum) is calculated. Then, the vector calculation unit 23 calculates a state feature vector (first vector) based on the operating spectrum calculated by the frequency spectrum calculation unit 22. Then, the angle calculation unit 24 calculates an angle formed between the reference vector (second vector) and the state feature vector calculated by the vector calculation unit 23. Then, the determination unit 25 monitors the state of the facility 10 based on the angle calculated by the angle calculation unit 24 and the determination threshold value.

  With the configuration of the state monitoring device 20, the component indicating the slight abnormal tendency and change included in the electrical signal in the operating state of the device 11 is not lost (filtered out) by the filtering, and the component of the state feature vector is not lost. It can be reflected as a component. And since the state of the installation 10 can be judged using the angle of such a state feature vector and a reference | standard vector, the precision of abnormality diagnosis and abnormality detection of the installation 10 can be improved.

<Other embodiments>
(1) In the first embodiment, as an example of the device 11 to be monitored, the device 11 is described as being an electric motor (motor) or a turbo compressor. However, the present invention is not limited to this. For example, the device 11 to be monitored may be a wide range of devices (equipment) using a high voltage cable such as a reciprocating compressor, a transformer, a refrigerator, or the high voltage cable itself.

  (2) The state monitoring device 20 described in the first embodiment can be realized with the following hardware configuration. That is, the state monitoring device 20 may have a network interface, a processor, and a memory. The acquisition unit 21 of the state monitoring device 20 described in the first embodiment may be realized by a network interface. In addition, the frequency spectrum calculation unit 22, the vector calculation unit 23, the angle calculation unit 24, and the determination unit 25 of the state monitoring device 20 described in the first embodiment read and execute a program stored in the memory by the processor. May be realized. The storage unit 26 may be realized by a memory.

  Note that the present invention is not limited to the above-described embodiment, and can be changed as appropriate without departing from the spirit of the present invention.

DESCRIPTION OF SYMBOLS 1 State monitoring system 10 Equipment 11 Equipment 12 Sensor 15 Control panel 20 State monitoring apparatus 21 Acquisition part 22 Frequency spectrum calculation part 23 Vector calculation part 24 Angle calculation part 25 Determination part 26 Storage part L1, L2, L3, L4 Power cable

Claims (1)

A state monitoring method for equipment including equipment,
Obtain an electrical signal that changes with the operation of the device from a sensor mounted on the facility,
By Fourier transforming the electrical signal acquired in the operating state of the device, the first frequency spectrum is calculated,
In the calculated first frequency spectrum, a first vector having a peak value corresponding to each frequency in a predetermined frequency group as each element is calculated,
Monitoring the state of the equipment based on the second vector corresponding to the reference state of the device, the angle formed by the calculated first vector, and a threshold;
Equipment condition monitoring method.

Images