ROTATING EQUIPMENT

TECHNICAL FIELD

[001] The present invention relates in general to condition monitoring of rotating equipment. More particularly, the present invention relates to estimating a running speed of the rotating equipment.

BACKGROUND

[002] Rotating equipment such as, motors, generators, pumps and the like are highly important in industrial applications. The rotating equipment are affected by environmental, mechanical, and other issues, most of which needs to be avoided if possible or rectified at the earliest in case the downtime is unavoidable. Due to these issues, condition monitoring of the rotating equipment has been an important factor in industries. The condition monitoring of rotating equipment is a process of monitoring parameters of a particular condition in the rotating equipment (such as, vibration, speed, etc.). The condition monitoring of rotating equipment typically involves prediction of conditions requiring maintenance and / or other support, from the monitored parameters of the rotating equipment.

[003] Multiple parameters of the rotating equipment are monitored during the condition monitoring. One of the parameters being running speed of the rotating equipment. Accurate speed estimation is an important step in condition analysis of the rotating equipment.

[004] Existing techniques for speed estimation of the rotating equipment offer a combined analysis of vibration and magnetic field. However, the vibration analysis-based speed estimation may sometime result in incorrect speed estimation due to higher vibrations at multiples of speed and not close to actual speed, or due to vibrations originating from sources other than the equipment itself. Similarly, magnetic field-based analysis may not be useful for loading below certain range, or for equipment having less information in terms of signal parameters. SUMMARY

[005] The present invention relates to a method and a condition monitoring device for estimating running speed of a rotating equipment. The rotating equipment can be one of, but not limited to, motors, generators, pumps and the like. In an embodiment, the condition monitoring device may be affixed to the rotating equipment. Alternatively, the condition monitoring device can be a portable device associated with the rotating equipment. The condition monitoring device comprises one or more sensors for measuring parameters associated with the rotating equipment. [006] The one or more sensors include at least one magnetometer. Optionally, the one or more sensors also include an accelerometer. The one or more sensors can be triaxial sensors. Thus, the parameters of the rotating equipment are measured at three different axes. The one or more sensors measure the parameters at periodic instants. The condition monitoring device includes a memory for storing the parameter values obtained at each instant and a processor for processing the parameter values measured using the one or more sensors for estimating the running speed of the rotating equipment.

[007] The method comprises obtaining the measurements from the one or more sensors of the condition monitoring device at periodic instants. The measurements correspond to a measurement location determined based on a placement of the condition monitoring device relative to the rotating machine. For example, the one or more sensors (for example, tri axial magnetometer or accelerometer) of the condition monitoring device can generate signals associated with a magnetic field or acceleration of the rotating equipment at a measurement location relative to the rotating equipment. The signals have measurements of the magnetic field or the acceleration along a first axis, second axis and a third axis. In an embodiment, the measurements obtained from the one or more sensors in at least one axis may be in a radial direction, based on the placement of the condition monitoring device. The measurements may alternatively be in axial or tangential directions. [008] At each instant, the method comprises determining a value for a parameter associated with the rotating equipment based on the obtained measurements. In an embodiment, the parameter is a magnetometer amplitude corresponding to supply frequency, when measurement is obtained from the magnetometer. Alternatively, the parameter is a root mean square (RMS) of one of acceleration and velocity, when the measurement is obtained from the accelerometer.

[009] Based on the value of the parameter, the method comprises obtaining an estimate of running slip of the rotating equipment for the instant using a regression model. The regression model comprises at least one of a linear relation and a polynomial relation between the running slip and the parameter, and the relation is determined based on least square fitting of the data points of the historic data. The regression model is generated using historic data collected from the one or more sensors and comprises a relation between the running slip and the parameter (the parameter being one of the magnetometer amplitude and the RMS of one of acceleration and velocity). The historic data comprises a plurality of paired data points. In an embodiment, each pair of data points comprises the magnetometer amplitude and the associated running slip. In another embodiment, each pair of data points comprises the RMS value and the associated running slip.

[010] Each of the data points are associated with a quality level and the data points are selected for generating the model based on the quality of the data point. The quality of each data point is determined based on a signal parameter associated with the one or more sensors. In an embodiment, the historic data comprises the data points of one or more of the three quality levels, and the signal parameter is a Signal to Noise Ratio (SNR), and the three quality levels are determined based on three SNR values.

[Oi l] In an embodiment, the regression model is generated offline and stored in the condition monitoring device, where model parameters may be updated from the server. In another embodiment, the regression model is stored in a server. The regression model can be updated or adapted if the location of the condition monitoring device changes relative to the rotating equipment.

[012] Thereafter, the method comprises estimating the running speed of the rotating equipment using the obtained running slip and the value of the parameter. The running speed is utilized for determining a performance indicator of the rotating equipment.

[013] Additionally, the method comprises providing the measurements obtained from the one or more sensors to a server, which stores the regression model. In response to providing the measurements, the method comprises receiving at least one model parameter from the server for estimating the running speed.

BRIEF DESCRIPTION OF DRAWINGS

[014] Figures 1A and IB illustrate monitoring of a motor with a condition monitoring device for estimating running speed of the motor, in accordance with different embodiments of the invention;

[015] Figure 2 shows a simplified representation of the condition monitoring device, in accordance with an embodiment of the invention;

[016] Figure 3 illustrates one or more axes for measurement with one or more sensors of the condition monitoring device, in accordance with an embodiment of the invention; [017] Figures 4 is a flowchart of a method for estimating running speed of a rotating equipment, in accordance with an embodiment of the invention;

[018] Figures 5 is a flowchart of a method for generating a regression model, in accordance with an embodiment of the invention; and

[019] Figure 6 is a sample plot illustrating a linear relationship between running slip and magnetometer amplitude of a rotating equipment.

DETAILED DESCRIPTION

[020] The invention provides a method and a condition monitoring device for estimating running speed of a rotating equipment. The rotating equipment may be one of several rotating equipment in an industrial environment such as a power plant, an oil and gas plant, a paper manufacturing unit etc. The rotating equipment may be one of, but not limited generators, motors, pumps and the like, used in an industrial environment. The method is performed with the condition monitoring device associated with the rotating equipment (e .g . a motor). The condition monitoring device may be a non-invasive condition monitoring device, or a portable device. The condition monitoring device estimates the running speed of the rotating equipment which is utilized for determining a performance indicator of the rotating equipment. [021] Reference is now made to Figure 1A showing a schematic diagram of a rotating equipment, such as a motor 100. As shown in Figure 1A, the motor 100 is affixed with a condition monitoring device 101. The condition monitoring device 101 may be a portable handheld device, which can be brought proximal to the motor for taking measurements. The condition monitoring device 101 is configured to monitor parameters of the motor 100 to assess a condition of the motor 100. The condition monitoring device 101 comprises one or more sensors (not shown in Figure 1 A explicitly) for monitoring the parameters of the motor 100.

[022] The one or more sensors measure the parameters based on a placement of the condition monitoring device 101 relative to the motor 100. In an embodiment, the one or more sensor includes a magnetometer. In another embodiment, the one or more sensors includes an accelerometer. Thus, the condition monitoring device 101 can have one of a magnetometer or an accelerometer or both. There may be also other sensors such as acoustic sensors, temperature sensors and so forth. The condition monitoring device 101 is configured with a model (not shown in Figure 1A, covered in Figure 2). This model is a regression model that is generated offline (or at a server having a model generator) based on historic data collected from the one or more sensors. The historic data from each of the one or more sensor is analysed for generating the regression model.

[023] The regression model is generated using a statistical modelling technique for determining relationship between parameters such as by using data fitting, curve fitting or other graphical analysis based-techniques. The regression model is generated using historic data which comprises a plurality of paired data points. In an embodiment, each pair of the data points include the magnetometer amplitude and an associated running slip. In another embodiment, each pair of the data points may include RMS value of one of acceleration and velocity and the associated running slip. The parameters (magnetometer amplitude, RMS or running slip) are estimated with processing of the measurements. The regression model includes a relation between the running slip and a parameter. For example, the parameter can be the magnetometer amplitude or the RMS value, and accordingly the model has a relation between the magnetometer amplitude and the running slip, or the RMS and slip. In an embodiment, the relation is a linear relation where the running slip is linearly proportional to increase (or decrease) of the magnetometer amplitude. In another embodiment, the relation is a polynomial relation. The relation is determined based on a technique such as, least square fitting of the data points of the historic data.

[024] In an embodiment, the regression model is stored in a server as shown in Figure IB. As shown in Figure IB, the condition monitoring device 101 associated with the motor 100 is connected with a server 105. The server 105 stores regression model 107 and historian data 109. The regression model 107 is used by the condition monitoring device 101 in real time for estimating the running seed of the rotating equipment. In such a scenario, according to relative placements, the one or more sensors may measure data and send processed data or raw measurements to the server 105 for processing. Thereafter, the condition monitoring device 101 receives information from the server 105 for estimating the running speed of the rotating equipment.

[025] A simplified representation of the condition monitoring device 101 is shown in Figure 2. The condition monitoring device 101 comprises a housing body 200 capable of being affixed to a body, or shell, or frame of the rotating equipment such as, a motor 100 (as shown in Figure 1A and IB). The housing body 200 houses one or more sensors including, but not limited to, a magnetometer 204, and an accelerometer 205. The one or more sensors may include other sensors 206 such as acoustic sensors, temperature sensors and so forth. The magnetometer 204 and the accelerometer 205 are collectively represented as one or more sensors throughout the invention. The one or more sensors record and transmit measurements. The measurements can be transmitted to a local storage on the condition monitoring device 101, or to a remote device (e.g. a server). According to the positioning of the condition monitoring device 101 relative to the rotating equipment, sensors of the condition monitoring device 101 take measurements.

[026] The condition monitoring device 101 also includes a memory 201 (local storage) for storing data. The data may include measurements taken by the one or more sensors for periodic instants while monitoring the rotating equipment (e.g. from few minutes upto an hour or so). The data may include the measurements from one of the magnetometer 204 and the accelerometer 205, which is used for determining a value of parameters associated with the rotating equipment. For instance, the data measured by the magnetometer 204 is used for determining the value of magnetometer amplitude corresponding to supply frequency (or line frequency) or corresponding to a harmonic (e.g. first or second harmonic of supply frequency). Similarly, the data measured by the accelerometer 205 is used for determining the value of root mean square (RMS) of one of acceleration and velocity. The memory 201 may also store the regression model 107 which is generated offline as explained in Figure 1A. This may be computationally less complex, considering limited processing capability at the condition monitoring device 101. Also, such a model may be updated from time to time with updated model parameters being determined at a remote location (server).

[027] The magnetometer 204 is used to measure the magnetic field or leakage magnetic flux around the rotating equipment, such as the motor 100. The magnetic flux measured in one or more axes is used to determine amplitude corresponding to supply frequency of the rotating equipment. For example, amplitude in radial direction may be more dominant and therefore utilized.

[028] The accelerometer 205 may be used in addition to the magnetometer 204. The accelerometer is used to measure acceleration/velocity of a component such as a rotor. The accelerometer 205 is used for measuring the acceleration/velocity of the rotating equipment for each measurement instant (which may be in minutes or a duration configured as per need), which then is used by the condition monitoring device 101 for detecting the RMS value of acceleration/velocity. Thus, at each periodic instant, data from the magnetometer 204 and the accelerometer 205 are measured for determining the value of parameter of the rotating equipment. It is to be noted that data from only magnetometer or accelerometer is sufficient for the invention, however having both measurements can help in increasing the confidence.

[029] In addition, the condition monitoring device 101 includes a processor 202. The processor is configured to execute various steps involved in the monitoring, and such steps may be stored in the memory as executable steps (or instructions) and utilized during run time along with other information in the memory.

[030] The processor is configured to have the measurements received from the one or more sensors in order to estimate the running speed of the rotating equipment. At the periodic instant, the processor 202 receives the measurements (e.g. with a data logger and logged in the memory) from one of the magnetometer 204 and the accelerometer 205 to determine the value of the parameter associated with the rotating equipment. Particularly, the processor 202 may determine the value of the magnetometer amplitude corresponding to supply frequency in case of the magnetometer 204 and / or the RMS of one of the acceleration and velocity in case of the accelerometer 205.

[031] In an embodiment, the measurements are taken in a radial direction, and the radial direction is determined based on the placement of the condition monitoring device 101. The measurements may also be taken in one of but not limited to, axial and tangential direction. Thus, measurements in one or different axis may be utilized.

[032] Figure 3 illustrates one or more axes for measurement with the condition monitoring device, in accordance with an embodiment of the invention. Figure 3 shows a sectional view of radial magnetic field around the motor 100, in accordance with various embodiments of the present invention. As shown in figure 3, the magnetic fields around the motor 100 can be categorized in radial direction. The radial direction indicates plane perpendicular to the axis of the motor 100. It has been determined through experimentation that the radial direction comprises higher signal strength from the one or more sensors.

[033] Referring back to Figure 2, the condition monitoring device 101 may include a network interface 203 configured for communication with an external device (such as, the server 105). In one embodiment, the network interface 203 is capable of communicating over wireless media such as Bluetooth, Wireless HART, etc. The network interface 203 may communicate the measurements using an antenna 207 as shown in Figure 2. In another embodiment, the network interface 203 may be configured to facilitate communication between the processor 202 and the server 105.

[034] The processor 202 based on the value of one of the magnetometer amplitude and the RMS of one of the acceleration and velocity may obtain an estimate of running slip for the instant using the regression model. The processor 202 may utilise the value of one of the parameter, i.e., the magnetometer amplitude and the RMS of one of the acceleration and velocity and check a relation between the running slip and the parameter. In case, the regression model is stored in the server 105 as shown in Figure IB, the processor 202 may obtain the running slip by providing the measurements (or the magnetometer amplitude or RMS as estimated) from the one or more sensors to the server 105, which based on the regression model 107 provides at least one model parameter to the processor 202 for estimating the running speed.

[035] The model parameter may include for instance, a curve value and offset parameters obtained based on a plot of the parameters obtained from the measurements (e.g. linear relation between magnetometer amplitude and running slip). Once the running slip is obtained, the processor 202 estimates the running speed of the rotating equipment using the running slip and the value of the parameter (e.g. amplitude). The estimated running speed is utilized for determining a performance indicator of the rotating equipment.

[036] The condition monitoring device 101 may be mounted on a frame of the rotating equipment for receiving measurements through the one or more sensors of the condition monitoring device 101. In such a scenario, according to their relative placements, the one or more sensors send measurements to the processor 202, which may process the measurement to determine the running speed of the rotating equipment.

[037] The running speed of the rotating equipment can serve as input for determining the performance indicator of the rotating equipment such as, analysis for power estimation, current estimation, condition monitoring, etc.

[038] The above Figures 1 - 3 are explained considering the rotating equipment as the motor 100. The present invention is not restricted to motor 100. The system can also be implemented in other rotating equipment such as, generators, pumps and the like. Also, the condition monitoring device may be a self-powered device which is powered with an energy source such as one or more batteries.

[039] Referring now to Figure 4, which is a flowchart of a method of a method for estimating running speed of a rotating equipment in accordance with an embodiment of the invention. Various steps of the method may be performed by the condition monitoring device 101, or at least in part by condition monitoring device 101.

[040] At 401, the measurements from the one or more sensors is obtained (e.g. with the processor 202). The measurements are obtained at periodic instants and the measurements correspond to the measurement location determined based on the placement of the condition monitoring device 101 relative to the rotating machine. [041] At 402, for each instant, the value for the parameter associated with the rotating equipment is determined (e.g. by the processor 202) based on the measurements. In an embodiment, the measurements include the value of the parameter in the radial direction based on the placement of the condition monitoring device 101. The parameter is one of the magnetometer amplitude, when the measurements are obtained from the magnetometer 204 and RMS of one of the acceleration and velocity, when the measurements are obtained from the accelerometer 205.

[042] At 403, the estimate of the running slip for the instant is obtained by the processor 202 using the regression model 107 and the value of the parameter. As mentioned, the regression model 107 comprises the relation between the running slip and the parameter. As explained in Figure IB, the regression model 107 is generated from the historic measurements collected from the one or more sensors.

[043] Figures 5 is a flowchart of a method for generating a regression model in accordance with an embodiment of the invention.

[044] As shown in Figure 5, at 405, the measurements corresponding to historic data is obtained from the one or sensors.

[045] At 406, quality level for each measurement is calculated (or utilized) based on a signal parameter. The signal parameter includes a Signal to Noise Ratio (SNR) for the rotating equipment.

[046] At 407, the SNR associated with each measurement is compared against a first threshold (first SNR value) associated with first quality level. Based on the comparison, if the SNR is greater than the first threshold, the amplitude, slip and SNR is added to historian data 109 and confidence level is set to“1” (at 411).

[047] Alternatively, (at 408) if the SNR is less than the first threshold SNR, the SNR is compared against a second threshold (second SNR value) associated with a second quality level. Based on the comparison, if the SNR is greater than the second threshold, the method moves to block 412. Otherwise, the method moves to block 409. [048] At block 409, the SNR is compared against a third threshold (third SNR value) associated with a third quality level. If the SNR is greater than the third threshold, the method moves to block 413 and checks if the count of data is reached. In case, the count has reached, at 414, the result is compared, and the corresponding amplitude, slip and SNR is added to historian data 109 and quality level is set as second level. Otherwise, if the count is not reached, at 415, the corresponding amplitude, slip, quality level is stored with confidence level =”0”.

[049] At block 412, count of the data is checked. In case, the count has reached, at 416, the result is compared, and the corresponding amplitude, slip and SNR is added to historian data 109 and quality level is set as second level. Otherwise, if the count is not reached, at 417, the corresponding amplitude, slip, quality level is stored with confidence level =”0”.

[050] At 418, count of data is checked to ensure sufficient data points are added for generating the regression model 107. In case, the data count has reached a predefined number, the data points are utilised to identify relationship between the amplitude and slip of the rotating equipment by using square fitting of the data points (at 419). The relationship is determined by plotting the data points and identifying curve and offset parameters. In an embodiment, the relationship may be at least a linear relation and a polynomial relation between the running slip and the amplitude.

[051] In one implementation, the relation is determined based on the least square fitting of the data points. Alternatively, in case the data count is not sufficient, the system is reset to determine additional data points. Thus, the data points selected are stored as historic data in the historian data 109. In an embodiment, if the one or more sensors are moved from its location, the method obtains new data points and regenerates the regression model 107 based on the new data points. Similar adaptation or tuning may be performed, when there is a significant drift in the relation between the parameters (e.g. due to ageing). The model has a relation between the various parameters, which is utilized to estimate the running speed.

[052] Figure 6 is a sample plot illustrating a linear relationship between running slip and magnetometer amplitude of a rotating equipment. As shown in the plot Figure 6, with increase in running slip (in Hz), the magnetometer amplitude corresponding to supply frequency in radial direction reduces linearly. It is to be noted that the relationship between the running slip and magnetometer is linearly decreasing, however the relationship may be increasing or have other behaviour depending on the application, load etc.

[053] Returning to Figure 4, at 404, the running speed of the rotating equipment is estimated using the running slip and the value of the parameter. The running speed is utilized for determining a performance indicator of the rotating equipment.

[054] The present invention offers improved accuracy in the speed estimation of the rotating equipment by identifying relationship between parameters with quality data collected with onboard sensors. Thus, data points which could affect or provide incorrect estimates are ignored, and a superior estimate of the running speed is obtained.

[055] The present invention accordingly helps in improving the Key Performance Indicator (KPI) computations such as analysis for power estimation, current estimation, condition monitoring, and so forth.

REFERRAL NUMERALS