Managing health of rotating systems

Abstract

The present disclosure relates to systems, apparatuses, and methods for managing health conditions of at least one rotating system. The method includes receiving, by a processing unit, operational data associated with the rotating system from one or more sensing units (115) in real-time. The operational data includes parameter values corresponding to the operation of the rotating system (110). Further, a virtual replica of the rotating system (110) is configured using the operational data. The behavior of the rotating system is simulated on a simulated instance of the rotating system (110) based on the configured virtual replica. The simulation results are analyzed to determine an anomaly in the health condition of the rotating system (110). The anomaly corresponds to a health state of an internal component of the rotating system (110). Further, a notification indicative of the anomaly is generated on a graphical user interface (160).

Description

[0001] The present invention relates to the field of rotating systems, and more particularly to managing the health of a rotating system.

[0002] Rotating systems, such as generators and electric machines, are used in various applications. As the rotating systems are used, they can develop a number of problems, which manifest as abnormalities in the health of the rotating system. Abnormalities related to external components of the rotating system can be easily detected and addressed. However, abnormalities related to internal components of the rotating system can be difficult to detect, as they are inside the rotating system. Eventually, undetected abnormalities can lead to breakdown of the rotating system. Breakdown of the rotating system can further cause damage to equipment operatively coupled with the rotating system, and can also endanger human life. When the rotating system is part of a larger setup, breakdown of the rotating system can cause downtime of the setup, and also cause loss of productivity. Current known techniques do not support determination of abnormalities associated with internal components of the rotating system, before breakdown of the rotating system.

[0003] WO2015149928 discloses a method for online evaluation of operating range and performance of a compressor, the method comprising at least the following offline steps: setting up a digital undegraded model of the compressor in an undegraded state of the compressor; calibrating and validating the undegraded model using historical data from a compressor operating line; calculating at least one undegraded performance map using the undegraded model; and the method comprising at least the following online steps: updating the undegraded model by operating data of the compressor determined by at least one sensor of the compressor to calculate a degradation-adaptive model; calculating at least one actual performance map of the operating state of the compressor using the degradation-adaptive model; detecting a breakdown of the compressor by comparing at least one undegraded performance map derived from the undegraded model and at least one actual performance map derived from the degradation-adaptive model via at least one mathematical function. However, this patent application does not address the problem of determining specific abnormalities related to internal components of the rotating system, which sensors cannot be positioned in.

[0004] EP3255512A2 discloses an adaptive model based method for quantifying degradation of a power generation system. According to EP3255512 A2, a correction factor is used to reduce the difference between an estimated parameter value and a measured parameter value. The correction factor is further used to generate a transfer function that relates the estimated and measured values of a second parameter. However, this transfer function fails to relate the estimated parameter value and specific abnormalities associated with internal components of the power generation system.

[0005] EP2975525A1 discloses a system analysis device and a system analysis method for generating a correlation graph. However, it fails to disclose how to apply such a correlation graph for determining a specific anomaly associated with an internal component of a rotating system.

[0006] US20130024416A1 discloses a method for determining a future operating state of an object. The method comprises obtaining reference data indicative of a normal operating state of the object and obtaining an input pattern array. Each input pattern array has a plurality of input vectors, while each input vector represents a point in time and has input values representing a plurality of parameters indicative of a current condition of the object. However, this patent application does not address the problem of determining an anomaly with respect to an internal component of a rotating system based on such reference data and input arrays.

[0007] US20190287005A1 discloses a method comprising receiving operating data associated with operating conditions of a pump; and determining, using a first predictive model, prognostic data using the received operating data. The prognostic data includes a plurality of metrics associated with the pump and predicted with respect to the operating data. The operating conditions relate to motor failure, pump failure, cable or motor lead extension failure, seal failure, shaft and / or coupling failure, etc. However, this patent application does not discuss how to determine an anomaly associated with an internal component of a rotating system.

[0008] In view of the above, there is a need for a method and a system for managing the health condition of a rotating system. More specifically, there is a need for determining in real time an anomaly due to the health condition of a specific internal component of a rotating system.

[0009] It is therefore an object of the present invention to provide a computer-implemented method, apparatus, system and computer program product for managing the health condition of a rotating system based on an accurate determination of an anomaly in the rotating system. This object is solved by the method according to claim 1, the apparatus according to claim 7, the system according to claim 8, and the computer program product according to claim 9. Further advantageous embodiments and improvements of the present invention are listed in the dependent claims. In the following, some aspects of the present invention are described which are helpful for understanding the present invention before the detailed description of embodiments of the present invention with reference to the accompanying drawings. However, it should be noted that the present invention is defined by the appended claims and any examples and embodiments not covered by these claims should also be understood as helpful for understanding the present invention.

[0010] According to one aspect, the present invention is a method for managing health of a rotating system, as disclosed herein. The term "rotating system" used herein can refer to any electrical machine operating on the principle of electromagnetic induction, including a rotor and a stator separated by an air gap. Non-limiting examples of the rotating system include an AC generator, a DC generator, an AC motor, a DC motor, a motor amplifier, a synchronizer, a metadyne, a flux clutch, a flux brake, a flux dynamometer, a hysteresis dynamometer, and a rotary transducer.

[0011] According to another aspect, the method for managing health of a rotating system includes receiving, by a processing unit, operational data associated with the rotating system in real-time from one or more sensing units. The operational data includes parameter values corresponding to the operation of the rotating system. Non-limiting examples of the parameter values can include vibration frequency, vibration amplitude, magnetic field strength, magnetic flux density, noise amplitude, noise frequency, current, voltage, temperature, and the like.

[0012] According to another aspect, the method further includes configuring a virtual replica of the rotating system using the operational data. The virtual replica is a virtual representation of the rotating system. In one aspect, the virtual replica can be based on one or more models of the rotating system. Non-limiting examples of the one or more models include a CAD model, a 1D model, a 2D model, a 3D model, a meta-model, a stochastic model, a parametric model, a machine learning model, an artificial intelligence (AI) based model, a neural network model, a deep learning model, and the like. The virtual replica of the rotating system is configured using the operational data by updating the virtual replica based on the operational data in real-time using simulation instances. In another aspect, the virtual replica can be created based on information received from an original equipment manufacturer (OEM) of the rotating system and stored in a device. In another aspect, the virtual replica can be stored in a user device, a personal computer, a removable storage device, a server, a cloud storage device, and the like. Further, the stored virtual replica can be accessed and downloaded on the device.

[0013] Advantageously, the virtual replica facilitates soft sensor techniques for determining anomalies in the rotating system. More specifically, the virtual replica determines anomalies associated with internal components of the rotating system based on the operational data in real-time.

[0014] According to another aspect, the method further includes generating simulation results by simulating behavior of the rotating system on at least one simulation instance of the rotating system based on the configured virtual replica of the rotating system. The simulation results are indicative of the behavior of the rotating system. More specifically, the simulation results are generated by executing the virtual replica on a simulation platform for the at least one simulation instance.

[0015] Advantageously, the virtual replica is executed using the operation data received in real time. The simulation results generated correspond to the real time behavior of the rotating system. Moreover, the simulation speed can be maximized by employing a model of the rotating system.

[0016] According to another aspect, the method further comprises determining an anomaly in the health condition of the rotating system based on the analysis of the simulation results. The anomaly in the health condition corresponds to a health state of at least one internal component of the rotating system. It must be understood that the term "anomaly" used herein refers to data associated with an abnormal condition. The abnormal condition can include, but is not limited to, air gap asymmetry, rotor vibration, rotor displacement, magnetic field asymmetry, air gap eccentricity, unbalanced forces, heating, bearing defects, rotor bar breakage, stator related issues, etc. Non-limiting examples of internal components include rotor, stator, bearing, stator coil, brush, etc. In one aspect, the method further comprises determining a deviation in the behavior of the rotating system by analyzing the simulation results with respect to an expected behavior of the rotating system. Moreover, identifying at least one correlation model based on the deviation in the behavior of the rotating system. Further, determining the anomaly in the health condition of the rotating system using the correlation model and one or more parameter values indicative of the deviation in the behavior of the rotating system.

[0017] According to another aspect, the method further comprises generating a notification on a graphical user interface indicating the anomaly in the health condition of the rotating system. In one aspect, a representative view of the anomaly can be presented on the graphical user interface. The representative view of the anomaly includes a real time representation of the health condition of the internal component associated with the anomaly. In a further aspect, the real time representation of the internal component associated with the anomaly is a color coded representation of the internal component in conjunction with the anomaly in the health condition of the rotating system. In addition to the representative view, an audio alert can also be generated to indicate the presence of the anomaly.

[0018] Advantageously, the present invention facilitates real time notification of a specific anomaly related to an internal component associated with the rotating system to a maintenance personnel. This enables the maintenance personnel to perform maintenance in a timely manner with less downtime of the rotating system.

[0019] In one aspect of the present invention, the method further comprises determining a root cause associated with the determined anomaly in the health condition of the rotating system and predicting one or more preventive maintenance actions to address the root cause associated with the determined anomaly in the health condition of the rotating system.

[0020] Advantageously, the present invention enables determination of an anomaly in the health condition of the rotating system and a root cause of the anomaly without disassembling the rotating system.

[0021] In one aspect of the application, the method further comprises predicting an impact of the anomaly in the rotating system on performance of a facility, wherein the rotating system is part of the facility. In one example, the impact can be associated with an operational quality of the facility. In another example, the impact can be associated with possible damage to a system operatively coupled to the rotating system.

[0022] In one aspect of the application, the method further comprises determining a remaining useful life of an internal component of the rotating system based on the anomaly in the health of the rotating system. In a further aspect, the method further comprises predicting a remaining life of the rotating system based on the remaining useful life of the internal component of the rotating system.

[0023] Advantageously, the present application facilitates determining the remaining useful life of the rotating system without manually testing the rotating system. Thus, the disclosed method provides more accurate results as compared to manual testing methods. Further, the disclosed method is also faster as compared to manual testing methods.

[0024] In one aspect of the application, the method further comprises scheduling a preventive maintenance activity by basing the impact of the anomaly in the health of the rotating system on performance of the facility to optimize a downtime of the facility.

[0025] Advantageously, the present application facilitates automatically scheduling a preventive maintenance activity without requiring any manual input. Further, a human operator can plan other activities in accordance with the scheduled preventive maintenance activity. Further, the optimization of the downtime helps increase the productivity of the facility.

[0026] According to another aspect, the present application discloses an apparatus for managing a health of a rotating system. The apparatus comprises one or more processing units, and a memory unit communicatively coupled to the one or more processing units. The memory unit comprises a health monitoring module stored in the form of machine-readable instructions executable by the one or more processing units. The health monitoring module is configured to perform the method steps described above. The execution of the health monitoring module can also be performed using a co-processor such as a graphics processing unit (GPU), a field-programmable gate array (FPGA), or a neural processing / computing engine.

[0027] According to an aspect of the present invention, the apparatus can be an edge computing device. As used herein, "edge computing" refers to a computing environment that can be executed on an edge device (e.g., which is connected at one end to a sensing unit in an industrial setting, and at another end to remote server(s), such as computing server(s) or cloud computing server(s)), which can be a compact computing device having a small form factor and resource constraints in terms of computing power. The apparatus can also be implemented using a network of edge computing devices. The network of edge computing devices can be referred to as a fog network.

[0028] In another aspect, the apparatus is a cloud computing system having a cloud computing based platform configured to provide cloud services for analyzing operational data. As used herein, "cloud computing" refers to a processing environment that includes configurable computing physical and logical resources (e.g., networks, servers, storage, applications, services, etc.), and data distributed over a network (e.g., the Internet). The cloud computing system provides on-demand network access to a shared pool of configurable computing physical and logical resources. The network is, for example, a wired network, a wireless network, a communication network, or a network formed by any combination of these networks.

[0029] Further, the present invention is a system comprising one or more sensing units including a sensing unit capable of providing operational data associated with a rotating system. The system further comprises an apparatus as described above, which is communicatively coupled to the one or more sensing units. The apparatus is configured to manage a health condition of the rotating system based on the operational data.

[0030] According to another aspect, the present invention is a computer program product having machine readable instructions stored therein, which when executed by a processor, cause the processor to perform a method as described above.

[0031] The above attributes, features, and advantages of the present invention and the manner of attaining them will become apparent, and the present invention will be better understood by a

[0032] The present invention is explained in detail below by taking a squirrel cage induction motor (hereinafter referred to as "motor") as an example of a rotating system.

[0033] The present invention is further described below with reference to the illustrated embodiments shown in the drawings, in which:

[0034] Figure 1An environment for a device for managing health of an electric machine is illustrated in accordance with one embodiment of the present application;

[0035] Figure 2 A test setup for calibrating a virtual replica of an electric machine is illustrated in accordance with one example embodiment of the present application;

[0036] Figure 3 A flowchart of a method for determining correlation of rotor displacement in an electric machine with shell vibration response of the electric machine is illustrated in accordance with one example embodiment of the present application;

[0037] Figure 4 Plots of simulated rotor displacement for angular misalignment, parallel misalignment, and no misalignment are illustrated;

[0038] Figure 5 Shell vibration response of an electric machine obtained from operational data for parallel misalignment is shown;

[0039] Figure 6 Shell vibration response of an electric machine obtained from operational data for angular misalignment is shown;

[0040] Figure 7 A system for managing health of a fleet of electric machines is illustrated in accordance with one embodiment of the present application; and

[0041] Figure 8 A flowchart of a method for managing health of an electric machine is illustrated in accordance with one embodiment of the present application.

[0042] In the following, embodiments for carrying out the present application are described in detail. Various embodiments are described with reference to the drawings, wherein like reference numerals are used throughout to designate like elements. In the following description, for the purposes of explanation, numerous specific details are set forth in order to provide a thorough understanding of one or more embodiments. It can be apparent, however, that such embodiments, can be practiced without

[0043] Reference is made to Figure 1 , an environment 100 for a device 105 for managing health of an electric machine 110 is illustrated in accordance with one embodiment of the present application. In one example, the environment can be a facility, where a facility is a part of the facility. In the present embodiment, the electric machine 110 is a squirrel cage induction machine.

[0044] The motor 110 is associated with one or more sensing units 115. The one or more sensing units 115 can include sensors operatively coupled to the motor 110 for measuring values of parameters corresponding to the operation of the motor 110 in real time. In one example, the one or more sensing units 115 can measure values associated with vibrations at different parts of the motor 110, such as the housing, shaft, bearings, and the like. In another example, the one or more sensing units 115 can measure values of stray magnetic flux generated by the motor 110. The one or more sensing units 115 are further communicatively coupled to the controller 120.

[0045] The controller 120 includes a transceiver 125, one or more first processing units 130, and a first memory 135. The transceiver 125 is configured to connect the controller 120 to a network interface 140. In one embodiment, the controller 120 receives the operational data from the one or more sensing units 115 and transmits the operational data to the device 105 through the network interface 140.

[0046] The device 105 includes a communication unit 145, one or more second processing units 150, a display 155, a graphical user interface (GUI) 160, and a second memory 165, which are communicatively coupled to each other. In a preferred embodiment, the communication unit 145 includes a transmitter (not shown), a receiver (not shown), and a gigabit Ethernet port (not shown). The second memory 165 can include a stacked 2 gigabyte random access memory (RAM) package on package (PoP), and a flash storage device. The one or more second processing units 150 are configured to execute computer program instructions defined in the modules. Further, the one or more second processing units 150 are also configured to execute instructions in the second memory 165 concurrently. The display 155 includes a high-definition multimedia interface (HDMI) display and a cooling fan (not shown). Additionally, a control personnel can access the device 105 through the GUI 160. The GUI 160 can include a web-based interface, a web-based downloadable application interface, and the like.

[0047] In one embodiment, the device 105 is configured on a cloud computing platform (not shown). The cloud computing platform can be implemented as a service for analyzing data.

[0048] The second memory 165 includes a plurality of modules: a calibration module 170, a correlation module 175, a simulation module 180, and a diagnostic module 185, which are hereinafter collectively referred to as health monitoring modules 190. The following description explains the functions of the modules when executed by the one or more second processing units 150.

[0049] The calibration module 170 calibrates the virtual replica of the motor 110 to replicate the substantially similar response of the motor 110 in real-time in the simulated scenario. The virtual replica can be based on metadata associated with the motor 110, historical data associated with the motor 110, and a model of the motor 110. The metadata can include current rating of the motor 110, housing material of the motor 110, hysteresis coefficient of different parts of the motor 110, thermal coefficient of different parts of the motor 110, and the like. The historical data can include historical information related to the performance, maintenance, and health condition of the motor 110. The model of the motor 110 can include physics-based model such as magnetic flux model, vibration model, or a combination thereof.

[0050] In particular, the calibration module 170 calibrates the virtual replica by updating the model to accurately represent the motor 110. In other words, the virtual replica is calibrated to ensure a certain degree of fidelity with the motor 110. The calibration of the virtual replica substantially involves adjustment of parameters associated with the model to accurately represent the response of the motor 110 for a given operating condition (e.g., load). In one embodiment, the virtual replica is calibrated using machine learning techniques including, but not limited to, supervised learning techniques, unsupervised learning techniques, and reinforcement learning techniques. The machine learning techniques can use stochastic simulation for calibrating the virtual replica based on the virtual replica and operating data obtained from the motor 110. In one example, the virtual replica can be calibrated using Bayesian calibration techniques. With the calibration, the response of the virtual replica such as at time t = 10 seconds can represent the response of the motor 110 at time t = 10 seconds under the same operating condition. The operating condition can be defined by the simulation instance. In another embodiment, the virtual replica is calibrated using artificial intelligence (AI) based techniques. In one example, the AI based techniques can involve deep learning. Reference is made to Figure 2 The process of calibrating the virtual replica is explained in detail using examples with reference to

[0051] The correlation module 175 is configured to determine one or more correlations between parameter values associated with the motor 110 and one or more anomalies in the motor 110. In one example, the one or more correlations are pre-determined by the correlation module 175 using correlation analysis techniques. The correlation analysis techniques can involve machine learning techniques, non-linear regression techniques, linear regression techniques, and the like. The determined one or more correlations are further used to generate a correlation model. An example method for determining the correlation between the housing vibration and the rotor displacement (i.e., displacement of the rotor) has been explained with reference to Figure 3 , 4 , 5 and 6.

[0052] The simulation module 180 is configured to provide a simulation platform of the virtual replica. The simulation platform enables the virtual replica to simulate real-time behavior of the electric machine 110 based on the operation data received in real-time. In the present embodiment, the operation data obtained from the one or more sensing units 115 is used to simulate the behavior of the electric machine 110 using the virtual replica. To simulate the behavior of the electric machine 110, first, the virtual replica is configured based on the operation data. More specifically, the virtual replica is updated based on the simulation instance associated with the electric machine 110 using the operation data. The simulation instance can define the operating or environmental conditions that can affect the behavior of the electric machine 110. After the update, the virtual replica, when executed on the simulation platform, can simulate the behavior of the electric machine 110 in real-time. Thus, in the preferred embodiment, during the simulation, the operation data is fed to the metamodel of the electric machine 110 in the virtual replica. In the case of simulation, the simulation module 180 generates simulation results.

[0053] The diagnostic module 185 is configured to analyze the simulation results generated by the simulation module 180. More specifically, the simulation results are analyzed with respect to the expected behavior of the electric machine 110 to determine the deviation in the behavior of the electric machine 110. For example, the deviation in the behavior of the electric machine 110 can be associated with the rotor displacement in the electric machine 110. The expected behavior of the electric machine 110 can be determined based on the predefined conditions stored in the second memory 165. For example, the predefined conditions can include rotor displacement of less than an upper threshold of 0.05 mm. If the actual value of the rotor displacement in the electric machine 110 is 0.08 mm, then the deviation can be 0.03 mm. Based on the deviation in the behavior of the electric machine 110, one or more correlation models are identified from among the correlation models generated by the correlation module 175. Further, the diagnostic module 185 determines the anomaly in the electric machine 110 using the identified one or more correlation models. For example, the anomaly can be one of rotor parallel misalignment, rotor angular misalignment, air gap asymmetry, and the like.

[0054] On determining the anomaly, the diagnostic module 185 can determine the root cause associated with the determined anomaly in the health condition of the rotating system. For example, the root cause associated with the rotor angular misalignment can be due to the deformation of the electric machine 110 due to improper coupling with the drive system. On determining the root cause, the diagnostic module 185 can predict one or more preventive maintenance actions to address the root cause associated with the determined anomaly in the health condition of the electric machine 110.

[0055] The diagnostic module 185 can further predict the impact of the abnormality in the health of the electric machine 110 on the performance of the facility. For example, the abnormality in the health of the electric machine 110 can affect the system driven by the electric machine 110, which can eventually lead to a failure of the drive system. The diagnostic module 185 further optimizes the downtime of the facility by scheduling preventive maintenance activities based on the impact of the abnormality in the health of the rotating system on the performance of the facility.

[0056] The diagnostic module 185 can further determine the remaining useful life of the internal component corresponding to the abnormality in the health of the electric machine 110. For example, in case of the rotor angular misalignment, the internal component can be the rotor, the bearing coupled to the rotor, and the like. Further, based on the remaining useful life of the internal component, the remaining life of the electric machine 110 can be predicted. Further, the diagnostic module can further generate a recommendation for preventive maintenance action to increase the remaining life of the electric machine 110.

[0057] The diagnostic module 185 further generates a notification on the GUI 160 indicating the abnormality in the health of the electric machine 110. The notification can include a representative view of the abnormality. The representative view of the abnormality can include a real-time representation of the health of the internal component associated with the abnormality. For example, when the abnormality is the rotor angular misalignment, a color-coded representation of the misaligned rotor's path of motion inside the air gap of the electric machine 110 and the resulting change in the magnetic flux distribution within the electric machine 110 can be shown. Further, the internal component corresponding to the abnormality, i.e., the rotor, can be highlighted in the color-coded representation. The notification can further include information associated with the impact of the abnormality in the electric machine 110 on the performance of the facility, the remaining useful life of the internal component, and the remaining life of the electric machine 110. The representative view can further include a proposed schedule of preventive maintenance activities and the corresponding downtime of the facility. Further, a human operator can interact with the GUI 160 to understand the nature of the abnormality and the required preventive maintenance action. In another example, the notification can further include an audio alert.

[0058] In another embodiment, the controller 120 performs the functions of the apparatus 105. The first memory 135 of the controller 120 can include modules similar to the health monitoring module 190.

[0059] In yet another embodiment, the apparatus 105 can take the form of a device that can be deployed on or near the electric machine 110. Further, the device can be communicatively coupled to a display device including a GUI (similar to the GUI 160). The display device can be located at a remote location, thereby enabling a human operator to remotely monitor the health of the electric machine 110.

[0060] REFERENCES Figure 2FIG. 1 illustrates a test setup 100 for calibrating a virtual replica of an electric machine 105 to represent real-time operation of the electric machine 105, in accordance with one example embodiment of the present application. More specifically, the setup 100 is used to calibrate a virtual replica of the electric machine 105 to accurately represent the electric machine 105 in real-time along with anomalies, if any. The virtual replica resides on a device 107 (similar to the device 105). In this example embodiment, the anomalies are associated with a rotor of the electric machine 105. The anomalies can include, but are not limited to, rotor imbalance, rotor parallel misalignment, rotor angular misalignment, rotor displacement, and broken rotor bars.

[0061] The setup 100 includes a first sensing unit 110, a second sensing unit 115, and a third sensing unit 120 to measure operating parameters associated with the electric machine 105. The first sensing unit 110 includes radial and axial vibration sensors mounted on a housing of the electric machine 105 for measuring vibrations in the housing. The radial and / or axial vibration sensors can be implemented using one of an accelerometer, a velocity meter, a displacement meter, and a non-contact sensor such as an eddy current sensor. The second sensing unit 115 includes a non-contact sensor, for example, an eddy current sensor, to measure shaft vibrations. The non-contact sensor can be positioned directly above a shaft of the electric machine 105. The second sensing unit 115 can further include a displacement amplifier to amplify the shaft vibrations. The third sensing unit 120 includes flux probes mounted on the housing of the electric machine 105. The third sensing unit 120 further includes a gauss meter communicatively and electrically coupled to the flux probes. The gauss meter along with the flux probes is configured to measure stray flux associated with the electric machine 105. In other words, the gauss meter is configured to perform a flux leakage test on the electric machine 105. Additionally, the output of each of the first sensing unit 110, the second sensing unit 115, and the third sensing unit 120 can be further processed by respective signal conditioning units. The operating data from the first sensing unit 110, the second sensing unit 115, and the third sensing unit 120 is further sent to the device 107. The device 107 further simulates behavior of the electric machine 105 using the operating data based on the virtual replica. The virtual replica includes metadata, historical data, and models associated with the electric machine 105.

[0062] In one embodiment, the model is a three-dimensional magnetic flux model, for example, a finite element (FE) model, of the electric machine 205. The magnetic flux model can initially represent the magnetic flux distribution in the electric machine 205 based on the magnetization characteristics of different components, such as the rotor, stator, bearing, housing, etc., and the contact conditions defined for each component. Further, the magnetic flux model can be used to simulate the magnetic flux distribution in different regions within the electric machine 205 for different load conditions and for different types of anomalies. Further, the magnetic flux model is validated based on the stray magnetic flux measured by the third sensing unit 220. More specifically, the magnetic flux model is validated by comparing the measured value of the stray magnetic flux density with the simulated value of the stray magnetic flux density. Based on the result of the validation, the magnetic flux model can be calibrated with respect to the stray magnetic flux to a predetermined accuracy. It must be understood that the stray magnetic flux is a parameter that can be measured externally using the third sensing unit 220. However, the measurement of the magnetic flux density at the critical regions within the electric machine 205, for example, at the rotor, stator coils, etc., is difficult. Therefore, it is very cumbersome to validate the simulated values of the magnetic flux density at the critical regions. Therefore, by applying a pre-defined correction factor corresponding to each critical region to the calibrated stray magnetic flux, the virtual replica can be calibrated with respect to the magnetic flux density at the critical regions within the electric machine 205. Similarly, other parameters associated with the magnetic flux model can also be calibrated. Further, the parameters can be calibrated continuously as described above in order to replicate the near real-time performance of the electric machine 205 using the virtual replica.

[0063] In another embodiment, the model can be a two-dimensional vibration model, for example, a finite element model. In one example, the two-dimensional vibration model can be based on a spring-mass-damper model of the electric machine 205. The spring-mass-damper model can include bearings modeled as springs with equivalent stiffness, fan cover modeled as a lumped mass, etc. The two-dimensional vibration model can be used to simulate the casing vibration and shaft vibration in the electric machine 205. Similar to the case of the magnetic flux model, the values of the simulated casing vibration and simulated shaft vibration are verified by comparing with the values of the casing vibration received from the first sensing unit 210 and the shaft vibration received from the second sensing unit 215, respectively. Based on the verification, the two-dimensional vibration model can be calibrated to a predetermined accuracy with respect to the casing vibration and shaft vibration. In a preferred embodiment, the external vibration response of the virtual replica (e.g., the casing vibration response, i.e., the vibration response associated with the casing) is calibrated using machine learning techniques. Further, the internal vibration response of the virtual replica can be calibrated based on the external vibration response by applying a correction factor. In one example, the internal vibration response can be the vibration response associated with the rotor (or rotor vibration response). The rotor vibration response thus calculated can be further used to simulate the rotor displacement using the virtual replica. For example, the rotor displacement can be simulated based on a predefined criterion that the rotor displaces by 0.05 mm during vibration.

[0064] Reference is made to Figure 3 in conjunction with Figure 2 and Figure 4 , 5 and 6, a flowchart of a method 300 for determining the correlation of rotor displacement with casing vibration response is shown, in accordance with one example embodiment of the present application. Figure 4 Simulated rotor displacements (in millimeters) with respect to multiple stations 1, 2...13 on the rotor are shown for rotor angular misalignment, rotor parallel misalignment, and rotor no misalignment. Further, Figure 5 and Figure 6 actual casing vibration responses of the electric machine 205 are shown for parallel misalignment and angular misalignment, respectively. The method 300 includes steps 305-325.

[0065] At step 305, rotor displacement is determined based on simulation using the calibrated virtual replica. Stations 1, 2...13 represent measurement points equally distributed along the length of the rotor, with station 7 representing the center measurement point on the rotor. More specifically, Figure 4 The simulation results of

[0066] At step 310, the casing vibration response is obtained based on the measured values of casing vibration from the first sensing unit 210. For the present example, Figure 5 and Figure 6 show the casing vibration response from test results for parallel misalignment and angular misalignment, respectively. The casing vibration response is obtained by applying a Fast Fourier Transform (FFT) to the measured values of casing vibration at a sampling frequency SF. The minimum frequency Δf of the FFT can be given by:

[0067] Δf = SF / N = 1 / (N. Δt)

[0068] where N is the number of FFT points and Δt is the minimum time step. The minimum time step Δt is the inverse of the minimum frequency Δf. When sampling the vibration values, the minimum frequency Δf is set such that the line frequency does not merge with the minimum frequency at any point. Based on the results of the FFT, the first 505, second 510, third 515 and fourth 520 harmonics of the casing vibration response for rotor parallel misalignment are determined, as shown in Figure 5 Similarly, the first 605, second 610, third 615 and fourth 620 harmonics of the casing vibration response for rotor angular misalignment are also determined based on the results of the FFT, as shown in Figure 6

[0069] Referring to Figure 4 and Figure 5 , for parallel misalignment, the simulated rotor displacement at station 7 is 0.013 mm and the corresponding test results show that the peak of the fourth harmonic 520 has an amplitude of 0.05 m / s 2 . Similarly, referring to Figure 4 and Figure 6 , for angular misalignment, the simulated rotor displacement at station 7 is 0.023 mm and the corresponding test results show that the peak of the fourth harmonic 620 has an amplitude of 0.1 m / s 2 . In other words, when the rotor displacement is increased by a factor of 2, the amplitude of the fourth harmonic 620 is also increased by a factor of 2. Thus, there is a strong correlation between the simulated rotor displacement and the peak amplitude of the fourth harmonic 620 (or 520) of the casing vibration response. This correlation is a direct result of the rotor displacement and the resulting air gap asymmetry that contributes to the casing vibration. At step 315, a correlation analysis technique is used to establish the correlation between the simulated rotor displacement and the peak of the fourth harmonic 620 (or 520). It is also possible to identify a correlation between the air gap flux density and the fourth harmonic 620 (520) when the rotor displacement causes a change in the air gap flux density. In another implementation, it is also possible to determine the correlation between the simulated rotor displacement and the measured air gap flux density.​

[0070] At step 320, a correlation model is generated for the correlation established in step 315 to determine rotor displacement based on the peak value of the fourth harmonic 620 (or 520) of the casing vibration response obtained in real-time from the electric machine 205. Similarly, a correlation model can be generated to determine anomalies indicated by other harmonics of the casing vibration response. For example, the peak value of the first harmonic 505 (or 605) can be related to imbalance in the rotor, the peak value of the third harmonic 515 (or 615) can be related to misalignment in the rotor, and so on.

[0071] As can be appreciated, the air gap asymmetry caused by the rotor displacement results in a change in the magnetic flux in the electric machine 205. The virtual replica can also determine the change in the magnetic flux at key locations inside the electric machine 205 based on the rotor displacement, as indicated by step 325. In one example, as explained above, a correlation model can be generated for mapping the casing vibration response to the magnetic flux density at the key locations inside the electric machine 205.

[0072] Reference Figure 7 In accordance with one embodiment of the present application, a system 700 for managing health of a fleet of electric machines 705-1, 705-2... 705-n (collectively, fleet 705) is shown. The system 700 includes a device 710 (similar to device 105).

[0073] The device 710 is communicatively coupled to one or more sensing units (not shown) associated with each electric machine of the fleet 705. Each of the one or more sensing units measures values associated with operating parameters of each electric machine of the fleet 705. The device 710 is further communicatively coupled to a server 715 through a network interface 720. The server 715 is further communicatively coupled to a display device 745 including a GUI.

[0074] The server 715 includes a communication unit 726, one or more processing units 728, and a memory 730. The memory 730 is configured to store computer program instructions defined by modules, for example, a health monitoring module 735 (similar to health monitoring module 190). Further, the memory 730 can also store at least one virtual replica corresponding to at least one electric machine of the fleet 705.

[0075] In one embodiment, the server 715 can also be implemented in a cloud computing environment where computing resources are delivered as a service over the network interface 720. As used herein, a "cloud computing environment" refers to a processing environment that includes configurable computing physical and logical resources (e.g., networks, servers, storage, applications, services, etc.) and data distributed through a network interface 720 (e.g., the Internet). A cloud computing environment provides on-demand network access to a shared pool of configurable computing physical and logical resources. The network interface 720 is, for example, a wired network, a wireless network, a communications network, or a network formed from any combination of these networks.

[0076] In the present embodiment, the device 710 acts as an edge device for collecting operational data associated with each electric machine in the fleet 705. The correlation between the operational data for each electric machine in the fleet 705 and anomalies in the health condition can vary. The variation in the correlation can be due to performance degradation, changes in the operating environment associated with each electric machine, defect tolerance of each electric machine, load profile of each electric machine, etc. In one example, each electric machine in the fleet 705 can be associated with a variability factor. The variability factor can account for the variations associated with the electric machine.

[0077] The memory 730 can further include a fleet calibration module 735 to calibrate a virtual replica corresponding to the fleet 705. The virtual replica can be calibrated using artificial intelligence (AI) based techniques. For example, the AI based techniques can dynamically calibrate the virtual replica of each motor using a deep learning based AI model based on the real-time received operational data and variability factors, if any. The AI model can be trained based on historical data associated with each motor to calibrate the virtual replica. The health monitoring module 740 determines one or more anomalies in each motor of the fleet 705 further based on the calibrated virtual replica. Further, a representative view of the one or more anomalies is presented on a GUI of the display device 745. For example, the representative view can indicate rotor displacement in one or more motors of the fleet 705, and magnetic flux variations in the one or more motors caused by the rotor displacement using color coding. The representative view can include a two-dimensional representation of the anomaly. In another example, the representative view can include a three-dimensional representation. Further, the anomaly in the health condition of a motor, such as the motor 705-1, can be further used to determine a remaining useful life associated with an internal component corresponding to the anomaly in the motor 705-1, a remaining life of the motor 705-1, an impact of the anomaly on performance of a facility in which the fleet 705 is a part, and the like. The health monitoring module 740 can also determine an impact of the anomaly in the motor 705-1 on other motors of the fleet 705. Further, the health monitoring module 740 can also schedule a preventive maintenance action for the fleet 705 based on the anomalies detected in each motor of the fleet 705. The health monitoring module 740 can also determine an optimized downtime of the facility based on the scheduled preventive maintenance activities.

[0078] Referring to Figure 8 A flow diagram of a method 800 for managing health condition of a motor, in accordance with one embodiment of the present application is shown. In one example, the method 800 can be implemented on the device 105. The method includes steps 805-830.

[0079] At step 805, operational data associated with the motor 110 is received in real-time from one or more sensing units 115. The operational data includes parameter values corresponding to the operation of the motor 110.

[0080] At step 810, a virtual replica of the motor is configured by the simulation module 180 using the operational data.

[0081] At step 815, the simulation module 180 generates simulation results by simulating behavior of the motor 110 on at least one simulated instance of the motor based on the configured virtual replica of the motor. The simulation results indicate behavior of the rotating system.

[0082] At step 820, the diagnostic module 185 determines an abnormality in the health of the electric machine based on the analysis of the simulation results. The abnormality in the health corresponds to a health state of at least one internal component of the electric machine 110.

[0083] At step 825, a notification is generated by the diagnostic module 185 to indicate the abnormality in the health of the electric machine on the GUI 160.

[0084] The present application can take the form of a computer program product which includes a program module accessible from a computer-usable or computer-readable medium storing program codes designed and constructed for use with or in conjunction with one or more computers, processors, or instruction execution systems. For the purpose of this description, a computer-usable or computer-readable medium is any apparatus that can contain, store, communicate, propagate, or transport the programs for use by or in connection with the instruction execution system, apparatus, or device. The medium can be electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system (or apparatus or device) or a propagation medium of these kinds, as well as any suitable combination of the foregoing, as the signal bearer is not included in the definition of the physical computer-readable medium, which includes semiconductor or solid-state memory, magnetic tape, removable computer diskette, random access memory (RAM), read-only memory (ROM), rigid magnetic disk, and optical disk such as a compact disk read-only memory (CD-ROM), compact disk read / write, and DVD. Both the processor and the program codes for implementing each aspect of the technology can be centralized or distributed (or a combination thereof), as known to those skilled in the art.

[0085] While the present application has been illustrated and described in detail in the preferred embodiments, the application is not limited to the disclosed examples. Other variations can be deduced by those skilled in the art without departing from the scope of the claimed application.

Claims

1. A computer-implemented method of managing health of at least one rotating system (110), the method comprising: receiving, by a processing unit, in real-time, from one or more sensing units (115), operational data associated with the rotating system, wherein the operational data comprises parameter values corresponding to operation of the rotating system (110); configuring a virtual replica of the rotating system (110) using the operational data, wherein the virtual replica is calibrated to accurately represent real-time response of the rotating system; generating simulation results by simulating behavior of the rotating system (110) on at least one simulation instance of the rotating system (110) based on the configured virtual replica of the rotating system (110), wherein the simulation results are indicative of the behavior of the rotating system (110); identifying at least one correlation model from a plurality of correlation models based on deviation in the behavior of the rotating system (110), wherein each correlation model determines one or more correlations between parameter values associated with the rotating system and one or more anomalies; determining an anomaly in the health of the rotating system (110) based on analysis of the simulation results; and generating a notification indicative of the anomaly in the health of the rotating system (110), wherein the simulation results are analyzed with respect to expected behavior of the rotating system (110) to determine the deviation in the behavior of the rotating system (110); and wherein the anomaly is determined in real-time using the at least one correlation model corresponding to the deviation in the behavior of the rotating system (110), wherein the at least one identified correlation model and parameter values indicative of the deviation are used to determine the anomaly; wherein the anomaly corresponds to a health of at least one internal component of the rotating system (110).

2. The method of claim 1, wherein the virtual replica is a virtual representation of the rotating system (110).

3. The method of claim 1, wherein using the operational data to configure a virtual replica of the rotating system (110) comprises: updating the virtual replica of the rotating system (110) based on the operational data in real-time using the simulation instance.

4. The method of claim 1, wherein determining an abnormality in a health of the rotating system (110) comprises: identifying the at least one correlation model from the plurality of correlation models.

5. The method of claim 1, wherein generating the notification indicative of the anomaly in the health of the rotating system (110) comprises: presenting a representative view of the anomaly on a graphical user interface (160), wherein the representative view of the anomaly comprises a real-time representation of the health of an internal component associated with the anomaly.

6. The method of claim 5, wherein the real-time representation of the internal component associated with the anomaly is a color-coded representation of the internal component in conjunction with the anomaly in the health of the rotating system (110).

7. An apparatus (105) for managing health of a rotating system, the apparatus (105) comprising: one or more processing units (150); and a memory (155) coupled to the one or more processing units (150), wherein the memory (155) stores instructions that, when executed by the one or more processing units (150), cause the one or more processing units (150) to perform operations comprising: receiving, in real-time, from one or more sensing units (115), operational data associated with the rotating system, wherein the operational data comprises parameter values corresponding to operation of the rotating system (110); configuring a virtual replica of the rotating system (110) using the operational data, wherein the virtual replica is calibrated to accurately represent real-time response of the rotating system; generating simulation results by simulating behavior of the rotating system (110) on at least one simulation instance of the rotating system (110) based on the configured virtual replica of the rotating system (110), wherein the simulation results are indicative of the behavior of the rotating system (110); identifying at least one correlation model from a plurality of correlation models based on deviation in the behavior of the rotating system (110), wherein each correlation model determines one or more correlations between parameter values associated with the rotating system and one or more anomalies; determining an anomaly in the health of the rotating system (110) based on analysis of the simulation results; and generating a notification indicative of the anomaly in the health of the rotating system (110), wherein the simulation results are analyzed with respect to expected behavior of the rotating system (110) to determine the deviation in the behavior of the rotating system (110); and wherein the anomaly is determined in real-time using the at least one correlation model corresponding to the deviation in the behavior of the rotating system (110), wherein the at least one identified correlation model and parameter values indicative of the deviation are used to determine the anomaly; wherein the anomaly corresponds to a health of at least one internal component of the rotating system (110). a memory unit (165) communicatively coupled to the one or more processing units (150), wherein the memory unit (165) comprises a health monitoring module (190) stored in the form of machine-readable instructions executable by the one or more processing units (150), wherein the health monitoring module (190) is configured to perform the method steps of any one of claims 1 to 6.

8. A system comprising: one or more sensing units (115) capable of providing operational data associated with a rotating system (110); and the apparatus (105) of claim 7 communicatively coupled to the one or more sensing units (115), wherein the apparatus (105) is configured to manage the health of the rotating system (110) based on operational data according to any one of the method claims 1 to 6.

9. A computer program product having machine-readable instructions stored therein, which when executed by one or more processing units, cause the processing units to perform the method of any one of claims 1 to 6.

Images