Analyzing the quality of force response data detected during testing and mitigating poor quality output

Abstract

A system and method for automated MIMO force response characterization of a structure under test. The structure under test is coupled to a plurality of exciter devices and a plurality of response sensors. An excitation signal is automatically and repetitively applied to each exciter device while sensor data is collected from each response sensor to collect response data for a plurality of different exciter-sensor combinations (i.e., response data collected by a single response sensor while the excitation signal is applied to the single exciter device). A signal quality test is applied to the collected sensor data, and in response to determining that the collected response data for a particular exciter-sensor combination is of insufficient quality, data collection for that exciter-sensor combination is automatically repeated. The excitation signal may be automatically adjusted prior to repeating data collection to improve the quality of the collected data for the exciter-sensor combination.

Description

[Technical Field]

[0001] Background technology The present invention relates to a system and method for testing the response of a device or structure to an applied excitation (eg, vibration) condition. Summary of the Invention [Means for solving the problem]

[0002] In one embodiment, the present invention provides a method for automated MIMO force-response characterization of a structure under test. The structure under test is coupled to a plurality of exciter devices and a plurality of response sensors. An excitation signal is automatically and repeatedly applied to each of the plurality of exciter devices, causing each exciter device (e.g., one exciter device at a time) to apply an excitation force to the structure under test. Sensor data is collected from each response sensor while the excitation signal is repeatedly applied to the exciter devices. The collected sensor data includes response data collected by a single response sensor while an excitation force is applied by a single exciter device to each of a plurality of different exciter-sensor combinations. A signal quality test is applied to the collected sensor data, and in response to determining that the collected response data for a particular exciter-sensor combination does not meet a defined signal quality condition, data collection for that exciter-sensor combination is automatically repeated. In some embodiments, the excitation signal applied to the exciter devices is adjusted before repeating the data collection to improve the quality of the collected data.

[0003] Other aspects of the invention will become apparent by consideration of the detailed description and accompanying drawings. [Brief explanation of the drawings]

[0004] [Figure 1A] FIG. 1 is a perspective view of a vibration testing system according to one implementation. [Figure 1B] FIG. 1B is a perspective view of an exciter device of the vibration testing system of FIG. 1A. [Figure 1C] FIG. 1C is a cross-sectional view of the exciter device of FIG. 1B. [Figure 2] 1B is a graph of the frequency response function measured by the system of FIG. 1A, expressed as a frequency-dependent accelerance function. [Figure 3] FIG. 1 is a schematic diagram of the mathematical relationship between multiple-input multiple-output (MIMO) system response function (SRF) measurements and the corresponding MIMO SRF matrix. [Figure 4] FIG. 1B is a block diagram of a control system for the test system of FIG. 1A configured to perform MIMO system vibration testing by sequentially performing single-input multiple-output (SIMO) SRF data acquisition and processing. [Figure 5] FIG. 1B is a block diagram of a control system for the test system of FIG. 1A configured to perform MIMO system vibration testing by sequentially performing single-input, single-output (SISO) SRF data acquisition and processing. [Figure 6] FIG. 6 is a block diagram of a control unit for the control system of FIGS. 4 and / or 5. [Figure 7] 7 is a flowchart of a method for performing SISO SRF data collection and processing using the control system of FIG. 5 and the control unit of FIG. 6. [Figure 8] 8 is a flowchart of a method for analyzing the quality of data collected during the execution of the method of FIG. 7. [Figure 9] 9 is a table listing examples of data quality tests that may be performed as part of the method of FIG. 8. [Figure 10] 9 is a graph of SRF reciprocity data collected during quality testing of the method of FIG. 8. [Figure 11] 9 is a series of graphs of response consistency data collected during quality testing of the method of FIG. 8. [Figure 12] 9 is a graph of Interface Integrity Criteria (ICC) data collected during quality testing of the method of FIG. 8. [Figure 13A] 9 is a schematic diagram of an example of adjusting the applied excitation signal during the method of FIG. 7 based on the results of a quality test of the method of FIG. 8. [Figure 13B] 9 is a graph of the applied excitation signal before and after adjustment of the excitation signal based on the results of a quality test of the method of FIG. 8. [Figure 13C] 10 is a graph of the force spectrum of an applied excitation signal before and after adjusting the excitation signal. [Figure 13D] 10 is a graph of the response spectrum before and after adjusting the excitation signal. DETAILED DESCRIPTION OF THE INVENTION

[0005] Before describing any embodiments of the invention in detail, it is to be understood that the invention is not limited in its application to the details of construction and the arrangement of components set forth in the following description or illustrated in the following drawings. The invention is capable of other embodiments and of being practiced or carried out in various ways.

[0006] Excitation tests can be performed on a device or structure (e.g., an automotive part) to model how the structure responds to different applied forces (e.g., sound, vibration, shock, etc.). In some implementations, the test can include applying an excitation force to the device or structure at a first location and measuring the response at a second location. For example, a vibration force can be applied at a first location and acceleration measured at another location. The difference between the applied excitation force and the measured response indicates the system response. In other implementations, the force excitation and response measurement can be juxtaposed (e.g., an exciter device attached to an accelerometer, which is connected to the structure under test (e.g., see sensor 107A in Figure 4 below)) to capture driving point SRF measurements.

[0007] To approach structure-borne sound and vibration problems in their full complexity (e.g., multipath and multi-degree-of-freedom (DoF) systems), characterization of the structural dynamics of machines is crucial. While feasible for simple structures, accurate numerical modeling of structural and / or vibroacoustic properties remains challenging for most complex technological components and assemblies. Instead, in some implementations, their dynamic behavior can be experimentally characterized by system response function (SRF) measurements, such as those used during modal testing. The nature of the system response function (SRF) can be quite comprehensive and can include structural, acoustic, vibroacoustic, or other descriptions of the structure's propagation characteristics (e.g., hydraulic (fluid) pressure response to an applied force excitation). Furthermore, the SRF may be expressed in the time domain, modal domain, state-space domain, physical domain, or frequency domain. One specific example of a method for determining the SRF is the measurement of frequency response functions (FRFs), such as compliance, mobility, or acceleration.

[0008] Experimentally, FRFs can be determined by employing some type of force excitation at a given input degree of freedom (DoF) and measuring the resulting system response at one or more spatial observer DoFs. For example, an instrumented hammer (modal hammer) and / or vibration shaker can be used as an excitation source to provide external force excitation to the structure, and kinematic sensors (e.g., displacement probes, velocity sensors, and / or accelerometers) are used to capture at least one system response per excitation test. Other types of sensors may also be used to account for other types of system responses, such as sound pressure, as in the case of vibro-acoustic FRF measurements.

[0009] For simple FRF tests (e.g., when the equipment setup and complexity are relatively simple and the number of SRFs to be measured is small), the experiment can be performed manually, for example, with all measurement equipment installed before the first measurement and all FRFs measured in parallel. Alternatively, in more sophisticated configurations or cases where the availability of measurement equipment is limited, the experiment may be performed in a “roving equipment” manner, in which at least some of the exciters and / or sensors are moved to different positions after each excitation test. Related measurement techniques may be referred to as “roving exciter (hammer, shaker)” or “roving sensor (accelerometer, etc.)” experiments. Roving measurement techniques are advantageous in terms of cost and / or availability of the required equipment. These techniques can significantly reduce the number of exciters and / or sensors required, as well as be able to be performed with low-channel-count DAQ systems. However, some disadvantages of roving measurement tests include the following: These include: (1) increased measurement time and effort as the complexity of the structure under test increases; (2) the introduction of experimental error due to the need to reposition instrumentation between different tests (e.g., consistent positioning of the device, proper mounting / application of force excitation, proper alignment of excitation / response DoF); and (3) increased susceptibility to errors during subsequent manual data processing steps and the lack of automatic synchronization between data collection and data post-processing steps.

[0010] For these reasons, in some implementations, SRF testing of more advanced multiple-input, multiple-output (MIMO) systems is performed in a partially or fully automated manner, with all necessary measurement equipment installed on the test structure before measurements begin. In some implementations, FRF measurements are then performed simultaneously for all response and force excitation DoFs. Simultaneous measurements significantly reduce data collection time; unfortunately, this reduction in data collection time comes at the expense of higher equipment costs, as high-channel-count DAQ systems are required to acquire all force excitations and system responses in parallel to use multi-criteria FRF post-processing methods, as well as the flexibility to excite the structure with arbitrary excitation functions (e.g., swept sine waves, random noise, shocks, etc.).

[0011] In some implementations, such “automated” simultaneous FRF measurement techniques may utilize specifically designed excitation sequences and associated post-processing methods, such as vibration testing using multi-criteria random methods (e.g., burst random excitation) and / or multi-sine sweep excitation. These types of methods may be utilized, for example, in equipment-intensive experimental studies such as aircraft ground vibration testing, where measurement equipment is installed directly on the test structure and dismantled after the test is completed. Therefore, the data collection time for “single-specimen” SRF measurements is relatively short compared to the total test time, which includes measurement and equipment setup time, data collection time, data post-processing steps, and dismantling the measurement equipment after the test. Similar considerations are valid for similarly sophisticated but dimensionally more compact structures, such as automotive components and / or component assemblies (e.g., electric power steering (EPS) systems, transmission gearboxes, engines, etc.). However, in automotive applications, repeated vibration testing on multiple parts may be required to evaluate overall noise, vibration, and harshness (“NVH”) performance for a statistically relevant number of test specimens. These "multi-sample" SRF measurements would greatly benefit from improved SRF measurement techniques that can quickly and efficiently collect high-quality SRF data in a fully automated manner, without the need for labor-intensive instrumentation steps and / or expensive DAQ systems to provide a sufficient number of separate measurement channels.

[0012] FIG. 1A illustrates an example of an SRF test system 100. A fixture 101 includes two arms configured to hold a device / structure under test 103. In the example of FIG. 1A, the fixture 101 is configured to hold an electric power steering system for SRF testing. A plurality of excitation units 105 are configured to controllably apply excitation forces to the device / structure under test 103 and are coupled to the fixture 101, which in turn couples the fixture to the device / structure under test 103, either directly or indirectly. Similarly, a plurality of response sensors 107 (e.g., accelerometers) are also coupled (directly or indirectly) to the device / structure under test.

[0013] As shown in more detail in FIG. 1B , each excitation unit 105 in the example of FIG. 1A includes a vibration exciter 109 (i.e., a “shaker”) and a load cell 111. As shown in FIG. 1C , the vibration exciter 109 is a magnetically driven exciter device including a cylindrical body 113 having a pair of electromagnetic coils 115 coupled to the cylindrical body 113. A piston 117 (e.g., also referred to as a “table”) having a permanent ring magnet 119 attached thereto is at least partially disposed within the cylindrical body 113 without being fixedly coupled to the cylindrical body 113. Thus, a magnetic field controllably applied by the electromagnetic coil 115 interacts with the permanent ring magnet 119 to cause movement of the cylindrical body 113 relative to the piston 117. For example, when one end of piston 117 is coupled to device / structure under test 103 and an alternating current is applied to electromagnetic coil 115, the resulting alternating magnetic field causes a corresponding alternating movement of cylindrical body 113 relative to piston 117, which in turn results in a controllable vibrational force being applied to device / structure under test 103 via piston 117. The amplitude and frequency of the applied vibrational force can be adjusted and regulated by controlling the amplitude and frequency of the current applied to electromagnetic coil 115, which in turn controls the amplitude and frequency of the magnetic field applied by electromagnetic coil 115 to permanent ring magnet 119. While the example of Figures 1A-1C shows the use of vibrational exciter 109, in some implementations other types of exciter devices may be used instead of or in addition to vibrational exciter 109. The exciter device may include, for example, a mechanical or electromechanical device configured to apply a vibratory, pulsating, reciprocating, transient, or other dynamically varying force, such as a speaker, a rotating eccentric mass actuator, an electric motor, a solenoid actuator, a piezoelectric actuator, or any other driving mechanism capable of dynamically exciting a physical structure.

[0014] Thus, an exciter device, such as vibration exciter 109, is used to generate a "known" dynamic force input to the device / structure under test 103; as described in more detail below, the force / load generated by the excitation source and applied to the device / structure under test 103 can be used as a reference signal for calculating a system response function (SRF) in a data post-processing step. The actual dynamic force / load applied by the exciter device (e.g., vibration exciter 109) is measured, for example, using a load cell 111 coupled between the exciter device and the device / structure under test 103 (as shown in FIG. 1B). In other implementations, other mechanisms for measuring the applied excitation force may be used, including, for example, one or more force transducers embedded within the excitation source. Other techniques and methods may be used to determine the actual dynamic force excitation imparted to the device / structure under test 103, including, for example, inverse measurement techniques (e.g., inverse force synthesis), model-based force reconstruction methods (e.g., analytical, numerical, experimental, and / or hybrid approaches), or measurements of other load-indicating quantities (e.g., current, voltage, strain, etc.) mapped (e.g., mathematically) to force / load excitation space. In the examples described herein, the term "load" is a generalization of the term "force" to account for other types of dynamic excitation that may be applied to the device / structure under test 103, such as sound pressure (in the case of an acoustic exciter).

[0015] The response sensor 107 is configured to sense / determine the actual response of the device / structure under test 103 to dynamic force excitation applied by one or more exciter devices (e.g., vibration exciter 109). In some implementations, the response sensor 107 can be thought of as a dynamic transducer mechanically coupled to the device / structure under test. The response sensor 107 measures how the device / structure under test 103 (and any test bench components coupled thereto, e.g., fixture 101) responds to the excitation signal. The response sensor 107 provides signals and / or data representing the output response to a connected data acquisition system (as described further below). Depending on the type of system response function (SRF) to be determined, different types of sensors may be used, including, for example, an accelerometer and / or a microphone. In the example of FIG. 1A, the response sensor 107 includes an accelerometer. In some other implementations, the test system 100 may be configured to measure / determine the response of the device / structure under test 103 without mechanically coupled sensors, for example, by using non-contact sensors such as laser-based measurement devices. In some implementations, the response of the device / structure under test may be determined at some locations without physical instrumentation, for example, by using system equivalent model mixing.

[0016] As shown in Figure 2, in some implementations, the frequency response function between two locations (e.g., the location of the vibration exciter 109 and the location of the response sensor 107) can be defined as the "accelerance," or in other words, the ratio between the frequency domain representation of the measured acceleration response and the frequency domain representation of the applied dynamic force. Figure 2 shows the frequency response function as both magnitude and phase as a function of frequency (i.e., a frequency-dependent accelerance function (A / F)).

[0017] As further shown in Figure 3, by calculating the acceleration between different combinations of input (i.e., applied dynamic excitation force) and output (i.e., response acceleration), several different frequency response functions can be calculated for the device / structure under test. The schematic diagram on the left side of Figure 3 shows a simplified example of a plate-like structure with multiple input forces (F) and multiple output responses (A). Each input and output provides another degree of freedom. To characterize the structure's ability to propagate structure-borne energy between the input and output degrees of freedom (DoF), a complex network of SRFs is measured / determined. In this example, the test structure is assumed to be linear and time-invariant; therefore, SRF characterization can be performed simultaneously. Alternatively, SRF characterization can be performed as a series of subsequent measurements, applying excitation one input DoF at a time and observing the associated dynamic response. In some implementations (as described in more detail below), the associated dynamic response of each input DoF is also measured one at a time (SISO), while in other implementations, multiple output DoFs are measured simultaneously (SIMO).

[0018] Figure 3 also shows an example of how the complete MIMO SRF matrix can be reconstructed from individual SISO or SIMO SRF experiments. In a SISO experiment, every element of the MIMO matrix corresponds to a separate measurement, whereas in a SIMO experiment, an entire column of the MIMO SRF matrix can be read simultaneously. Thus, no additional assumptions are made regarding the force excitations utilized to achieve the appropriate reference signals necessary to construct the mathematical SRF ratios defined in complex space.

[0019] 4 and 5 illustrate examples of systems configured to perform fully automated system response function (SRF) measurements, including, for example, structural and vibro-acoustic frequency response functions (FRFs) for advanced multiple-input, multiple-output (MIMO) vibration problems. In particular, these systems are configured to perform advanced MIMO SRF measurements by sequentially performing single-input, multiple-output (SIMO) data acquisition (as shown in FIG. 4) and single-input, single-output (SISO) SRF data acquisition (as shown in FIG. 5). Post-processing routines are then applied to the collected data to reconstruct the relevant data to obtain the complete MIMO SRF measurement, for example, by loading it into the MIMO SRF matrix shown in FIG. 3.

[0020] In both examples (FIGS. 4 and 5), the test system includes a single-channel (SISO) signal generator 501 configured to provide continuous, transient, and / or other types of dynamic signals. The SISO signal generator 501 is coupled to a SISO power amplifier 503 configured to output an appropriately conditioned driver signal. The driver signal is then routed to each vibration exciter 109, one at a time, by a SIMO exciter router 505. The SIMO exciter router 505 includes a controllable switching device and / or other control logic arrangement configured to selectively couple the output from the SISO power amplifier 503 to a different one of the vibration exciters 109 during each vibration test in the series, while electronically isolating other unused vibration exciters 109 to ensure optimal signal / data conditioning and quality. As mentioned above, each vibration exciter 109 in this example is coupled to a load cell 111 to determine the actual dynamic load imparted to the device / structure under test 103 by the “operational” vibration exciter 109.

[0021] All of the multiple load cells 111 are communicatively coupled to a MISO load router 507, which is configured to electronically and / or digitally connect the "operational" load cells 111 (i.e., the load cells 111 currently coupled to the "operational" vibration exciter 109) to a single input channel of the data acquisition system 509 / 603 and electronically isolate other unused load cells 111. The MISO load router 507 helps ensure optimal signal / data conditioning and quality while reducing the number of data acquisition input channels required for the data acquisition system 509 / 603.

[0022] During SRF testing using the system of FIG. 4 or FIG. 5, an operating vibration exciter 109 generates a dynamic force such that vibrations are imparted to the device / structure under test 103 and any structure connected thereto (e.g., fixture 101). The provided structure-borne energy is then propagated through (and radiated by) different paths in the coupled device / structure under test 103, causing a kinematic response at each of multiple response locations (each location providing an additional response DoF for the SRF experiment). The kinematic response at each different location is measured by a corresponding one of the multiple response sensors 107. In the example of FIG. 4, each response sensor 107 of the multiple sensors is coupled to a different input channel of the data acquisition system 509; therefore, the outputs of the multiple response sensors 107 are measured in parallel with measurements from a single load cell 111 during each data collection step. However, in the example of FIG. 5, each response sensor 107 is coupled to the MISO sensor router 601. The MISO sensor router 601 is configured to electronically and / or digitally (if a numerical method is used to determine the response) connect a single responding sensor 107 to the data acquisition system 603 and electronically isolate other unused responding sensors 107. Thus, multiple responding sensors 107 can be selectively and controllably coupled to the same input channel of the data acquisition system 603, thereby reducing the number of required input channels. In other implementations, multiple MISO sensor routers 601 may be used to simultaneously measure responses from a subset of the responding sensors 107 that are being used. In such cases, the data acquisition system 603 may include multiple channels for measuring a subset of the sensors 107 in parallel, but it is not necessary for the data acquisition system 603 to have enough input channels to simultaneously measure all sensors 107.

[0023] As shown in Figure 6, controller 701 includes an electronic processor 703 and one or more non-transitory computer-readable memories 705. Memory 705 stores data (e.g., data collected from sensors during an SRF experiment) and computer-executable instructions that are accessed and executed by electronic processor 703 to provide functionality of controller 701, including, for example, functionality described herein. Controller 701 is communicatively coupled to data acquisition systems 509 / 603 of the systems of Figures 4 and 5, respectively, and configured to receive response signal data collected during the SRF experiment. Controller 701 is also communicatively coupled to SISO signal generator 501 and configured to generate control signals that define and / or modify the signal generated by SISO signal generator 501. The controller 701 is also communicatively coupled to the SIMO exciter router 505 and the MISO load router 507 and configured to send control signals to the routers 505, 507 to selectively control which vibration exciters 109 are coupled to receive excitation signals and which corresponding load cells 111 are coupled to input channels of the data acquisition system 509 / 603. Similarly, in the system of Figure 5, the controller 701 is coupled to the MISO sensor router 601 and configured to send control signals to the MISO sensor router 601 to selectively control which response sensors 107 are coupled to input channels of the data acquisition system 509 / 603.

[0024] 7 illustrates a method executed by a controller 701 to operate the system of FIG. 5 to perform a multiple-input, multiple-output (MIMO) system vibration test by sequentially executing multiple single-input, single-output (SISO) data acquisition and processing routines. The controller 701 causes the SISO signal generator 501 to generate an excitation signal (step 801). The controller then causes the SISO exciter router 505 to couple the excitation signal to the first vibration exciter 109 and the MISO load router 507 to couple the first load cell 111 to a load cell input channel of the data acquisition system 603 (step 803). Similarly, the controller causes the MISO sensor router 601 to couple the first response sensor 107 to a response input channel of the data acquisition system 603 (step 805).

[0025] Once the first vibration exciter 109 is coupled to receive the excitation signal and the first load cell 111 and first response sensor 107 are coupled to their respective input channels of the data acquisition system 603, the controller 701 begins collecting and storing data from the data acquisition system (step 807). In some implementations, the controller 701 is configured to use a trigger mechanism to wait until one or more predetermined conditions are met before recording the received data. Triggers may also be used to control data capture based on detection of specific events in the applied or sensed vibration or other type of signal. Examples may include exceeding a specific force level when applying excitation with an impact hammer or shaker during modal testing, or exceeding or receiving a specific vibration level.

[0026] In some implementations, the data acquisition system 603 is configured to utilize a trigger mechanism such that data collection and processing does not begin (or stop) until some signal level (e.g., voltage) is detected within an input channel. After the trigger is armed, the controller 701 is initialized to wait for a signal event (e.g., the rising flank of the trigger signal exceeding 50% of the maximum channel range) to occur before acquiring / storing data and / or to continue storing data received from the data acquisition system 603 until a predetermined stop condition is met. In some implementations, the stop condition is defined as a duration (e.g., the controller 701 stores 5 seconds' worth of data from when the signal event is detected) or other stop trigger criteria (e.g., the falling flank of the trigger signal falls below 50% of the maximum channel range). In some implementations, the trigger can be set to automatically re-arm after each trigger so that several measurements can be performed one after the other (e.g., to automate the aggregation and storage of multiple measurements).

[0027] In some implementations, the controller 701 is configured to utilize a trigger that is controlled based on the ability of the device / structure under test 103 to vibrate in response to the applied vibration. For example, it may take the device / structure under test 103 some time to respond linearly enough to the applied excitation. Therefore, starting data acquisition at the same time that the excitation signal is applied to the vibration exciter 109 will not provide high-quality SRF data. Instead, the trigger start event may be configured to delay data acquisition relative to the excitation start time, thus giving the device / structure under test 103 more time to respond before the measurement begins. The same applies to triggering the stop of the measurement. If the measurement is stopped at the same time that the vibration exciter 109 is disconnected from the excitation signal, the structure may still not be able to respond to the latest excitation signal. Instead, in some implementations, the controller 701 may be configured to stop the measurement while the excitation signal is still applied to the vibration exciter 109. This stop condition trigger also prevents nonlinear behavior of the vibration exciter 109 during abrupt start or stop from completely corrupting the measurement. In some implementations, synchronization triggering may be performed automatically by control logic based on excitation type, frequency range, or other test-specific criteria.

[0028] 7, after the data from the coupled response sensors is read and stored in memory 705, the controller 701 increments the MISO sensor router 601 (step 805) to couple the next response sensor 107 to the response input channel of the data acquisition system 603. Thus, this process of collecting response sensor data (step 807) and incrementing the MISO sensor router 601 (step 805) is repeated until sensor data has been captured from each response sensor 107 while the excitation signal is applied by the same individual vibration exciter 109 (step 813). After sensor data has been collected by all response sensors 107 in the sequence (step 813), the controller 701 increments the SIMO exciter router 505 and the MISO load router 507 (step 803), causing the excitation signal to be applied to a different vibration exciter 109. In this example, as the SIMO exciter router 505 and the MISO load router 507 are incremented to apply excitation signals to different vibration exciters 109, the increment of the MISO sensor router 601 is reset to capture sensor data from the first sensor in the sequence. This process continues until excitation signals have been applied to all vibration exciters 109 and sensor data has been recorded from each responding sensor 107 while the excitation signal is applied to the vibration exciter 109.

[0029] For example, consider a configuration in which three vibration exciters 109 and three response sensors 107 are coupled to a device / structure under test 103. The controller 701 applies an excitation signal to a first vibration exciter 109 and collects sensor data from each of the three response sensors 107. The controller 701 then applies an excitation signal to a second vibration exciter 109 and again collects sensor data from each of the three response sensors 107. Finally, the controller 701 applies an excitation signal to a third vibration exciter 109 and again collects sensor data from each of the three response sensors 107. Thus, the controller 701 captures nine different sensor data sets, one for each different possible combination of vibration exciter 109 and response sensor 107.

[0030] Once sensor data has been recorded for all response sensors 107 in the sequence (step 813) while the excitation signal is applied to the last vibration exciter 109 in the sequence (step 819), one or more post-processing routines are applied to the collected data (step 821) to reconstruct and / or update associated MIMO SRF models / measurements indicative of the dynamic behavior of the entire device / structure under test 103. For example, in some implementations, the controller 701 is configured to generate a MIMO SRF matrix using frequency response functions (FRFs) for each different combination of vibration exciter 109 / load cell 111 and response sensor 107, as shown in FIG. 3 above.

[0031] In some implementations, to ensure consistently high data quality, the data acquisition system (DAQ), in cooperation with the controller 701, receives appropriate parameterization before each individual measurement (or, in some implementations, before each group or sequence of measurements). The DAQ itself can be configured to provide advanced functionality (formulas / algorithms) for appropriately adjusting the individual transducers (e.g., load cells 111 and response sensors 107) connected to each DAQ measurement channel. In some implementations, the DAQ provides the ability to automatically adjust range settings based on the “strength” of the incoming measurement signal. This process may be referred to as “auto-ranging” and, in some implementations, is fully automated by the DAQ, occurring immediately before a measurement is taken. This may be done, for example, by operating one or more exciter devices 109 under conditions that produce the highest vibration levels for a period of time (e.g., a few seconds) while detecting and adjusting the connected sensor signals to avoid undesired overloads (clipping). The purpose of a parameterization procedure such as “auto-ranging” is to maximize the sensitivity of the measurement chain (transducers, cables, DAQ) without adversely affecting data quality. For example, in some implementations, the DAQ system is configured to measure a ±5V input signal (discretized at 24 bits) corresponding to a vibration amplitude (e.g., 10 m / s^2). When the vibration amplitude is smaller (e.g., 1 m / s^2), the effective range observed by the DAQ decreases to ±1V (auto-ranging). In this case, the 24-bit resolution remains equivalent for higher measurement accuracy (e.g., a clearer distinction between the signal and the sensitivity threshold of the measurement device). In some implementations, the DAQ features a communication and / or control interface that enables communication with other types of devices via a compatible protocol, such as a serial bus, Ethernet, USB, or other communication interface.

[0032] In some implementations, the controller 701 is also configured to analyze data received from the load cell 111 and / or the response sensor 107 to determine whether the captured data meets certain data quality conditions. In response to determining that the captured data for a particular response sensor 107, load cell 111, or various combinations thereof does not meet the data quality conditions, the controller 701 is configured to repeat data collection for one or more combinations of vibration exciter 109 and response sensor 107. In some implementations, the controller 701 may repeat data collection by applying the same excitation signal to the vibration exciter 109 and collecting data from the response sensor 107. In some implementations, the controller 701 may be configured to adjust the excitation signal applied to the vibration exciter 109 in addition to or instead of repeating data collection with the same excitation signal previously applied.

[0033] For example, as shown, the controller 701 may be configured to apply one or more quality tests (step 809) to the collected sensor data after each sensor reading step (step 807). As described in more detail below, in some implementations, the quality tests may be designed to analyze data for individual vibration exciter 109 / response sensor 107 combinations, while in other implementations, the quality tests may be designed to analyze data for a vibration exciter 109 / response sensor 107 combination against sensor data collected for other vibration exciter 109 / response sensor 107 combinations.

[0034] 7 , in some implementations, in addition to or instead of applying one or more quality tests after each response sensor reading (step 809), the controller 701 is also configured to apply one or more quality tests (step 815) after collecting data from all sensors in a sequence to which an excitation signal was applied by the same vibration exciter 109. In other words, the one or more quality tests (step 815) are applied before incrementing the SIMO exciter router 505 and MISO load router 507 (step 803) to apply an excitation signal to the next vibration exciter 109 and connect the corresponding load cell 111 to the data acquisition system 603. Again, in some implementations, in response to determining that the collected data fails one or more of the quality tests (step 815), the controller 701 is configured to adjust the excitation signal applied to the vibration exciter 109 (step 817) and repeat data collection from one or more of the response sensors in sequence before proceeding to the next vibration exciter 109.

[0035] Furthermore, in some implementations, the controller 701 may be configured to apply one or more quality tests to the collected data after collecting all of the SRF data (i.e., after collecting data from all vibration exciter 109 / response sensor 107 combinations), in addition to or instead of the quality checks performed after each individual sensor reading and after each complete sequence of sensor readings. Furthermore, in some implementations, as described above, the controller 701 is configured to send control signals to the SIMO exciter router 505, the MISO load router 507, and the MISO sensor router 601 to selectively operate any particular combination of vibration exciter 109 and response sensor 107. Thus, in some implementations, the controller 701 is configured to analyze the data captured after collecting data for all vibration exciter 109 / response sensor 107 combinations, identify one or more combinations that do not meet data quality conditions, and repeat data collection only for those combinations identified as having insufficient data quality. In some implementations, data collection for these insufficient combinations is performed by adjusting the excitation signal, operating the SIMO exciter router 505 to couple the adjusted excitation signal to the identified vibration exciters 109 of the combinations with missing data, operating the MISO sensor router 601 to couple the identified response sensors 107 of the combinations with missing data to response input channels of the data acquisition system 603, and then collecting data from the coupled response sensors 107. This process is then repeated for each combination identified as having insufficient data quality.

[0036] As described above, in some implementations, multiple data quality checks may be applied to the collected data simultaneously (i.e., after data is collected from an individual sensor, after data is collected from the entire sequence of sensors, and / or after data is collected from all vibration exciter 109 / response sensor 107 combinations). FIG. 8 illustrates an example of a method performed by the controller 701 for applying multiple different quality tests to the collected data. After input data is received (e.g., from one or more load cells 111 and / or multiple response sensors 107) (step 901), the controller 701 applies a first signal quality test (step 903). If the collected data does not pass the first signal quality test (step 905), a first adjustment is applied to the excitation signal (step 907) and data acquisition is repeated (step 909). However, if the collected data passes the first signal quality test (step 905), the controller 701 proceeds to apply a second signal quality test (step 911). Again, if the collected data does not pass the second signal quality test (step 913), a second adjustment is applied to the excitation signal (step 915), and data acquisition is repeated (step 909). However, if the collected data passes the second signal quality test (step 913), controller 701 proceeds to apply a third signal quality test (step 917). If the collected data does not pass the third signal quality test (step 919), a third adjustment is applied to the excitation signal (step 921), and data acquisition is repeated (step 909). However, if the collected data passes all three signal quality tests, controller 701 proceeds to the next data acquisition step, or if all data has been collected, controller 701 proceeds to apply applicable post-processing routines to the collected data (e.g., to load into a MIMO SRF matrix) (step 923).

[0037] In some implementations, the controller 701 is configured to perform each test sequentially, such that the controller 701 proceeds to the second signal quality test only if the collected data passes the first signal quality test, and similarly, proceeds to the third signal quality test only if the collected data passes both the first and second signal quality tests. However, in other implementations, the controller is configured to perform each test in parallel, such that the controller 701 proceeds to the second signal quality test regardless of whether the collected data passes the first signal quality test (as shown by the dashed line between steps 907 and 911 in FIG. 8 ), and similarly, proceeds to the third signal quality test regardless of whether the collected data passes the first and / or second signal quality tests (as shown by the dashed line between steps 915 and 917 in FIG. 8 ). In some implementations, the controller 701 is configured to apply multiple signal quality tests as a combination of serial and parallel tests. For example, in some implementations, the controller 701 may be configured to perform both the first signal quality test (step 903) and the second signal quality test (step 911) on the collected data regardless of whether the collected data passes the first signal quality test or the second signal quality test, but to perform the third signal quality test (step 917) only if the collected data passes the second signal quality test (or, in some implementations, only if the collected data passes both the first signal quality test and the second signal quality test).

[0038] Finally, in some implementations, the controller 701 may be configured to apply the same type of signal conditioning regardless of which signal quality test fails (e.g., the first excitation signal conditioning (step 907), the second excitation signal conditioning (step 915), and the third excitation signal conditioning (step 921) are the same type of signal conditioning). However, in other implementations, the controller 701 is configured to apply different types of signal conditioning to the excitation signal depending on which signal quality test fails. For example, the controller 701 may be configured to change the excitation signal from a random signal to a logarithmic sweep signal (as described below with reference to FIG. 13B) in response to determining that the collected data fails a first signal quality test, and to apply a low-pass (or high-pass) filter to the excitation signal (as described below with reference to FIG. 13C) in response to determining that the collected data fails a second signal quality test. Thus, in some implementations, in response to determining that the collected data fails both the first and second signal quality tests, the controller 701 is configured to both (1) change the excitation signal to a logarithmic sweep signal and (2) apply a low-pass (or high-pass) filter to the logarithmic sweep excitation signal.

[0039] In various implementations, the data quality test (e.g., the signal quality test in the example of FIG. 8 ) may include data processing and quality assurance steps in which the data quality of each individual SIMO or SISO experiment sequence is verified, each individual SIMO or SISO experiment is converted into some kind of SRF, the individual SRFs may be temporarily arranged into one or more partial MIMO matrices, and / or the data quality of the entire MIMO matrix (or multiple partial MIMO matrices) is verified. In implementations in which the individual SRFs are temporarily arranged into one or more partial MIMO matrices, controller 701 may also be further configured to perform additional data conversion, manipulation, or expansion routines, including numerical, model-based, and / or hybrid techniques. Such techniques may include, for example, geometric reduction and / or transformation techniques (e.g., finite difference approximations and virtual point transformations), data manipulation techniques that numerically modify the boundary conditions under which the SRF experiments were performed (e.g., dynamic substructuring for coupling and uncoupling), data dimensionality reduction and / or regularization techniques (e.g., principal component analysis, truncation, or Tikhonov-regularized singular value decomposition), and augmentation techniques that model additional data at degrees of freedom (e.g., locations on the device / structure under test) for which no physical measurements were performed (e.g., system equivalent model mixing (SEMM) techniques in which experimental data is combined with other measured / simulated data). Some specific examples of metrics and calculations that may be used in various implementations to ensure high-quality experimental data during the automated SRF measurement process are outlined in the table of FIG. 9. However, some implementations include fewer calculations, metrics, and tests than those listed in the table of FIG. 9, and some implementations may include other calculations, metrics, and tests in addition to or instead of those listed in the table of FIG. 9.

[0040] In some implementations, the controller 701 is configured to perform data processing and quality assurance steps using additional data (e.g., data specific to the test setup). Such data may include, for example, geometric relationships between excitation and response DoFs (e.g., Euclidean distances, Euler angles, etc.), experiment-related metadata, and / or metadata required for recording purposes (e.g., data required to populate MIMO matrices).

[0041] In various implementations, the analytical results of the data processing and quality assurance steps applied by controller 701 may or may not be provided to the user as feedback (e.g., via a display screen of user interface 707) during the automated SRF measurement process. In implementations in which quality test feedback is provided to the user, the results may be summarized in a variety of ways, ranging from a single numerical value to sophisticated graphical representations of data quality, to assist the user in interpreting individual and / or overall data quality. Figures 10, 11, and 12 show example mechanisms for displaying data quality feedback information to the user via a display screen of user interface 707.

[0042] 10 shows an example of a graph that may be displayed to a user via a display screen of user interface 707, illustrating the SRF reciprocity between automated measurement locations as a single-valued matrix representation. In the example of FIG. 10, the relative darkness of each displayed square indicates the relative quality of the data (e.g., darker squares represent higher quality data). In some implementations, this analysis is part of the set quality test (FIG. 7, step 815).

[0043] FIG. 11 shows an example series of graphs that may be displayed to a user via the display screen of user interface 707, illustrating measures of response consistency. The top graph shows overall response consistency as a frequency spectrum, the middle graph shows individual specific response consistencies as frequency spectrums, and the bottom graph shows specific response consistencies as single values. In the graphs of FIG. 11, values ​​closer to 100% indicate higher quality data. In some implementations, controller 701 may be configured to display all three graphs of FIG. 11 simultaneously on the screen; in other implementations, controller 701 may be configured to display only one or two of the graphs of FIG. 11 and / or to selectively and interchangeably display different graphs of FIG. 11.

[0044] 12 shows an example of a graph that may be displayed to a user via the display screen of user interface 707, showing the Interface Integrity Criterion (ICC) as a frequency spectrum. Values ​​close to one ("1") indicate frequencies that should be further analyzed if the MIMO system is well described and measurements are valid.

[0045] 8 and 9, in some implementations, the controller 701 is configured to adjust the applied excitation signal to further improve data quality. In some implementations, the controller 701 is configured to attempt to apply an excitation signal that imparts a force input to the device / structure under test such that a desired vibration output is obtained that meets as many of the specified quality conditions as possible. Adjustments to the excitation signal may include, for example, (1) duration (longer measurements can provide clearer (e.g., less noisy) results due to averaging effects, but increase the total data acquisition time), (2) type (e.g., impulse / transient excitation, burst / random excitation, pseudo-random excitation, swept sine wave, stepped sine wave, etc.), (3) magnitude (e.g., amplitude scaling of the entire signal or time / frequency dependent amplitude scaling), (4) frequency content (e.g., frequency dependent filtering / equalization (e.g., amplification of excitation at lower frequencies to obtain sufficient signal-to-noise ratio (SNR) in the elastic coupling assembly)), or (5) a combination of some or all of the above (e.g., transmission loss compensation by transmittance filtering).

[0046] In some implementations, the controller 701 is configured to determine whether the applied excitation signal requires adjustment based on the following information: (1) feedback-free adjustment using an integrated “exciter calibration database,” (2) feedback from a connected data acquisition system (DAQ), and / or (3) feedback from a MIMO data processing system. In some implementations, the “exciter calibration database” used by the controller 701 to perform feedback-free adjustment includes information about the connected vibration exciters along with knowledge of which vibration exciters are currently operating (e.g., a predetermined exciter-specific equalization filter is applied to the excitation signal to linearize the exciter output each time the exciter operates). In some implementations, the DAQ itself may be configured to perform internal data quality checks to analyze the most recent individual SIMO / SISO experiments, and the results are communicated to the controller 701 via an appropriate protocol. Finally, in some implementations, the controller 701 (or a separate MIMO data processing system) is configured to process and analyze the individual SIMO / SISO measurements and / or the entire set of measurements reconstructed as one or more MIMO matrices. As mentioned above, in some implementations, systems such as those shown in Figures 4-6 are configured to address individual and overall data quality / consistency simultaneously.

[0047] 13A-13D illustrate an example of a test system (e.g., the system of FIGS. 4-6 ) configured to adjust an applied force (X) (i.e., the actual vibration applied to a device / structure under test in response to an excitation signal) in response to a sensed vibration response (Y). The test system in this example is configured to apply a signal quality test to evaluate a signal-to-noise ratio (SNR). The example of FIG. 13A includes a separate MIMO data processing system 1301 configured to analyze collected data and control logic 1303 configured to adjust the excitation signal in response to feedback received by control logic 1303 from MIMO data processing system 1301. In some implementations, a single controller (e.g., controller 701) may be configured to provide the functionality of both MIMO data processing system 1301 and control logic 1303, as described in this example. In other implementations, a controller (e.g., controller 701) may be configured to provide the functionality of control logic 1303, and data acquisition system 509 / 603 may be configured to provide the functionality of MIMO data processing system 1301. Other combinations of logical components and distribution of functionality are possible in other implementations.

[0048] As shown in FIG. 13A (and as described in other examples above), SISO signal generator 501 is configured to generate excitation signals in response to control signals received from control logic 1303, and in this example, is configured to normalize the excitation signal output (i.e., excitation signals having amplitudes of ±1) for each individual vibration exciter 109. An example of a normalized signal x(t) generated by SISO signal generator 501 is shown in FIG. 13B. The normalized excitation signal x(t) is then provided as an input to SISO power amplifier 503, which outputs a conditioned excitation signal that is routed to the vibration exciter 109. The vibration exciter 109 applies a vibration excitation force x(t) to the device / structure under test, and the actual vibration is measured by a corresponding load cell 111. An example of the actual applied vibration X measured by load cell 111 is shown in a frequency-domain representation (X(ω)) in FIG. 13C. The applied force is transmitted through the device / structure under test 103, and actual vibrations at other locations on the device / structure under test are routed to data acquisition system 509 / 603, and data inputs indicative of the sensed vibration y(t) are collected by or transmitted to MIMO data processing system 1301. An example of the sensed vibration Y measured by response sensor 107 is shown in FIG. 13D in a frequency domain representation (i.e., response spectrum Y(ω)).

[0049] 13A, MIMO data processing system 1301 is configured to apply one or more signal quality tests to the acquired data, including at least one signal quality test designed to determine whether the collected data exhibits a sufficient signal-to-noise ratio. MIMO data processing system 1301 sends feedback to control logic 1303 indicating the results of the signal quality tests (i.e., that the SNR is sufficient), and based on the received feedback, control logic 1303 determines whether adjustments to the excitation signal are appropriate.

[0050] To demonstrate the adjustment mechanism in the example of FIG. 13A in more detail, FIGS. 13B-13D show various example signals both before and after adjustments are made to the excitation signal by the control logic. In FIG. 13B, the top graph shows the original excitation signal, and the bottom graph shows the adjusted excitation signal. In FIG. 13C, the top graph shows the force spectrum X(ω) of the actual force applied to the device / structure under test (as measured by load cell 111) in response to the original excitation signal, and the bottom graph shows the force spectrum X(ω) of the actual force applied in response to the adjusted excitation signal. In FIG. 13D, the top graph shows the response spectrum Y(ω) based on the output of response sensor 107 when the original excitation signal is applied, and the bottom graph shows the response spectrum Y(ω) based on the output of response sensor 107 when the adjusted excitation signal is applied.

[0051] The original excitation signal (shown in the top graph of FIG. 13B) has a random (broadband white noise) time signature x(t). The applied force in response to this original excitation signal (shown in the top graph of FIG. 13C) is characterized by a moderately broadband frequency spectrum X(ω) with a significant force dropoff in the higher frequency range. In this example, the excitation frequency range spans a low-frequency region of frequencies not relevant to this particular experiment (shown as "unimportant frequency range" in FIG. 13C). However, this low-frequency excitation component may limit the ability of the vibration exciter 109 to provide a high-energy output within more relevant frequency regions (e.g., the higher frequency range where the signal in the top graph of FIG. 13C begins to drop off). In reality, the illustrated dropoff in force toward higher frequencies may be due to internal damping from simultaneously exciting all frequencies (e.g., using a random white noise excitation signal). Thus, the vibration exciter 109 is attempting to overcome large strokes at low frequencies while simultaneously superimposing small stroke motions at higher frequencies. The applied vibration causes the device / structure under test 103 to vibrate linearly and time-invariantly (LTI) in response to the applied force, which is then sensed by the response sensor 107 .

[0052] In the top graph of FIG. 13D , the spectrum of the detected response vibration Y(ω) is plotted along with the spectrum of the sensor noise floor. In some implementations, the sensor noise floor may be measured with the same setup when no excitation is applied to the device / structure under test 103. As discussed above, the “quality” of the detected vibration can be analyzed in a variety of different ways. However, in this example, the top graph of FIG. 13D illustrates a signal-to-noise problem in the region where the detected vibration spectrum Y(ω) is close to the sensor noise floor. In some implementations, a data quality condition may be defined as a detected signal at least 10 dB higher than the sensor noise floor at this frequency. However, in the example of the top graph of FIG. 13D , the detected vibration Y(ω) suffers from poor signal-to-noise performance at the expected anti-resonance of the structure and generally at high frequencies due to the relatively low force excitation (as a result of the high internal damping of the vibration exciter).

[0053] In this example, MIMO data processing system 1301 is configured to analyze the detected vibration to identify insufficient data quality (i.e., insufficient SNR) and send a feedback message to control logic 1303. In response, control logic 1303 initiates automatic adjustments to the applied vibration to mitigate the detected signal quality problem. In some implementations, control logic 1303 is configured to identify appropriate adjustments to the excitation signal based on detailed feedback regarding the frequency ranges where the excitation is insufficient, along with additional metadata provided by the user regarding the target frequency range and / or time requirements for the measurement.

[0054] In this particular example, the control logic 1303 is configured such that, based on feedback from the MIMO data processing system 1301, the control logic determines that the "type" of excitation signal should be changed from a random broadband white noise excitation signal (shown in the top graph of FIG. 13B) to a logarithmic sweep signal (shown in the bottom graph of FIG. 13B). Doing so causes the vibration exciter to generate vibrations at a single frequency at a time, thus preventing the vibration exciter from internally damping while simultaneously maximizing force output. In this example, a logarithmic frequency sweep is used instead of a linear sweep to provide the device / structure under test 103 with sufficient time to respond to the applied vibrations, especially when it is more difficult to excite the lower frequency ranges.

[0055] In this example, comprising control logic 1303, the control logic also determines, based on feedback from MIMO data processing system 1301, that additional adjustments to the excitation frequency range may be appropriate (e.g., to prevent damage to the vibration exciter due to excessive stroke at low frequencies beyond the frequency range of interest and / or to minimize the duration of measurements). In this example, this is achieved by having SISO signal generator 501 apply a high-pass filter during the signal generation process. This high-pass filter prevents the vibration exciter from operating at low frequencies, and as shown in the bottom graph of FIG. 13C, the applied force spectrum X(ω) after adjustments to the excitation signal is optimized to favor the frequency range of interest. Note also that equivalent signal amplification is sufficient to achieve significantly higher exciter force output by operating the exciter at one frequency at a time. Also, in some implementations, a similar filter can be applied as a low-pass filter to block higher frequencies to avoid high shaker temperatures during continuous operation. Other adjustments to the force signal x(t) may include changes in signal amplitude. For example, the sinusoidal signature of the signal in Figure 13B can be amplified or attenuated at certain frequencies to "shape" the force characteristic X(ω) in Figure 13C. In some implementations, the signal generator is configured to apply these adjustments. This allows, for example, to apply higher forces at lower frequencies, which may be useful, for example, for structures coupled using rubber isolators where high excitation energy is required to produce a low-frequency response signal on the structure due to energy dissipation in the rubber isolators.

[0056] The vibration spectrum X(ω) applied in response to the adjusted excitation signal results in better overall excitation of the device / structure under test, as shown by the bottom graph of Figure 13D, which shows a significantly higher vibration response spectrum Y(ω) with increased separation between the sensed signal and the sensor noise floor. Thus, when the MIMO data processing system 1301 receives the updated sensed signal and applies a similar signal quality test, the output of the signal quality test indicates that the signal quality issue (i.e., insufficient SNR) has been sufficiently mitigated and that even the anti-resonances of the device / structure under test (i.e., the lowest measured amplitude levels of the response signal) are clearly separated from the sensor noise floor. The higher-quality response spectrum, along with the measured force spectrum of the applied vibration, can be fully used to determine a high-quality SRF between the corresponding excitation and response degrees of freedom (i.e., the corresponding vibration exciter 109 / response sensor 107 combination).

[0057] 13A-13D, the excitation signal adjustment strategy is fully automated and executed by MIMO data processing system 1301 and control logic 1303 (e.g., in some implementations, both are implemented in controller 701) without direct input from a user. However, in some implementations, control logic 1303 may be configured to output feedback to a user indicative of quantified signal quality and to receive user feedback indicative of adjustments to the excitation signal. In some such implementations, control logic 1303 may be configured to prioritize this direct user feedback over the automatic adjustment strategy.

[0058] Similar excitation signal adjustment strategies may be implemented in some implementations based on received feedback while adjusting the applied vibration, for example, taking into account one or more data quality indicator criteria based on (i) only the applied vibration (e.g., force), (ii) only the sensed vibration (e.g., acceleration), or (iii) both the applied and sensed vibration. In some implementations, the output of the SISO signal generator 501 or the SISO power amplifier 503 is utilized as a basis for adjusting the applied vibration downstream, for example, using numerical data representing a normalized or adjusted desired signal output along with additional performance-related criteria (e.g., the amplifier's frequency response function or the vibration exciter's "output vibration" to "input current" transfer function) to estimate and adjust the expected applied vibration.

[0059] Thus, the systems and methods described herein facilitate decomposing complex experimental MIMO vibration tests into multiple sequentially executed SIMO or SISO SRF experiments according to a fully automated process, as shown in the example of Figure 7. After initial test system setup, including instrumenting the device / structure under test 103 and / or specially designed fixture 101, the controller 701 can autonomously execute all experiments, including data quality checks, complex data-based decision making, and data post-processing. For example, a MIMO SRF vibration test according to one implementation may include the following process control steps: (1) system setup and instrumentation (one-time setup; special fixtures may be used to facilitate measurements); (2) routing of excitation hardware (vibration excitation source and associated load cell, and / or use of additional / alternative methods to determine actual force excitation); (3) routing of response sensor hardware and / or use of additional / alternative methods to determine the actual response of the test structure to the applied force excitation; (4) parameterization of the data acquisition (DAQ) system with the current SIMO or SISO experiment sequence (e.g., sensor calibration, dynamic range, sampling frequency, sampling duration, filtering, triggering, etc.); (5) collection of all SIMO or SISO experiment sequence data; (6) (in some implementations) data post-processing of the current SIMO or SISO experiment sequence, including quality assurance methods and associated process flow control (e.g., repeat experiment with improved excitation strategy or data collection parameterization if quality assurance methods indicate poor results); (7) post-processing (and storage) of successful SIMO or SISO SRF experiment data and initiation of subsequent SIMO or SISO experiments; and (8) reconfigured / updated SIMO or SISO experiment sequence data. Post-processing and provision of SRF measurements (which may involve additional numerical / model-based / hybrid data manipulation / transformations).

[0060] Thus, various implementations of the systems and methods described herein may offer the following advantages over other MIMO testing approaches: (1) one signal generator and power amplifier may be shared by all exciters; (2) the number of excitation and response DoFs is not limited by the number of DAQ channels; (3) any excitation signal (transient, continuous, etc.) may be used while ensuring optimal phase reference and improved signal-to-noise ratio (SNR); (4) an automatic calibration process may be used to adjust the measurement range for each excitation used; and (5) difficult measurements may be split into multiple parts, allowing the exciter to cool down between measurements (data post-processing methods may include fragmentation). (6) Control strategies allow subsequent difficult measurements to be performed using different excitation sources (data post-processing methods can be used to reconstruct / merge different data (e.g., different exciters at different frequency ranges)); (7) reciprocal calibration methods can be used to perform in-situ exciter output calibration and / or reduce the number of required exciter output sensors (e.g., load cells); (8) detection of local nonlinearities using appropriate signal generation and / or post-processing methods; and (9) signal generation independent of the number of SRFs to be determined.

[0061] Other features and advantages of the invention are set forth in the following claims.

Claims

1. 1. A method for automated MIMO force response characterization of a structure under test, comprising: The method comprises: coupling the structure under test to a plurality of exciter devices and a plurality of response sensors; generating an excitation signal; automatically applying the excitation signal repeatedly to each exciter unit of the plurality of exciter units, wherein applying the excitation signal to an exciter unit causes the exciter unit to apply an excitation force to the structure under test; collecting sensor data from each response sensor of the plurality of response sensors while repeatedly applying the excitation signal to the plurality of exciter devices, the collected sensor data including response data for each of a plurality of different exciter-sensor combinations, the response data for each of the plurality of different exciter-sensor combinations including the sensor data collected by a single response sensor while the excitation force is applied to the structure under test by a single exciter device; applying a signal quality test to the collected sensor data to determine if the collected response data for a first exciter-sensor combination of the plurality of different exciter-sensor combinations fails to meet a defined signal quality condition; repeating data collection for the first exciter-sensor combination in response to determining that the collected response data for the first exciter-sensor combination does not satisfy the defined signal quality condition, wherein repeating the data collection for the first exciter-sensor combination comprises applying the excitation signal to a first exciter device of the first exciter-sensor combination and collecting updated sensor data from a first response sensor of the first exciter-sensor combination while the excitation signal is applied to the first exciter device; Including, repeating the data acquisition for the first exciter-sensor combination further includes adjusting the excitation signal; Repeating the data collection for the first exciter-sensor combination includes: applying the adjusted excitation signal to the first exciter device by applying the excitation signal to the first exciter device; collecting the updated sensor data from the first response sensor by collecting the updated sensor data from the first response sensor while the adjusted excitation signal is applied to the first exciter device; Including, applying the signal quality test includes analyzing a frequency domain magnitude of the response data for the first exciter-sensor combination within a defined frequency range; Repeating the data collection for the first exciter-sensor combination includes: repeating the data collection in response to determining that the frequency domain magnitude of the response data for the first exciter-sensor combination within the defined frequency range exceeds a threshold; and adjusting the excitation signal by applying a frequency filter to the excitation signal, the frequency filter comprising at least one selected from the group consisting of a low-pass filter, a high-pass filter, and a band-pass filter, the frequency filter configured to filter out components of the excitation signal corresponding to the defined frequency range used in the signal quality test; A method comprising:

2. repeatedly applying the excitation signal to each exciter unit of the plurality of exciter units includes applying a white noise excitation signal; applying the signal quality test includes analyzing a frequency spectrum signal-to-noise ratio of the response data for the first exciter-sensor combination; Repeating the data collection for the first exciter-sensor combination includes: repeating the data collection for the first exciter-sensor combination in response to determining that the frequency spectrum signal-to-noise ratio of the response data for the first exciter-sensor combination is less than a threshold; and adjusting the excitation signal by changing the type of the excitation signal from the white noise excitation signal to at least one selected from the group consisting of a logarithmic sweep excitation signal and a linear sweep excitation signal; The method of claim 1 , comprising:

3. Adjusting the excitation signal comprises: applying a multi-sine sweep excitation signal; adjusting at least one parameter of the multi-sine sweep excitation signal; The method of claim 1 , comprising at least one selected from the group consisting of:

4. 10. The method of claim 1, wherein applying the signal quality test comprises applying at least one selected from the group consisting of a signal conditioning statistical analysis, a signal-to-noise ratio analysis, a coherence function analysis, a data consistency analysis, and an interface integrity criteria analysis.

5. calculating a frequency response function for each exciter-sensor combination of the plurality of different exciter-sensor combinations; the frequency response function comprises a ratio of the excitation power applied by the exciter device of the exciter-sensor combination to the sensed excitation power sensed by the response sensor of the exciter-sensor combination; Calculating the frequency response function for each exciter-sensor combination of the plurality of different exciter-sensor combinations includes: determining a frequency response function for the first exciter-sensor combination based on the updated sensor data collected by the first response sensor after repeating the data collection for the first exciter-sensor combination; determining a system response function based on a plurality of frequency response functions, including the calculated frequency response function for each exciter-sensor combination; Including, The method of claim 1 , wherein determining the system response function based on the plurality of frequency response functions comprises loading the plurality of frequency response functions into a system response function matrix.

6. applying a set quality test to a set of collected response data to determine if the set of collected response data fails to meet a set quality condition, the set of collected response data including the collected response data for the first exciter-sensor combination and collected response data for a plurality of exciter-sensor combinations collected before the collected response data for the first exciter-sensor combination; repeating data collection for one or more exciter-sensor combinations of the plurality of exciter-sensor combinations in response to determining that the set of collected response data does not satisfy the set quality condition; and The method of claim 1 further comprising:

7. the response data for each of the plurality of different exciter-sensor combinations; acceleration response data collected by the single response sensor while the excitation force is applied to the structure under test by the single exciter device; applied force data collected by a load cell corresponding to the single exciter device, the applied force data indicating the actual excitation force applied to the structure under test by the single exciter device; and The method of claim 1 , comprising:

8. 2. The method of claim 1, wherein collecting the sensor data from each response sensor of the plurality of response sensors while repeatedly applying the excitation signal to the plurality of exciter devices comprises repeatedly coupling each response sensor of the plurality of response sensors to an input channel of a data acquisition system while the excitation signal is applied to the same exciter device of the plurality of exciter devices.

9. 1. A system for automated MIMO force response characterization of a structure under test, comprising: The system comprises: a plurality of exciter devices; a plurality of response sensors; an excitation signal generator; an exciter router configured to selectively couple an excitation signal generated by the excitation signal generator to selected exciter devices of the plurality of exciter devices; an electronic controller, operating the exciter router to automatically apply the excitation signal repeatedly to each exciter unit of the plurality of exciter units, wherein applying the excitation signal to an exciter unit causes the exciter unit to apply an excitation force to the structure under test; collecting sensor data from each response sensor of the plurality of response sensors while repeatedly applying the excitation signal to the plurality of exciter devices, wherein the collected sensor data includes response data for each of a plurality of different exciter-sensor combinations, the response data for each of the plurality of different exciter-sensor combinations including the sensor data collected by a single response sensor while the excitation force is applied to the structure under test by a single exciter device; applying a signal quality test to the collected sensor data to determine if the collected response data for a first exciter-sensor combination of the plurality of different exciter-sensor combinations fails to meet a defined signal quality condition; and in response to determining that the collected response data for the first exciter-sensor combination does not satisfy the defined signal quality condition, repeating data collection for the first exciter-sensor combination, wherein the electronic controller repeats the data collection for the first exciter-sensor combination by applying the excitation signal to a first exciter device of the first exciter-sensor combination and collecting updated sensor data from a first response sensor of the first exciter-sensor combination while the excitation signal is applied to the first exciter device. an electronic controller configured to Equipped with The electronic controller adjusting the excitation signal; applying the adjusted excitation signal to the first exciter device; collecting the updated sensor data from the first response sensor while the conditioned excitation signal is applied to the first exciter device; repeating the data collection for the first exciter-sensor combination by the electronic controller is configured to apply the signal quality test by analyzing a frequency domain magnitude of the response data for the first exciter-sensor combination within a defined frequency range; The electronic controller repeating the data collection in response to determining that the frequency domain magnitude of the response data for the first exciter-sensor combination within the defined frequency range exceeds a threshold; adjusting the excitation signal by applying a frequency filter to the excitation signal, wherein the frequency filter comprises at least one selected from the group consisting of a low-pass filter, a high-pass filter, and a band-pass filter, and the frequency filter is configured to filter out components of the excitation signal corresponding to the defined frequency range used in the signal quality test. and repeating the data collection for the first exciter-sensor combination by:

10. the electronic controller is configured to repeatedly apply a white noise excitation signal to each exciter device of the plurality of exciter devices by applying the excitation signal; the electronic controller is configured to apply the signal quality test by analyzing a frequency spectrum signal-to-noise ratio of the response data for the first exciter-sensor combination; The electronic controller repeating the data collection for the first exciter-sensor combination in response to determining that the frequency spectrum signal-to-noise ratio of the response data for the first exciter-sensor combination is less than a threshold; Adjusting the excitation signal by changing the type of the excitation signal from the white noise excitation signal to at least one selected from the group consisting of a logarithmic sweep excitation signal and a linear sweep excitation signal.

10. The system of claim 9, configured to repeat the data collection for the first exciter-sensor combination by:

11. The electronic controller applying a multi-sine sweep excitation signal; adjusting at least one parameter of the multi-sine sweep excitation signal; 10. The system of claim 9, configured to adjust the excitation signal by at least one selected from the group consisting of:

12. 10. The system of claim 9, wherein the electronic controller is configured to apply the signal quality test by applying at least one selected from the group consisting of a signal conditioning statistical analysis, a signal-to-noise ratio analysis, a coherence function analysis, a data consistency analysis, and an interface integrity criteria analysis.

13. The electronic controller further configured to calculate a frequency response function for each exciter-sensor combination of the plurality of different exciter-sensor combinations; the frequency response function comprises a ratio of the excitation power applied by the exciter device of the exciter-sensor combination to the sensed excitation power sensed by the response sensor of the exciter-sensor combination; The electronic controller calculating a frequency response function for each exciter-sensor combination of the plurality of different exciter-sensor combinations by determining a frequency response function for the first exciter-sensor combination based on the updated sensor data collected by the first response sensor after repeating the data collection for the first exciter-sensor combination; determining a system response function based on a plurality of frequency response functions, including the calculated frequency response function for each exciter-sensor combination; It is configured as follows: The system of claim 9 , wherein the electronic controller is configured to determine the system response function based on the plurality of frequency response functions by loading the plurality of frequency response functions into a system response function matrix.

14. The electronic controller applying a set quality test to the set of collected response data to determine whether the set of collected response data fails to satisfy a set quality condition, wherein the set of collected response data includes the collected response data for the first exciter-sensor combination and collected response data for a plurality of exciter-sensor combinations collected prior to the collected response data for the first exciter-sensor combination; repeating data collection for one or more exciter-sensor combinations of the plurality of exciter-sensor combinations in response to determining that the set of collected response data does not satisfy the set quality condition. The system of claim 9 further configured to:

15. the response data for each of the plurality of different exciter-sensor combinations; acceleration response data collected by the single response sensor while the excitation force is applied to the structure under test by the single exciter device; applied force data collected by a load cell corresponding to the single exciter device, the applied force data indicating the actual excitation force applied to the structure under test by the single exciter device; and The system of claim 9 , comprising:

16. a data acquisition system; a sensor router configured to selectively couple selected ones of the plurality of response sensors to a shared input channel of the data acquisition system; Furthermore, 10. The system of claim 9, wherein the electronic controller is configured to collect the sensor data from each of the plurality of response sensors while repeatedly applying the excitation signal to the plurality of exciter devices by operating the sensor router to repeatedly couple each of the plurality of response sensors to the shared input channel of the data acquisition system while the excitation signal is applied to a same exciter device of the plurality of exciter devices.

Images