JP7097338B2 - Equipment diagnostic equipment, equipment diagnostic methods, and equipment diagnostic programs - Google Patents

Description

The present invention relates to an equipment diagnostic device, an equipment diagnostic method, and an equipment diagnostic program.

In plants and factories, a large number of equipment, devices, and equipment (hereinafter, collectively referred to as "equipment") that are driven by rotating a rotating shaft using an electric motor as a power source are used. Examples of such equipment include motors, pumps, transport rollers, fans, blowers, reactors, stirrers and the like.

Diagnosis of such equipment is often performed by providing a plurality of sensors for measuring vibration and temperature and capturing changes in the state of bearings that support the rotation shaft. This is because deterioration or abnormality of equipment in which the rotating shaft is rotationally driven (for example, shaft system imbalance, shaft misalignment, bearing deterioration, etc.) causes changes in the bearings that support the rotating shaft.

The following Non-Patent Document 1 discloses various diagnostic techniques in steelmaking equipment. For example, a diagnostic technique using a self-generated wireless vibration sensor is disclosed. Further, there is disclosed a diagnostic technique that not only manages the absolute value of the current flowing through the drive motor but also improves the ability to detect equipment deterioration by frequency analysis.

Toyoki Yamamoto, 2 others, "Technical Papers Current Status and Future of Diagnostic Technology in Steelmaking Equipment", Nippon Steel & Sumitomo Metal Technical Report, No. 402, 2015

By the way, in the above-mentioned diagnostic technique for capturing a change in the state of a bearing that supports a rotating shaft, there is a problem that a large number of sensors for measuring vibration and temperature are required. Further, depending on the equipment, there is a problem that it may be difficult to install the sensor. It is considered that a large number of sensors will not be required if the diagnostic technique disclosed in Non-Patent Document 1 described above (technique for analyzing the current flowing through the drive motor) is used, but accurate diagnosis is performed unless the load fluctuation becomes large to some extent. There is a problem that it cannot be done.

The present invention has been made in view of the above circumstances, and provides an equipment diagnosis device, an equipment diagnosis method, and an equipment diagnosis program capable of accurately diagnosing equipment even if the load fluctuation is small, without requiring a large number of sensors. The purpose is to provide.

In order to solve the above problems, the equipment diagnostic apparatus according to one aspect of the present invention is an equipment diagnostic apparatus (1, 1A, 2) that diagnoses equipment having a rotary shaft (AX) that is rotationally driven by an induction motor (IM). 3), the current detection unit (10) for detecting the drive current supplied to the induction motor, the rotation detection unit (20, 20A) for detecting the rotation of the rotation shaft, and the current detection unit. Using the detection result and the detection result of the rotation detection unit, the calculation unit (30, 30A) for obtaining the load correlation value, which is a value having a correlation with the load resistance acting on the rotation axis, and the load correlation value are output. The output unit (40) is provided.

Further, in the equipment diagnostic apparatus according to one aspect of the present invention, the calculation unit uses the detection result of the current detection unit and the detection result of the rotation detection unit to obtain the slip ratio of the rotation shaft. (33), the effective value calculation unit (34) for obtaining the effective value of the drive current from the detection result of the current detection unit, the calculation result of the slip ratio calculation unit, and the calculation result of the effective value calculation unit. It is provided with a load correlation value calculation unit (35) for obtaining the load correlation value by using the device.

Here, in the equipment diagnostic apparatus according to one aspect of the present invention, the output unit outputs the slip ratio of the rotating shaft in addition to the load correlation value.

Further, in the equipment diagnosis device according to one aspect of the present invention, the calculation unit includes a frequency calculation unit (31) for obtaining the frequency of the drive current from the detection result of the current detection unit, and the output unit is the output unit. In addition to the load correlation value and the slip ratio of the rotating shaft, the frequency of the driving current and the rotating speed of the rotating shaft obtained from the detection result of the rotation detecting unit are output.

Further, in the equipment diagnostic apparatus according to one aspect of the present invention, the load correlation value is ST, the frequency of the drive current is ω [rad / s] or [Hz], and the rotation speed of the rotation shaft is ω _m [rad /]. Assuming that s] or [Hz], the effective value of the drive current is _Ia , and the slip ratio of the rotating shaft is s [%], the load correlation value calculation unit performs the calculation represented by the following equation (1). The load correlation value ST is obtained.

Further, in the equipment diagnostic apparatus according to one aspect of the present invention, at least the load correlation value output from the output unit is compared with the threshold value (TH) set in the load correlation value, and the equipment is used. A determination unit (50) for determining whether or not it is abnormal is provided.

Further, the equipment diagnostic apparatus according to one aspect of the present invention is said to be based on at least the load correlation value output from the output unit and the learning result of the load correlation value output from the output unit in the past. A diagnostic unit (60) for diagnosing the presence or absence of an abnormality in the equipment is provided.

Further, in the equipment diagnostic apparatus according to one aspect of the present invention, the diagnostic unit obtains the data of the load correlation value output from the output unit from the acquisition unit (61) and the data acquired by the acquisition unit. , A cutout section (62) that cuts out data based on predetermined conditions, a feature extraction section (63) that extracts a feature vector from the data cut out by the cutout section, and a feature extraction section that extracts the feature vector. The learning unit (64a) that analyzes the feature vector and generates a model of the feature vector based on the analysis result, and the degree of deviation between the feature vector and the model of the data newly acquired by the acquisition unit. It is provided with a deviation degree calculation unit (64c) for calculation.

The equipment diagnosis method according to one aspect of the present invention is an equipment diagnosis method for diagnosing equipment having a rotating shaft (AX) rotationally driven by an induction motor (IM), and a drive current supplied to the induction motor is used. Using the current detection step (S11) to be detected, the rotation detection step (S14) to detect the rotation of the rotation shaft, the detection result of the current detection step, and the detection result of the rotation detection step, the rotation shaft can be set. It has a calculation step (S17) for obtaining a load correlation value which is a value having a correlation with the acting load resistance, and an output step (S18) for outputting the load correlation value.

The equipment diagnosis program according to one aspect of the present invention functions as an equipment diagnosis device (1, 1A, 2, 3) for diagnosing equipment having a rotation shaft (AX) rotationally driven by an induction motor (IM). This is an equipment diagnosis program that causes the computer to detect the detection result of the current detection unit (10) that detects the drive current supplied to the induction motor and the rotation detection unit (20, 20A) that detects the rotation of the rotating shaft. A calculation means (30, 30A) for obtaining a load correlation value which is a value having a correlation with the load resistance acting on the rotation axis, and an output means (40) for outputting the load correlation value. , And make it work.

According to the present invention, there is an effect that the equipment can be diagnosed with high accuracy even if the load fluctuation is small without requiring a large amount of sensors.

It is a block diagram which shows the main part structure of the equipment diagnostic apparatus according to 1st Embodiment of this invention. It is a flowchart which shows the equipment diagnosis method by 1st Embodiment of this invention. It is a figure which shows the display example of the load correlation value and the like in 1st Embodiment of this invention. It is a block diagram which shows the main part structure of the modification of the equipment diagnostic apparatus according to 1st Embodiment of this invention. It is a block diagram which shows the main part structure of the equipment diagnostic apparatus according to 2nd Embodiment of this invention. It is a block diagram which shows the main part structure of the equipment diagnostic apparatus according to the 3rd Embodiment of this invention. It is a block diagram which shows the internal structure of the diagnostic part in 3rd Embodiment of this invention. It is a flowchart which shows the operation example of the diagnostic unit in 3rd Embodiment of this invention.

Hereinafter, the equipment diagnosis device, the equipment diagnosis method, and the equipment diagnosis program according to the embodiment of the present invention will be described in detail with reference to the drawings.

[First Embodiment]
<Equipment diagnostic equipment>
FIG. 1 is a block diagram showing a main configuration of an equipment diagnostic device according to the first embodiment of the present invention. As shown in FIG. 1, the equipment diagnostic apparatus 1 of the present embodiment includes a current transformer 10 (current detection unit), an encoder 20 (rotation detection unit), a calculation unit 30 (calculation means), and an output unit 40 (output means). To prepare for. The equipment diagnostic device 1 having such a configuration diagnoses equipment having a rotary shaft AX that is rotationally driven by an induction motor IM.

Here, the induction motor IM includes a stator having a coil and a rotor having a cage-shaped structure, for example, and the rotor is rotated by a rotating magnetic field formed by the coil of the stator. The induction motor IM may be driven by a single-phase alternating current or may be driven by a three-phase alternating current. In the present embodiment, it is assumed that the induction motor IM is driven by a three-phase alternating current.

The rotary shaft AX is, for example, a columnar (rod-shaped) member, and is rotationally driven by the rotation of the rotor of the induction motor IM. The rotary shaft AX may be connected to the rotor of the induction motor IM via a speed reducer having a predetermined reduction ratio. When such a speed reducer is provided, for example, when a drive current of a predetermined frequency (for example, 10 [Hz]) is supplied to the induction motor IM, the rotary shaft AX is in a state of no load. It rotates at a predetermined rotation speed (for example, 100 [rpm]). The rotation shaft AX may be directly attached to the rotor of the induction motor IM so as to be coaxial with the rotor.

The current transformer 10 detects the drive current supplied to the induction motor IM. The current transformer 10 may detect all phases (three phases) of the drive current supplied to the induction motor IM, or may detect only a specific one phase. In the present embodiment, it is assumed that the current transformer 10 detects only a specific one of the three phases. The detection result of the current transformer 10 is output to the calculation unit 30.

The encoder 20 detects the rotation of the rotation axis AX. Specifically, the encoder 20 detects the amount of rotation (rotational position) of the rotation axis AX, and outputs a number of pulses according to the detection result. The encoder 20 may be a mechanical rotary encoder or an optical rotary encoder. The detection result of the encoder 20 is output to the calculation unit 30.

The calculation unit 30 includes a frequency calculation unit 31, a rotation speed calculation unit 32, a slip rate calculation unit 33, an effective value calculation unit 34, and an ST calculation unit 35 (load correlation value calculation unit). Using the detection result of the current transformer 10 and the detection result of the encoder 20, the calculation unit 30 having such a configuration has a load correlation which is a value having a correlation with the slip coefficient of the rotation axis AX and the load resistance acting on the rotation axis AX. Obtain the value ST (sliding torque coefficient) and the like.

The frequency calculation unit 31 obtains the frequency ω [rad / s] or [Hz] of the drive current supplied to the induction motor IM from the detection result of the current transformer 10. For example, the frequency calculation unit 31 obtains the frequency ω of the drive current supplied to the induction motor IM each time the rotation axis AX rotates by a predetermined number of rotations N (N is an integer of 1 or more). May be. The timing and period for obtaining the frequency ω of the drive current supplied to the induction motor IM by the frequency calculation unit 31 can be arbitrarily set.

The rotation speed calculation unit 32 obtains the rotation speed ω _m [rad / s] or [Hz] of the rotation axis AX from the detection result of the encoder 20. For example, the rotation speed calculation unit 32 may obtain the rotation speed ω _m of the rotation axis AX every time a predetermined time (for example, 1 [s]) elapses. The timing and period for the rotation speed calculation unit 32 to obtain the rotation speed ω _m of the rotation axis AX can be arbitrarily set.

The slip rate calculation unit 33 obtains the slip rate of the rotation axis AX by using the detection result of the current transformer 10 and the detection result of the encoder 20. Specifically, the slip ratio calculation unit 33 is a rotation speed calculation unit 32 using the frequency ω of the drive current obtained by the frequency calculation unit 31 using the detection result of the current transformer 10 and the detection result of the encoder 20. The slip ratio of the rotation axis AX is obtained by using the obtained rotation speed ω _m of the rotation axis AX. More specifically, the slip rate calculation unit 33 performs the calculation shown in the following equation (2) to obtain the slip rate s [%] of the rotation axis AX.

The effective value calculation unit 34 obtains the effective value I _a [A] of the drive current supplied to the induction motor IM from the detection result of the current transformer 10. For example, the effective value calculation unit 34 may obtain the effective value _Ia of the drive current each time the rotation axis AX rotates by a predetermined rotation speed N, similarly to the frequency calculation unit 31. The timing and period for the effective value calculation unit 34 to obtain the effective value _Ia of the drive current can be arbitrarily set.

The ST calculation unit 35 obtains the load correlation value ST by using the calculation result of the slip rate calculation unit 33 and the calculation result of the effective value calculation unit 34. Specifically, the ST calculation unit 35 is obtained by the frequency ω of the drive current obtained by the frequency calculation unit 31, the rotation speed ω _m of the rotation axis AX obtained by the rotation speed calculation unit 32, and the effective value calculation unit 34. Using the effective value I _a of the drive current and the slip rate s of the rotation axis AX obtained by the slip rate calculation unit 33, the load correlation value ST is obtained by performing the calculation shown in the following equation (3).

Here, consider a case where the model of the induction motor IM in the stationary coordinate system (αβ coordinate system) is converted into the dq coordinate system. Assuming that the number of poles of the induction motor IM is P, the mutual inductance between the stator and the rotor is M, the maximum value of the current on the _d -axis is Id, and the DC resistance value of the rotor is _Rr , the induction motor IM is in steady state. The generated torque τ _e in is expressed by the following equation (4).

Further, assuming that the braking coefficient (= load) is R _m and the Coulomb friction is T _l , the generated torque τ _e in the steady state of the induction motor IM is expressed by the following equation (5).

The following equation (6) can be obtained from the above equations (4) and (5).

Now, assuming that the mutual inductance M between the stator and the rotor, the DC resistance value R _r of the rotor, and the Coulomb friction T _l are constant, the following two conditions are assumed.
・ The rate of change of the current on the d-axis is almost the same and is proportional to the effective value _Ia of the drive current. ・ The rotation speed ω _m of the rotation axis AX does not change significantly. It is expressed by an expression.

From the above equation (7), the braking coefficient R _m indicating the load is a value obtained by dividing the frequency ω of the drive current by the rotation speed ω _m of the rotation axis AX, the square of the effective value I _a of the drive current, and It can be seen that it is proportional to the product of the slip ratio s of the rotation axis AX. Therefore, the load correlation value ST, which is a value having a correlation with the load resistance acting on the rotation axis AX, can be expressed as the above equation (3).

The output unit 40 outputs various values obtained by the calculation unit 30 to the outside. The output unit 40 outputs at least the load correlation value ST obtained by the ST calculation unit 35 to the outside. In addition to the load correlation value ST, the output unit 40 may output the slip rate s of the rotation axis AX obtained by the slip rate calculation unit 33. In the output unit 40, in addition to the load correlation value ST and the slip ratio s of the rotation axis AX, the frequency ω of the drive current obtained by the frequency calculation unit 31 and the rotation of the rotation axis AX obtained by the rotation speed calculation unit 32. The speed ω _m may be output.

The output unit 40 may output the load correlation value ST, the slip ratio s of the rotation axis AX, the frequency ω of the drive current, and the rotation speed ω _m of the rotation axis AX as data. For example, the output unit 40 may perform wired communication or wireless communication with an external device (for example, a data collection device) and output the above data to the external device. Further, the output unit 40 may output by writing the above data to the attached external memory.

Alternatively, the output unit 40 displays the values of the load correlation value ST, the slip ratio s of the rotation axis AX, the frequency ω of the drive current, and the rotation speed ω _m of the rotation axis AX on a display device such as a liquid crystal display device. It may be output by. The output unit 40 may numerically display the above-mentioned load correlation value ST, the slip ratio s of the rotation axis AX, the frequency ω of the drive current, and the rotation speed ω _m of the rotation axis AX, and display the change with time. It may be displayed in the graph shown. The output unit 40 may be a display unit.

The functions of the equipment diagnostic apparatus 1 (functions of the calculation unit 30 and the output unit 40) may be realized by software, for example, by a computer reading and installing a program recorded on a recording medium. Alternatively, the computer may be realized by software by installing a program downloaded via a network (not shown). Alternatively, it may be realized by using hardware such as FPGA (Field-Programmable Gate Array), LSI (Large Scale Integration), and ASIC (Application Specific Integrated Circuit).

<Equipment diagnosis method>
FIG. 2 is a flowchart showing an equipment diagnosis method according to the first embodiment of the present invention. The flowchart shown in FIG. 2 is repeated, for example, at a preset constant cycle. Here, it is assumed that a drive current is supplied to the induction motor IM from a drive device (not shown), and the rotary shaft AX is rotationally driven by the induction motor IM.

When the processing of the flowchart shown in FIG. 2 is started, first, the current transformer 10 detects the drive current supplied to the induction motor IM (step S11: current detection step). The drive current detected by the current transformer 10 is output to the calculation unit 30. Then, the frequency calculation unit 31 performs a process of obtaining the frequency ω of the drive current supplied to the induction motor IM from the detection result of the current transformer 10 (step S12). Further, the effective value calculation unit 34 performs a process of obtaining the effective value _Ia of the drive current supplied to the induction motor IM from the detection result of the current transformer 10 (step S13).

Next, the encoder 20 detects the rotation of the rotation axis AX (step S14: rotation detection step). The detection result of the encoder 20 is output to the calculation unit 30. Then, the rotation speed calculation unit 32 performs a process of obtaining the rotation speed ω _m of the rotation axis AX from the detection result of the encoder 20 (step S15).

Next, a process of obtaining the slip ratio of the rotating shaft AX using the frequency ω of the drive current obtained by the frequency calculation unit 31 and the rotation speed ω _m of the rotating shaft AX obtained by the rotating speed calculation unit 32 is performed. This is performed by the calculation unit 33 (step S16). Specifically, the slip rate calculation unit 33 performs a process of obtaining the slip rate s [%] of the rotation axis AX by performing the calculation shown in the above-mentioned equation (2).

Subsequently, the frequency ω of the drive current obtained by the frequency calculation unit 31, the rotation speed ω _m of the rotation axis AX obtained by the rotation speed calculation unit 32, and the effective value I of the drive current obtained by the effective value calculation unit 34. Using the slip rate s of the rotation axis AX obtained by _a and the slip rate calculation unit 33, the ST calculation unit 35 performs a process of obtaining the load correlation value ST (step S17: calculation step). Specifically, the ST calculation unit 35 performs a process of obtaining the load correlation value ST by performing the calculation shown in the above-mentioned equation (3).

When the above processing is completed, the output unit 40 performs a process of outputting the load correlation value ST or the like obtained by the ST calculation unit 35 (step S18: output step). Specifically, the load correlation value ST obtained by the ST calculation unit 35, the slip rate s of the rotation axis AX obtained by the slip rate calculation unit 33, the frequency ω of the drive current obtained by the frequency calculation unit 31, and the frequency ω of the drive current obtained by the frequency calculation unit 31. The output unit 40 performs a process of outputting the rotation speed ω _m of the rotation axis AX obtained by the rotation speed calculation unit 32.

In the flowchart shown in FIG. 2, for convenience of explanation, an example in which the processes of steps S14 and S15 are performed after the processes of steps S11 to S13 are completed is shown. However, the processes of steps S14 and S15 may be performed before the processes of steps S11 to S13, or may be performed in parallel with the processes of steps S11 to S13.

FIG. 3 is a diagram showing a display example of a load correlation value or the like in the first embodiment of the present invention. In the example shown in FIG. 3, a graph showing the time-dependent change of the load correlation value ST, a graph showing the time-dependent change of the slip rate s of the rotation axis AX, a graph showing the time-dependent change of the drive current frequency ω, and the rotation of the rotation axis AX. A total of four types of graphs showing the time course of the velocity ω _m are shown.

Note that FIG. 3A shows a case where the load resistance acting on the rotating shaft AX is almost zero. 3 (b) to 3 (d) show a case where a load resistance acts on the rotary shaft AX by applying a load to the rotary shaft AX. Specifically, FIG. 3B shows a case where a tool (for example, a spanner) is strongly pressed against the rotary shaft AX, and FIG. 3C shows a case where the rotary shaft AX is grasped by hand. 3 (d) shows the case where the tool is lightly brought into contact with the rotary shaft AX.

First, referring to FIG. 3A, the load correlation value ST, the slip ratio s of the rotation axis AX, the frequency ω of the drive current, and the rotation speed ω _m of the rotation axis AX are almost the same, although there are some fluctuations due to noise. It turns out to be constant. Next, referring to FIG. 3B, the load correlation value ST, the slip ratio s of the rotary shaft AX, and the rotational speed ω _m of the rotary shaft AX are large in the period T1 in which the tool is strongly pressed against the rotary shaft AX. You can see that it is changing.

Next, referring to FIG. 3 (c), it can be seen that the frequency ω of the drive current changes significantly over the entire period T2 in which the rotation axis AX is grasped by hand. Further, in FIG. 3C, at the end of the period T2 (when the hand is released from the rotation axis AX), the load correlation value ST, the slip ratio s of the rotation axis AX, and the rotation speed ω _m of the rotation axis AX are shown. You can also see the change in. Subsequently, referring to FIG. 3D, the load correlation value ST, the slip ratio s of the rotation axis AX, and the rotation speed ω _m of the rotation axis AX change during the period T3 in which the tool is lightly contacted with the rotation axis AX. Can be seen to appear.

As described above, although there are differences depending on the nature of the load applied to the rotary shaft AX (load resistance acting on the rotary shaft AX), the load correlation value ST, the slip ratio s of the rotary shaft AX, the frequency ω of the drive current, and the rotation The rotation speed ω _m of the shaft AX changes according to the load applied to the rotation shaft AX. In particular, even when the tool is lightly brought into contact with the rotary shaft AX, the load correlation value ST, the slip ratio s of the rotary shaft AX, the frequency ω of the drive current, and the rotational speed ω _m of the rotary shaft AX change.

Therefore, by outputting (displaying) the load correlation value ST, the slip ratio s of the rotation axis AX, the frequency ω of the drive current, and the rotation speed ω _m of the rotation axis AX, the user (for example, the plant manager) can output (display). It is possible to diagnose deterioration or abnormality of equipment in which the rotary shaft AX is rotationally driven (for example, shaft system imbalance, shaft misalignment, bearing deterioration, etc.). As described above, in the present embodiment, a large amount of sensors such as a vibration sensor and a temperature sensor are not required, and the equipment can be diagnosed with high accuracy even if the load fluctuation is small.

In the diagnosis of the equipment, it is not always necessary to consider all of the above load correlation value ST, the slip ratio s of the rotating shaft AX, the frequency ω of the drive current, and the rotating speed ω _m of the rotating shaft AX. Considering at least one of the above load correlation value ST, the slip ratio s of the rotation axis AX, the frequency ω of the drive current, and the rotation speed ω _m of the rotation axis AX (for example, only the load correlation value ST). The equipment may be diagnosed.

<Modification example>
FIG. 4 is a block diagram showing a main configuration of a modified example of the equipment diagnostic apparatus according to the first embodiment of the present invention. As shown in FIG. 4, the equipment diagnosis device 1A according to the present modification has a configuration in which the encoder 20 of the equipment diagnosis device 1 shown in FIG. 1 is replaced with a tachometer 20A (rotation detection unit) and the calculation unit 30 is replaced with a calculation unit 30A. Is. The tachometer 20A detects the rotation speed ω _m [rad / s] or [Hz] of the rotation axis AX. The tachometer 20A may be a mechanical type or an electric type.

The calculation unit 30A has a configuration in which the rotation speed calculation unit 32 of the calculation unit 30 shown in FIG. 1 is omitted, and the detection result of the tachometer 20A is directly input to the slip rate calculation unit 33 and the output unit 40 of the calculation unit 30A. That is, since the tacometer 20A can directly detect the rotation speed ω _m of the rotation axis AX, the rotation speed calculation unit 32 for obtaining the rotation speed ω _m of the rotation axis AX from the detection result of the encoder 20 is omitted.

The flowchart showing the equipment diagnosis method according to this modification is almost the same as the flowchart shown in FIG. Specifically, in the flowchart showing the equipment diagnosis method according to this modification, "detection of rotation of the rotating shaft" in step S14 in FIG. 2 is read as "detection of the rotation speed of the rotating shaft", and step S15 is omitted. It will be the one that was done.

[Second Embodiment]
FIG. 5 is a block diagram showing a configuration of a main part of the equipment diagnostic apparatus according to the second embodiment of the present invention. As shown in FIG. 5, the equipment diagnosis device 2 of the present embodiment has a configuration in which a determination unit 50 is added to the equipment diagnosis device 1 shown in FIG. In the equipment diagnosis device 1 shown in FIG. 1, the user diagnoses the equipment by referring to the output (display) of the output unit 40. On the other hand, in the equipment diagnosis device 2 of the present embodiment, the determination unit 50 refers to the output of the calculation unit 30 to determine whether or not the equipment is abnormal.

The determination unit 50 is provided between the calculation unit 30 and the output unit 40, and determines whether or not the equipment is abnormal by using various values obtained by the calculation unit 30. The determination unit 50 at least compares the load correlation value ST obtained by the ST calculation unit 35 with the threshold value TH set in the load correlation value ST, and determines whether or not the equipment is abnormal.

In addition to the comparison between the load correlation value ST and the threshold value TH set in the load correlation value ST, the determination unit 50 sets the slip rate s of the rotation axis AX and the threshold value TH set in the slip rate s of the rotation axis AX. A comparison may be made to determine if the equipment is abnormal. Further, in addition to the above comparison, the determination unit 50 compares the frequency ω of the drive current with the threshold value TH set to the frequency ω of the drive current, and the rotation speed ω _m of the rotation axis AX and the rotation speed of the rotation axis AX. It may be determined whether or not the equipment is abnormal by comparing with the threshold value TH set to ω _m .

The determination unit 50 outputs the load correlation value ST obtained by the calculation unit 30, the slip ratio s of the rotation axis AX, the frequency ω of the drive current, and the rotation speed ω _m of the rotation axis AX to the output unit 40. The determination unit 50 outputs the above determination result (determination result of whether or not the equipment is abnormal) to the output unit 40. The determination unit 50 sets a threshold value TH set in each of the output unit 40 for the load correlation value ST, the slip ratio s of the rotation axis AX, the frequency ω of the drive current, and the rotation speed ω _m of the rotation axis AX to the output unit 40. You may output it.

The function of the determination unit 50 may be realized by software, for example, by reading and installing the program recorded on the recording medium by a computer, similarly to the functions of the calculation unit 30 and the output unit 40. Alternatively, the computer may be realized by software by installing a program downloaded via a network (not shown). Alternatively, it may be realized by using hardware such as FPGA, LSI, and ASIC.

The output unit 40 may output the above-mentioned determination result or threshold value TH output from the determination unit 50. For example, when the output unit 40 displays a graph showing the load correlation value ST, the slip ratio s of the rotation axis AX, the frequency ω of the drive current, and the rotation speed ω _m of the rotation axis AX, the threshold value TH May also be displayed. By performing such a display, the user can see that the output unit 40 has the load correlation value ST, the slip ratio s of the rotation axis AX, the frequency ω of the drive current, and the rotation speed ω _m of the rotation axis AX exceeding the threshold value TH. It can be easily confirmed whether or not it is. The user can also confirm whether the determination result of the determination unit 50 is correct or not.

The flowchart showing the equipment diagnosis method according to the present embodiment is basically the same as the flowchart shown in FIG. Specifically, in the flowchart showing the equipment diagnosis method according to the present embodiment, the determination unit 50 compares the ST calculation unit 35 and the like with the threshold value TH between steps S17 and S18 in FIG. 2, and the equipment is installed. An additional step is added to determine if is abnormal.

In the present embodiment, the load correlation value ST, the slip ratio s of the rotary shaft AX, the frequency ω of the drive current, and the rotation speed ω _m of the rotary shaft AX, the load correlation value ST, the slip ratio s of the rotary shaft AX, and the drive current. The frequency ω and the rotation speed ω _m of the rotation axis AX are compared with the threshold values individually set to determine whether or not the equipment is abnormal. Therefore, in the present embodiment, it is possible to automatically determine whether or not the equipment in which the rotary shaft AX is rotationally driven is abnormal.

Further, since the equipment diagnosis device 2 of the present embodiment is obtained by adding the determination unit 50 to the equipment diagnosis device 1 of the first embodiment, the basic configuration is the same as that of the equipment diagnosis device 1 of the first embodiment. Therefore, also in this embodiment, a large amount of sensors such as a vibration sensor and a temperature sensor are not required, and the equipment can be diagnosed with high accuracy even if the load fluctuation is small.

[Third Embodiment]
FIG. 6 is a block diagram showing a configuration of a main part of the equipment diagnostic apparatus according to the third embodiment of the present invention. As shown in FIG. 6, the equipment diagnostic apparatus 3 of the present embodiment has a configuration in which a diagnostic unit 60 is added to the equipment diagnostic apparatus 1 shown in FIG. In the equipment diagnosis device 1 shown in FIG. 1, the user diagnoses the equipment by referring to the output (display) of the output unit 40. On the other hand, in the equipment diagnostic apparatus 3 of the present embodiment, the diagnostic unit 60 is based on the learning results of various data output from the output unit 40 and various data output from the output unit 40 in the past. It is for diagnosing the presence or absence of abnormalities in. The output unit 40 may be omitted, and various values obtained by the calculation unit 30 may be output to the diagnosis unit 60.

FIG. 7 is a block diagram showing an internal configuration of a diagnostic unit according to the third embodiment of the present invention. As shown in FIG. 7, the diagnosis unit 60 includes an acquisition unit 61, a cutting unit 62, a feature extraction unit 63, a model processing unit 64, and an output unit 65. The cutout control signal C11 and the learning control signal C12 are input to the diagnosis unit 60. The cutout control signal C11 is a signal for controlling the cutout / cutout process of the data input to the diagnostic unit 60. The learning control signal C12 is a signal that controls whether or not learning is performed in the diagnostic unit 60. The on / off of the cutout control signal C11 and the learning control signal C12 is operated by, for example, a user or an external system.

The acquisition unit 61 acquires various data output from the output unit 40. The acquisition unit 61 acquires at least the load correlation value ST output from the output unit 40. The acquisition unit 61 may acquire the slip ratio s of the rotation axis AX in addition to the load correlation value ST. The acquisition unit 61 may acquire the frequency ω of the drive current and the rotation speed ω _m of the rotation axis AX in addition to the load correlation value ST and the slip ratio s of the rotation axis AX.

The acquisition unit 61 may acquire various data by performing wired communication or wireless communication with the output unit 40. The acquisition unit 61 associates the acquired data with the time information (time stamp). The acquisition unit 61 may acquire a data file associated with the time information from an external device. The format of the data file is, for example, a CSV (Comma-Separated Values) format.

The data X (t) acquired by the acquisition unit 61 is represented by the following equation (8). Here, t is a variable representing the discrete time (sampling time). d is the number of data acquired by the acquisition unit 61.

The cutout unit 62 cuts out the data (data continuously in time series) acquired by the acquisition unit 61 according to the cutout control signal C11. Specifically, the cutting unit 62 cuts out each continuous data (segment) in time series for a period in which the cutting control signal C11 is on, for example. For example, when the load correlation value ST, the slip ratio s of the rotation axis AX, the frequency ω of the drive current, and the rotation speed ω _m of the rotation axis AX are acquired by the acquisition unit 61, the cutting unit 62 may obtain the load correlation value ST, the rotation speed ω of the rotation axis AX, and the rotation speed ω m of the rotation axis AX. Each of these data is cut out for the period during which the cutout control signal C11 is on.

Each data S (k) cut out by the cutout portion 62 is represented by the following equation (9). Here, k is the serial number of the cut out data. n _k is the number of samples (hereinafter referred to as “segment length”) included in the kth cut-out data. The segment length does not have to be constant.

The cutting unit 62 may cut out each continuous data in time series from the time when the start end of the cutting control signal C11, which is a pulse signal, is detected to the time when a predetermined time elapses. The cutting unit 62 may cut out each continuous data in time series with respect to the time from the time when the end of the cutting control signal C11 which is a pulse signal is detected to the time when a predetermined time elapses.

Even if the cutting unit 62 cuts out each continuous data in time series from the time when the cutting control signal C11 which is a pulse signal is detected to the time when the next cutting control signal C11 is detected. good. The cutting unit 62 may cut out each continuous data in time series for a period from the time when the predetermined data becomes equal to or more than the threshold value to the time when the data becomes less than the threshold value. Further, the cutting unit 62 may cut out each data at a fixed cycle without relying on an external signal.

The feature extraction unit 63 extracts a feature vector from the cut out data S (k) by analyzing the data cut out by the cutout unit 62. That is, the feature extraction unit 63 generates the feature vector z (k) of the cut out data as shown in the following equation (10).

Here, z is a feature vector and is a vertical vector having an m-dimensional real value as a domain. The method by which the feature extraction unit 63 extracts the feature vector is not limited to a specific method. For example, the feature extraction unit 63 calculates statistics such as an average, a standard deviation, a minimum value, and a maximum value with respect to the device data cut out with the serial number k. When the feature extraction unit 63 calculates these four statistics, the feature extraction unit 63 generates an m (= 4 × d) -dimensional feature vector. The feature extraction unit 63 may extract the power of each periodic signal whose device data is decomposed by the Fourier transform as a feature vector. The feature extraction unit 63 may extract the order statistic of the device data as a feature vector.

The model processing unit 64 analyzes the feature vector extracted by the feature extraction unit 63. The model processing unit 64 generates a model of the feature vector (reference pattern model) based on the analysis result of the feature vector. The model processing unit 64 includes a learning unit 64a, a storage unit 64b, and a deviation degree calculation unit 64c.

When the learning control signal C12 is on, the learning unit 64a acquires the extracted feature vector from the feature extraction unit 63. The learning unit 64a analyzes (learns) the feature vector for each extracted feature vector. The learning unit 64a generates a model of the feature vector based on the analysis result of the feature vector. In this way, the learning unit 64a sequentially generates a model of the feature vector.

The model of the feature vector is not limited to the model of a specific form, but for example, a constraint condition (relationship) such as the following equation (11) that a continuous variable should satisfy between each element of the feature vector z is given. It is a model represented by the simultaneous equations. The learning unit 64a obtains the constraint condition as shown in the following equation (11) by executing the principal component analysis. The learning unit 64a may obtain constraint conditions using a neural network.

The model of the feature vector may be, for example, a model represented by the probability density distribution p (z) of the feature vector z. For example, the learning unit 64a generates a model represented by a multivariate normal distribution of the feature vector z. The multivariate normal distribution of the feature vector z is represented by the mean vector μ ∈ R ^m of the feature vector z and the covariance matrix Σ ∈ R ^{m × m} . The learning unit 64a calculates the multivariate normal distribution of the feature vector z using the following equations (12) and (13).

Here, z _sum represents the cumulative sum of the feature vectors. Z _sum represents the cumulative sum of the matrix product of the matrix of feature vectors and the vector to which the feature vectors are transposed. Each of these elements constitutes a sufficient statistic in a multivariate normal distribution. That is, the mean vector and covariance matrix of the multivariate normal distribution can be calculated based on z _sum and Z _sum .

The learning unit 64a has a cumulative sum of feature vectors z _sum (k-1) based on the first to (k-1) th cut out data, and a feature vector z (k) of the kth cut out data. Based on the above, the cumulative sum _zsum (k) of the feature vectors based on the cut-out data from the first to the kth is calculated using the following equation (14).

In the learning unit 64a, the cumulative sum Z _sum (k-1) of the matrix product of the feature vector based on the cut-out data from the first to the (k-1) th and the vector into which the feature vector is translocated, and k. Cumulative matrix product of a feature vector based on the first to kth cut out data and a vector into which the feature vector is transposed, based on the matrix Z (k) of the feature vector of the second cut out data. The sum Z _sum (k) is calculated using the following equation (15).

The learning unit 64a does not store the data acquired by the acquisition unit 61 in a database having a large storage capacity, and the cumulative sum of the mean value vector μ (k), the covariance matrix Σ (k), and the feature vector z _sum (k). ) And the cumulative sum of the matrix products of the feature vectors Z _sum (k). The multivariate normal distribution determined by Z sum (k) is calculated as a model of the feature vector z.

The learning unit 64a may calculate a probability distribution other than the multivariate normal distribution as long as it is a probability distribution that approximates the acquired data. For example, a probability distribution other than a multivariate normal distribution is a mixed multivariate normal distribution, an empirical density function based on multiple samples of data. The learning unit 64a records the generated model of the feature vector in the storage unit 64b.

The storage unit 64b is a non-volatile storage device (non-temporary recording medium) such as a flash memory. The storage unit 64b stores a model of the feature vector. The model of the feature vector is represented by, for example, mathematical data. The feature vector model is updated with the feature vector model generated by the learning unit 64a. The storage unit 64b may store the program. The storage unit 64b may further include a volatile recording medium such as a RAM (Random Access Memory).

The deviation degree calculation unit 64c acquires the feature vector z _test of the data newly acquired after the model of the feature vector is generated from the feature extraction unit 63. The deviation degree calculation unit 64c calculates the deviation degree (abnormality) between the feature vector z _test and the model of the feature vector stored in the storage unit 64b.

For example, when the model of the feature vector is represented by the probability distribution p (z) of the feature vector z, the deviation degree calculation unit 64c calculates the negative log-likelihood of the z _test as the deviation degree. For example, the divergence degree calculation unit 64c may calculate the divergence degree by executing a predetermined arithmetic process on the parameters of the mathematical formula representing the model of the feature vector. For example, in a model based on a constraint condition such as Eq. (4) between each element of the feature vector z, the deviation degree calculation unit 64c determines the degree of inconsistency of the constraint condition (for example, the L2 norm of f (z)). , May be calculated as the degree of divergence.

The output unit 65 outputs the calculated degree of deviation. The output unit 65 may output the calculated degree of deviation as data. For example, the output unit 65 may perform wired communication or wireless communication with an external device (for example, a data collection device) and output the above data to the external device. Further, the output unit 65 may output by writing the above data to the attached external memory. Alternatively, the output unit 65 may output by displaying the calculated degree of deviation on a display device such as a liquid crystal display device. The output unit 65 may display the calculated degree of deviation as a numerical value or may display it as a graph showing a change with time.

Similar to the functions of the arithmetic unit 30 and the output unit 40, the function of the diagnostic unit 60 may be realized by software, for example, by a computer reading and installing a program recorded on a recording medium. Alternatively, the computer may be realized by software by installing a program downloaded via a network (not shown). Alternatively, it may be realized by using hardware such as FPGA, LSI, and ASIC.

FIG. 8 is a flowchart showing an operation example of the diagnostic unit according to the third embodiment of the present invention. The processing of the flowchart shown in FIG. 8 is repeated, for example, at a preset constant cycle. When the processing of the flowchart shown in FIG. 8 is started, the processing of acquiring data is performed by the acquisition unit 61 (step S21). For example, the process of acquiring the load correlation value ST output from the output unit 40 shown in FIG. 6, the slip ratio s of the rotation axis AX, the frequency ω of the drive current, and the rotation speed ω _m of the rotation axis AX is performed. Will be.

Next, whether or not the condition for executing the cutting process is satisfied is determined by the cutting unit 62 (step S22). For example, when the cutout control signal C11 is on, the cutout unit 62 determines that the condition for executing the cutout process is satisfied. When it is determined that the condition for executing the cutting process is not satisfied (when the determination result in step S22 is "NO"), the processing in step S22 is performed again by the cutting unit 62.

On the other hand, when it is determined that the condition for executing the cutting process is satisfied (when the determination result in step S22 is "YES"), a predetermined condition is determined from the data acquired in step S21. The process of cutting out data based on the above is performed by the cutting unit 62 (step S23). For example, for the period during which the cutout control signal C11 is on, each data is continuous in time series of the load correlation value ST, the slip ratio s of the rotation axis AX, the frequency ω of the drive current, and the rotation speed ω _m of the rotation axis AX. The process of cutting out the data is performed by the cutting unit 62. Subsequently, the feature extraction unit 63 performs a process of extracting the feature vector z from each of the cut out data (step S24).

Subsequently, the learning unit 64a determines whether or not the condition for executing the learning process is satisfied (step S25). For example, when the learning control signal C12 is on, the learning unit 64a determines that the condition for executing the learning process is satisfied. When it is determined that the condition for executing the learning process is not satisfied (when the determination result in step S25 is "NO"), the process of step S28 is performed by the deviation degree calculation unit 64c.

On the other hand, when it is determined that the condition for executing the learning process is satisfied (when the determination result in step S25 is "YES"), the extracted feature vector z is analyzed and based on the analysis result. The process of generating the model of the feature vector is performed by the learning unit 64a (step S26). Then, the learning unit 64a performs a process of updating the model of the feature vector stored in the storage unit 64b to the newly generated model of the feature vector (step S27).

Subsequently, the process of acquiring the feature vector of the newly acquired data from the feature extraction unit 63 is performed by the deviation degree calculation unit 64c. Then, the process of calculating the degree of deviation between the feature vector of the newly acquired data and the model of the feature vector stored in the storage unit 64b is performed by the deviation degree calculation unit 64c (step S28). Finally, the output unit 65 performs a process of outputting the deviation degree calculated by the deviation degree calculation unit 64c (step S29). By referring to this degree of deviation, it is possible to diagnose the presence or absence of an abnormality in the equipment. This completes the series of processes shown in FIG. It is possible to exchange the execution order between the processes from step S25 to step S27 and the processes in step S28.

As described above, in the present embodiment, the diagnostic unit 60 is provided, and the presence or absence of an abnormality in the equipment is based on the learning results of various data output from the output unit 40 and various data output from the output unit 40 in the past. Is diagnosing. Specifically, various data output from the output unit 40 are cut out based on the cutout control signal C11, a feature vector is extracted from the cutout data, the feature vector is analyzed, and a feature vector model is created. There is. Then, the degree of deviation between the feature vector of the newly acquired data and the model of the feature vector is calculated. By referring to this degree of deviation, it is possible to diagnose the presence or absence of an abnormality in the equipment.

Although the equipment diagnosis device and the equipment diagnosis method according to the embodiment of the present invention have been described above, the present invention is not limited to the above embodiment and can be freely changed within the scope of the present invention. For example, the above-mentioned second and third embodiments may be applied to the above-mentioned modifications of the first embodiment.

Further, in the above-described embodiment, an example has been described in which the output unit 40 outputs the load correlation value ST, the slip ratio s of the rotation axis AX, the frequency ω of the drive current, and the rotation speed ω _m of the rotation axis AX. In addition to these, the output unit 40 may output the effective value _Ia of the drive current obtained by the effective value calculation unit 34. In the equipment diagnostic apparatus 3 of the third embodiment, when the output unit 40 outputs the effective value I _a of the drive current, the diagnostic unit 60 also uses the effective value I _a of the drive current to determine the presence or absence of an abnormality in the equipment. You may make a diagnosis.

In the equipment diagnostic apparatus 2 of the second embodiment, the determination unit 50 is set to the load correlation value ST, the slip ratio s of the rotation axis AX, the frequency ω of the drive current, and the rotation speed ω _m of the rotation axis AX. The comparison with the threshold value TH was performed. The determination unit 50 may also use the effective value _Ia of the drive current to determine whether or not the equipment is abnormal. Specifically, the effective value _Ia of the drive current and the threshold value TH set in the effective value _Ia of the drive current may be compared to determine whether or not the equipment is abnormal.

Further, in each of the above-described embodiments, the load value may be calculated using the load correlation value ST obtained by the ST calculation unit 35. Specifically, the load correlation value ST may be converted into the load value D by performing the operation D = A · ST + B. Here, the variable A in the above equation defines the scaling (scale ratio) of the load correlation value ST, and the variable B in the above equation defines the offset of the load correlation value ST. By performing such a conversion, for example, when the rotary shaft AX is the rotary shaft of the stirrer, the load of the substance being agitated by the stirrer can be obtained. The obtained load value D may be smoothed by using a filter such as a first-order lag filter (low-pass filter).

1,1A Equipment diagnosis device 2 Equipment diagnosis device 3 Equipment diagnosis device 10 Current transformer 20 Encoder 20A Tachometer 30, 30A Calculation unit 31 Frequency calculation unit 33 Slip rate calculation unit 34 Effective value calculation unit 35 ST calculation unit 40 Output unit 50 Judgment Part 60 Diagnosis part 61 Acquisition part 62 Cutting part 63 Feature extraction part 64a Learning part 64c Deviation degree calculation part AX Rotation axis IM Induction motor TH threshold

Claims (10)

An equipment diagnostic device that diagnoses equipment with a rotating shaft that is driven to rotate by an induction motor.
A current detector that detects the drive current supplied to the induction motor, and
A rotation detection unit that detects the rotation of the rotation axis, and
Using the detection result of the current detection unit and the detection result of the rotation detection unit, a calculation unit for obtaining a load correlation value which is a value having a correlation with the load resistance acting on the rotation axis, and a calculation unit.
An output unit that outputs the load correlation value and
Equipped with
The calculation unit is a value obtained by dividing the frequency of the drive current obtained from the detection result of the current detection unit by the rotation speed of the rotation shaft obtained from the detection result of the rotation detection unit, a value obtained by the current detection unit. The load correlation value is the square of the effective value of the drive current obtained from the detection result, and the product of the slip ratio of the rotating shaft obtained by using the detection result of the current detection unit and the detection result of the rotation detection unit. Ask as,
Equipment diagnostic equipment. The calculation unit includes a frequency calculation unit that obtains the frequency of the drive current from the detection result of the current detection unit, and a frequency calculation unit.
Using the detection result of the current detection unit and the detection result of the rotation detection unit, the slip rate calculation unit for obtaining the slip rate of the rotation axis, and the slip rate calculation unit.
An effective value calculation unit that obtains the effective value of the drive current from the detection result of the current detection unit, and
A load correlation value calculation unit for obtaining the load correlation value using the calculation result of the frequency calculation unit, the calculation result of the slip rate calculation unit , and the calculation result of the effective value calculation unit.
The equipment diagnostic apparatus according to claim 1. The equipment diagnostic device according to claim 2, wherein the output unit outputs the slip ratio of the rotating shaft in addition to the load correlation value. The output unit outputs, in addition to the load correlation value and the slip ratio of the rotation shaft, the frequency of the drive current and the rotation speed of the rotation shaft obtained from the detection result of the rotation detection unit. The equipment diagnostic apparatus according to 3. The load correlation value is ST, the frequency of the drive current is ω [rad / s] or [Hz], the rotation speed of the rotation axis is ω _m [rad / s] or [Hz], and the effective value of the drive current. Is I _a , and the slip rate of the rotating shaft is s [%], the load correlation value calculation unit performs the calculation represented by the following equation (1) to obtain the load correlation value ST, claim 2 . The equipment diagnostic apparatus according to any one of claims 4 .

The claim includes at least a determination unit for comparing the load correlation value output from the output unit with a threshold value set for the load correlation value to determine whether or not the equipment is abnormal. The equipment diagnostic apparatus according to any one of claims 1 to 5. A claim including at least a diagnostic unit for diagnosing the presence or absence of an abnormality in the equipment based on the load correlation value output from the output unit and the learning result of the load correlation value output from the output unit in the past. The equipment diagnostic apparatus according to any one of items 1 to 5. The diagnostic unit includes an acquisition unit that acquires data of the load correlation value output from the output unit, and an acquisition unit.
A cutting unit that cuts out data from the data acquired by the acquisition unit based on predetermined conditions, and a cutting unit.
A feature extraction unit that extracts a feature vector from the data cut out by the cutout unit, and a feature extraction unit.
A learning unit that analyzes the feature vector extracted by the feature extraction unit and generates a model of the feature vector based on the analysis result.
A deviation degree calculation unit that calculates the deviation degree between the feature vector and the model of the data newly acquired by the acquisition unit, and
Equipped with
The presence or absence of abnormality in the equipment is diagnosed with reference to the degree of deviation calculated by the degree of deviation calculation unit.
The equipment diagnostic apparatus according to claim 7. It is an equipment diagnosis method for diagnosing equipment having a rotating shaft that is rotationally driven by an induction motor.
A current detection step for detecting the drive current supplied to the induction motor, and
A rotation detection step for detecting the rotation of the rotation axis, and
Using the detection result of the current detection step and the detection result of the rotation detection step, a calculation step for obtaining a load correlation value which is a value having a correlation with the load resistance acting on the rotation axis, and a calculation step.
An output step that outputs the load correlation value and
Have,
The calculation step is a value obtained by dividing the frequency of the drive current obtained from the detection result of the current detection step by the rotation speed of the rotation shaft obtained from the detection result of the rotation detection step, and the current detection step. The load correlation value is the square of the effective value of the drive current obtained from the detection result, and the product of the slip ratio of the rotation axis obtained by using the detection result of the current detection step and the detection result of the rotation detection step. Ask as,
Equipment diagnosis method. It is an equipment diagnosis program that makes a computer function as an equipment diagnosis device that diagnoses equipment having a rotating shaft that is rotationally driven by an induction motor.
The load resistance acting on the rotating shaft of the computer using the detection result of the current detecting unit that detects the drive current supplied to the induction motor and the detection result of the rotation detecting unit that detects the rotation of the rotating shaft. A calculation means for obtaining a load correlation value, which is a value having a correlation with
An output means for outputting the load correlation value and
To make it work
The calculation means is a value obtained by dividing the frequency of the drive current obtained from the detection result of the current detection unit by the rotation speed of the rotation shaft obtained from the detection result of the rotation detection unit, a value obtained by the current detection unit. The load correlation value is the square of the effective value of the drive current obtained from the detection result, and the product of the slip ratio of the rotating shaft obtained by using the detection result of the current detection unit and the detection result of the rotation detection unit. Ask as,
Equipment diagnostic program.

Images