Substructure modal identification method, device and equipment capable of being measured step by step and multiple times

Abstract

The invention discloses a substructure modal identification method, device and equipment capable of being measured step by step and multiple times. The method comprises the following steps of: performing signal demodulation processing on response data of each measurement point in each substructure of an engineering structure by using a nonlinear frequency modulation mode decomposition algorithm and taking a minimum objective function as an objective to obtain at least one-order target single-component data; each substructure at least has a common measuring point with one other substructure; obtaining spectrum data based on the target single-component data; obtaining the vibration mode amplitude of each order of each substructure according to the ratio of the same-order frequency spectrum amplitude of each measurement point in the frequency spectrum data; according to the vibration mode amplitude of each order of each substructure, the vibration mode amplitude of the common measurement point of each substructure is adjusted to be consistent, and the vibration mode amplitude of each order of all the measurement points is obtained; according to the vibration mode amplitude of each order of all the measurement points, minimizing and judging the direction of the vibration mode based on a modal confidence criterion; and obtaining the modal shape of the engineering structure according to the direction of the shape.

Description

Technical Field

[0001] The present application belongs to the field of modal identification technology, and in particular, relates to a substructure modal identification method, device, and equipment capable of performing multiple measurements in steps. Background Art

[0002] As a structural dynamics analysis technology, modal identification determines the modal parameters of the structural system, such as damping, frequency, vibration mode, etc., based on the response data of the sensor. This technology is widely used in monitoring, damage diagnosis and optimization design of engineering structures. It can help staff better understand the health status of the structural system and discover potential problems in time, thereby ensuring the safety of the structure.

[0003] Traditional modal identification methods require that all response data must be collected synchronously to ensure the accuracy of modal identification. However, for particularly complex structural systems, this usually requires the deployment of a large number of sensors in the structural system at the same time. This approach not only greatly increases the difficulty of obtaining response data, but also significantly increases the cost of modal identification. Summary of the invention

[0004] The embodiments of the present application provide a substructure modal identification method, device, and equipment that can perform multiple measurements in steps, aiming to solve the problems that the existing modal identification methods not only increase the difficulty of obtaining response data, but also increase the cost of modal identification.

[0005] In a first aspect, an embodiment of the present application provides a substructure modal identification method that can be measured in steps and multiple times, including: obtaining response data of each measurement point in each substructure of an engineering structure; the response data of all measurement points of the same substructure are response data collected synchronously at the same time; the response data of measurement points of different substructures are response data collected in steps and multiple times at different times; each substructure has at least one common measurement point with another substructure; for each response data, using a nonlinear frequency modulation modal decomposition algorithm, with the objective of minimizing an objective function, performing signal demodulation processing on the response data to obtain at least first-order target single component data ; The objective function is used to evaluate the signal complexity of at least one single component data; for each target single component data, the spectrum data is obtained based on the target single component data; for each substructure, the amplitudes of each order vibration mode of each substructure are obtained according to the ratio of the same order spectrum amplitudes of each measurement point in the spectrum data; according to the amplitudes of each order vibration mode of each substructure, the vibration mode amplitudes of the common measurement points of each substructure are adjusted to be consistent, and the amplitudes of each order vibration mode of all measurement points of the engineering structure are obtained; according to the amplitudes of each order vibration mode of all measurement points of the engineering structure, the direction of the vibration mode is determined based on the minimization of the modal confidence criterion; according to the direction of the vibration mode, the modal vibration mode of the engineering structure is obtained.

[0006] In one of the embodiments of the first aspect, the method also includes: determining multiple sets of initial adjustment parameters in a nonlinear frequency modulation mode decomposition algorithm; optimizing the multiple sets of initial adjustment parameters using a parameter-free heuristic optimization algorithm to obtain target adjustment parameters; the target adjustment parameters are used to obtain at least first-order target single component data.

[0007] In one of the embodiments of the first aspect, a plurality of groups of initial adjustment parameters are optimized using a parameter-free heuristic optimization algorithm to obtain target adjustment parameters, including: in a current iteration number, determining the fitness corresponding to each group of initial adjustment parameters based on an objective function; constructing an initial position vector for each group of initial adjustment parameters, and determining a core position vector with the smallest fitness among multiple initial position vectors; updating each initial position vector based on the core position vector to obtain a candidate position vector corresponding to each initial position vector; if a current iteration number does not meet an iteration number condition, updating multiple initial position vectors using multiple candidate position vectors until the current iteration number meets the iteration number condition, and then determining the parameter in the position vector with the smallest fitness among the multiple candidate position vectors as the target adjustment parameter.

[0008] In one of the embodiments of the first aspect, each initial position vector is updated based on the core position vector to obtain a candidate position vector corresponding to each initial position vector, including: taking the actual position difference being less than a preset position difference as a goal, each initial position vector is updated according to the core position vector, a weight value, and a balance index to obtain a candidate position vector corresponding to each initial position vector; wherein the position difference is used to characterize the position deviation rate of the candidate position vector compared to the core position vector, and the balance index is used to characterize the movable range of the initial position vector.

[0009] In one embodiment of the first aspect, with the actual position difference being less than the preset position difference as a goal, the initial position vector is updated according to the core position vector, the weight value and the balance index to obtain a candidate position vector, including: determining an initial value of the balance index within a first index interval and a weight value of each initial position vector in a plurality of initial position vectors; using the core position vector, the initial value and the weight value, updating each initial position vector within a preset search domain to obtain a first position vector corresponding to each initial position vector; for each initial position vector, determining a candidate value of the balance index, and updating the initial value using the candidate value and updating the initial position vector using the first position vector until the candidate value is within a second index interval, and determining the first position vector as the second position vector; wherein the candidate value is less than the initial value, and the first index interval is adjacent to and greater than the second index interval; if each second position vector If the position deviation rates compared to the core position vector are all within the preset deviation range, it is determined that the actual position difference is less than the preset position difference, and the core position vector is updated to a position vector corresponding to the core position vector among multiple second position vectors, and for each second position vector, the second position vector is updated using the core position vector and the weight value to obtain a candidate position vector; if there are one or more second position vectors whose position deviation rates compared to the core position vector are outside the preset deviation range, then for each second position vector, the second position vector is iteratively updated using the core position vector, the weight value and the candidate value until the position deviation rates of each second position vector compared to the core position vector are all within the preset deviation range, and the core position vector is updated to a position vector corresponding to the core position vector among multiple second position vectors, and for each second position vector, the second position vector is updated using the core position vector and the weight value to obtain a candidate position vector.

[0010] In one of the embodiments of the first aspect, multiple initial position vectors are updated using multiple candidate position vectors, including: determining the fitness of each candidate position vector based on an objective function; if there is a candidate position vector with a fitness smaller than the fitness of the candidate position vector corresponding to the core position vector among other candidate position vectors, updating the multiple candidate position vectors to multiple initial position vectors; the other candidate position vectors include position vectors other than the candidate position vector corresponding to the core position vector among the multiple candidate position vectors; if the fitness of the candidate position vector corresponding to the core position vector is greater than or equal to the fitness of other candidate position vectors, updating each candidate position vector based on the Levy flight algorithm until there is a candidate position vector with a fitness smaller than the fitness of the candidate position vector corresponding to the core position vector among other candidate position vectors, and updating the multiple candidate position vectors to multiple initial position vectors.

[0011] In one of the embodiments of the first aspect, the fitness corresponding to each group of initial adjustment parameters is determined based on the objective function, including: for each group of initial adjustment parameters, performing signal demodulation processing on the response data according to a nonlinear frequency modulation mode decomposition algorithm that adopts the initial adjustment parameters to obtain at least one initial single component data; using the objective function to determine the signal complexity of each initial single component data; determining the minimum signal complexity in at least one initial single component data as the fitness corresponding to the initial adjustment parameter.

[0012] In a second aspect, an embodiment of the present application provides a substructure modal identification device capable of performing multiple measurements in steps, comprising:

[0013] The acquisition module is used to acquire the response data of each measurement point in each substructure of the engineering structure; the response data of all measurement points of the same substructure are response data collected synchronously at the same time; the response data of measurement points of different substructures are response data collected multiple times at different times in steps; each substructure has at least one common measurement point with other substructures.

[0014] The decomposition module is used to perform signal demodulation processing on each response data using a nonlinear frequency modulation mode decomposition algorithm with the objective of minimizing the objective function to obtain at least first-order target single component data; the objective function is used to evaluate the signal complexity of at least one single component data.

[0015] The transformation module is used to obtain spectrum data based on each target single component data.

[0016] The identification module is used to obtain the amplitude of each order of vibration mode of each substructure according to the ratio of the amplitude of the same order spectrum of each measurement point in the spectrum data for each substructure.

[0017] The identification module is also used to adjust the vibration mode amplitudes of the common measurement points of each substructure to be consistent according to the vibration mode amplitudes of each substructure, so as to obtain the vibration mode amplitudes of each order of all measurement points of the engineering structure; to determine the direction of the vibration mode based on the minimization of the modal confidence criterion according to the vibration mode amplitudes of each order of all measurement points of the engineering structure; and to obtain the modal vibration mode of the engineering structure according to the direction of the vibration mode.

[0018] In a third aspect, an embodiment of the present application provides a terminal device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the method for substructure modal identification that can be measured in steps and multiple times is implemented as provided in the first aspect above.

[0019] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the substructure modal identification method that can be measured in steps and multiple times as provided in the first aspect above is implemented.

[0020] It can be understood that the beneficial effects of the second aspect mentioned above can be found in the relevant description of the first aspect mentioned above, and will not be repeated here.

[0021] Compared with the prior art, the beneficial effects of the embodiments of the present application are as follows: since the prior art requires synchronous acquisition of the response data of the structure, it is not only necessary to deploy a large number of sensors at the same time, but also necessary to ensure that these sensors can synchronously measure the response data of the structure, which increases the identification cost; at the same time, it also increases the requirements for the data acquisition accuracy of the sensor, which causes the problem of excessive difficulty in acquiring the response data. Therefore, the response data of the measurement points of different substructures acquired by the present application can be response data collected multiple times at different times, so as to greatly reduce the number of sensors that need to be synchronously deployed in large civil engineering structures. The requirements for data synchronization are greatly reduced, thereby reducing the difficulty of acquiring response data. Combined with the nonlinear frequency modulation mode decomposition algorithm, the response data is subjected to signal demodulation processing with the goal of minimizing the objective function, so as to decompose the response data into at least first-order target single component data. In this way, by extracting single-component data of the response data and establishing the relationship between the vibration mode amplitude and the target single-component signal spectrum amplitude, the vibration mode amplitudes of the common measurement points of each substructure are adjusted to be consistent. This no longer limits the synchronization of the collection time of the response data of all measurement points of the engineering structure. It only needs to ensure the synchronization of the data collection time between all measurement points in the same substructure. Even if two sensors are used to divide the engineering structure into multiple substructures containing two measurement points, and the response data of multiple substructures are obtained separately, the accuracy of the recognition of the modal vibration mode of the engineering structure can still be guaranteed, so that the obtained modal vibration mode can accurately reflect the health status of the engineering structure.

[0022] In addition, usually, noise will be dispersed into multiple frequency components. In multi-component data, cross-interference will occur between different components, which will reduce the signal-to-noise ratio of the signal and affect the accurate extraction and analysis of the signal. The single-component data in this application only contains modal information of a specific order, so that the impact of noise will be relatively small. In this way, by performing spectral analysis on each target single-component data, the frequency response of each component can be more finely characterized, thereby effectively reducing the errors caused by signal aliasing or interference. This refined analysis method enables the method to more accurately extract the modal vibration shape of the engineering structure during the modal identification process, which not only helps to improve the accuracy of modal vibration shape identification, but also can greatly reduce the amount of calculation and calculation time, and improve the overall analysis efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 It is a structural schematic diagram of a substructure modal identification system capable of multiple step-by-step measurements provided by an embodiment of the present application;

[0024] Figure 2 It is a flowchart of a substructure modal identification method capable of multiple step-by-step measurements provided by an embodiment of the present application;

[0025] Figure 3a It is a structural schematic diagram of an engineering structure provided by an embodiment of the present application;

[0026] Figure 3b It is a schematic diagram of an example of response data in an engineering structure provided in an embodiment of the present application;

[0027] Figure 3c is a schematic diagram of an example of decomposing response data provided by an embodiment of the present application;

[0028] Figure 3d is a schematic diagram of an example of a time-frequency curve of decomposed response data provided by an embodiment of the present application;

[0029] Figure 3e is a schematic diagram of an example of a wavelet spectrum for decomposing response data provided by an embodiment of the present application;

[0030] Figure 3f is a schematic diagram of an example of a spectrum curve of decomposed response data provided by an embodiment of the present application;

[0031] Figure 4 It is a flowchart of another substructure modal identification method capable of multiple step-by-step measurements provided by an embodiment of the present application;

[0032] Figure 5 It is a flowchart of another substructure modal identification method capable of multiple step-by-step measurements provided by an embodiment of the present application;

[0033] Figure 6 It is a flowchart of another substructure modal identification method capable of multiple step-by-step measurements provided by an embodiment of the present application;

[0034] Figure 7a It is a schematic diagram of an example of a balance indicator in an iterative process provided by an embodiment of the present application;

[0035] Figure 7b It is a schematic diagram of an example of a fitness convergence curve in an iterative process provided by an embodiment of the present application;

[0036] Figure 8 This is a flow chart of optimizing and adjusting parameters provided by an embodiment of the present application;

[0037] Fig. 9 It is a schematic diagram of an example of a 4th-order vibration mode provided in an embodiment of the present application;

[0038] Fig.10 This is a schematic diagram of an example of a 4th-order orthogonality verification provided by an embodiment of the present application;

[0039] Fig.11 It is a structural schematic diagram of a substructure modal identification device capable of multiple step-by-step measurements provided by an embodiment of the present application;

[0040] Fig.12 It is a structural block diagram of a terminal device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0041] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0042] Traditional modal identification methods require that all response data must be collected synchronously to ensure the accuracy of modal identification. However, for particularly complex structural systems, this usually requires the deployment of a large number of sensors in the structural system at the same time. This approach not only greatly increases the difficulty of obtaining response data, but also significantly increases the cost of modal identification.

[0043] Based on this, the embodiment of the present application provides a substructure modal identification method, device, and equipment that can be measured in steps and multiple times. The technical principle of this method is: divide the engineering structure into substructures to obtain multiple substructures containing common measurement points; obtain the response data of each measurement point in each substructure; the response data of all measurement points of the same substructure must be response data collected synchronously at the same time; the response data of measurement points of different substructures can be response data collected in steps and multiple times at different times, and each substructure has at least one common measurement point with other substructures. The nonlinear frequency modulation modal decomposition algorithm is used to minimize the objective function and decompose the response data into at least first-order single-component data. In the decomposition process, a non-parameterized heuristic optimization algorithm is used to optimize the adjustment parameters used in the nonlinear frequency modulation modal algorithm, which can more accurately determine the most suitable adjustment parameters for the nonlinear frequency modulation modal algorithm to improve the accuracy of the single-component data decomposition. After that, the frequency spectrum data of the single component signals in each substructure is calculated by Fourier transforming each order of the single component signal, and then the frequency and amplitude corresponding to the spectrum peak in the spectrum data are obtained. Further, according to the ratio of the amplitude of each order vibration mode of the engineering structure is the ratio of the amplitude of the same order spectrum at each measurement point, the vibration mode amplitude of each substructure is adjusted to be consistent, and the vibration mode amplitude of each order vibration mode at each measurement point and the frequency corresponding to each order single component signal are obtained in turn, thereby obtaining the relative ratio of the amplitude of each order vibration mode of the engineering structure at each measurement point; then, according to the relative ratio of each order vibration mode at each measurement point, the direction of each order vibration mode of the engineering structure is obtained using the modal confidence criterion, that is, first determine that the direction of the first-order vibration mode is all positive, and traverse the positive and negative directions of each measurement point of the second-order vibration mode in turn, calculate the orthogonality of the first-order vibration mode with all possible second-order vibration mode directions, and determine the direction of the second-order vibration mode based on the principle of minimum modal confidence criterion. Similarly, the vibration mode directions of other orders can be obtained. Thus, the modal vibration mode of the structure is obtained. In this way, by extracting single-component data of the response data, the collection time of the response data of multiple measurement points is no longer limited. Even if two sensors are used to obtain the response data of multiple measurement points separately, the accuracy of the identification of the modal vibration shape of the engineering structure can still be guaranteed, so that the obtained modal vibration shape can accurately reflect the health status of the engineering structure.

[0044] Figure 1 The schematic diagram of the structure of a substructure modal identification system capable of multiple step-by-step measurements provided in an embodiment of the present application. The substructure modal identification method capable of multiple step-by-step measurements provided in an embodiment of the present application can be applied to the substructure modal identification system 100 capable of multiple step-by-step measurements. Figure 1 , the modality recognition system 100 includes one or more sensors 101 and a terminal device 102 .

[0045] Optionally, one or more sensors 101 are used to transmit response data of measurement points of a detected structure (eg, a building structure) to the terminal device 102, so that the terminal device 102 performs modal recognition of the structure based on the response data of each measurement point.

[0046] In one possible implementation, a communication connection is established between the sensor 101 and the terminal device 102 via a wired or wireless network. Optionally, the wireless network or wired network uses standard communication technology and / or protocols. The network is typically the Internet, but may be any network, including but not limited to a local area network (LAN), a metropolitan area network (MAN), a wide area network (WAN), a mobile, wired or wireless network, a private network or any combination of a virtual private network.

[0047] Optionally, the sensor 101 is an acceleration sensor, a vibrating string sensor, etc. The acceleration sensor may be, but is not limited to, any of the following: a piezoelectric acceleration sensor, a capacitive acceleration sensor, a laser interferometer acceleration sensor, and a MEMS acceleration sensor. Acceleration sensors have different characteristics and applicability in the field of modal recognition of structures. They can be selected based on factors such as the required measurement range, accuracy, resolution, frequency response range, and environmental conditions. The embodiments of the present application do not limit this.

[0048] Optionally, the terminal device 102 may be a terminal device in various AI application scenarios. For example, the terminal device 102 may be a laptop, a tablet computer, a desktop computer, a vehicle-mounted terminal, a mobile device, etc. The mobile device may be, for example, a smart phone, a portable device, a personal digital assistant, and other types of terminals, which are not specifically limited in the embodiments of the present application.

[0049] Furthermore, the terminal device 102 may run an operating system, which may include but is not limited to Android system, IOS system, Linux system, Unix, Windows system, etc., and may also include a user interface (UI) layer, which can provide a display of modal recognition results to the outside through the UI layer.

[0050] For ease of understanding and explanation, the following Figures 2 to 11 The substructure modal identification, device, and equipment that can be measured multiple times in steps provided by the embodiments of the present application are described in detail.

[0051] Figure 2 FIG. 2 is a flow chart of a substructure modal identification method that can be measured multiple times in steps according to an embodiment of the present application. Figure 2 As shown, the method includes:

[0052] 201. Obtain response data of each measurement point in each substructure of the engineering structure; the response data of all measurement points of the same substructure are response data collected synchronously at the same time; the response data of measurement points of different substructures are response data collected multiple times at different times in steps; each substructure has at least one common measurement point with other substructures.

[0053] Optionally, the engineering structure may be a structural system for which modal vibration shapes need to be identified, and the structural system may be a system composed of various structures, such as a structural system for civil engineering, a structural system for mechanical engineering, and the like.

[0054] The response data can be acceleration, velocity or displacement data generated by the engineering structure under external load excitation (i.e. system input excitation, such as white noise). The response data can be a time-varying signal whose frequency changes with time. External load excitation can be used to apply external force to the engineering structure to make it vibrate. The response data can be used to describe the vibration form of the structural system under external load excitation, which can be described specifically by a vibration curve graph.

[0055] In order to study the structural damage of an engineering structure or to optimize the structure of an engineering structure, it is necessary to determine the vibration condition of the engineering structure, and then to perform modal identification on the engineering structure. By deploying sensors at various key nodes of the engineering structure, the response data of the engineering structure under environmental excitation can be collected, thereby obtaining the response data of the engineering structure.

[0056] It should be noted that the sensors in the embodiment of the present application can be two or more, forming a synchronous measurement substructure, and each substructure can be deployed in turn at different positions (i.e., measurement points) of the engineering structure to measure the response data at the measurement points of the engineering structure.

[0057] In a possible implementation, the measurement of the engineering structure is divided into multiple substructures according to a preset number of measurement layers, each substructure includes at least two layers of structure, and a substructure has at least one overlapping measurement point with other substructures of the multiple substructures excluding the substructure, which is a common measurement point. The multiple substructures are measured in sequence to obtain at least one response data of each measurement point in the engineering structure.

[0058] As an example, assuming that the engineering structure includes m1, m2, m3, and m4 from high to low, and two sensors are located in two different layers, that is, the two layers are measured simultaneously as a substructure, the substructure division of the engineering structure can be, for example, [m1, m2], [m2, m3], [m3, m4], or [m1, m2], [m1, m3], [m1, m4], or [m1, m2], [m1, m3], [m3, m4], and so on.

[0059] It can be understood that the number of groups of response data obtained at each measurement point is determined based on the substructure assigned to the measurement point. In combination with the above example, the division forms of the substructures [m1, m2], [m2, m3], and [m3, m4] are measured to obtain 1 group of response data of the m1 layer, 2 groups of response data of the m2 layer, 2 groups of response data of the m3 layer, and 1 group of response data of the m4 layer.

[0060] 202. For each set of response data, a nonlinear frequency modulation mode decomposition algorithm is used to perform signal demodulation processing on the response data with the goal of minimizing the objective function to obtain at least first-order target single component data; the objective function is used to evaluate the signal complexity of at least first-order single component data.

[0061] It is understandable that the single component data only includes a signal of one frequency value. For example, the single component data may be a first-order frequency signal or a second-order frequency signal.

[0062] The response data of an engineering structure can be composed of the superposition of several orders of modal vibration shapes, and then the nonlinear frequency modulation modal decomposition algorithm is used to decompose the structural response data into a single component signal corresponding to each order of mode.

[0063] Optionally, a nonlinear chirp mode decomposition (NCMD) algorithm is used to demodulate the response data to achieve decomposition of the response data. Specifically, a wide bandwidth time-varying signal is frequency demodulated, thereby converting the time-varying signal into at least one narrowband signal.

[0064] Specifically, the nonlinear chirp mode (NCM) can be represented by the following function:

[0065] (Formula 1)

[0066] in, represents the nonlinear frequency modulation mode, It can be expressed as instantaneous amplitude, t is the time variable, can be expressed as the instantaneous frequency, It can be expressed as the initial phase.

[0067] The response data can be the superposition of multiple nonlinear frequency modulation modes, which can be specifically expressed by the following formula: (Formula 2)

[0068] in, Indicates the response data. It can be Gaussian white noise, Q represents the total number of modes, and i represents the modal order. Frequency demodulation can be expressed by the following formula: (Formula 3)

[0069] in, as well as can be the demodulated signal, can be expressed as a frequency function of the demodulation operator, Indicates the instantaneous amplitude.

[0070] Among them, the instantaneous amplitude It can be reconstructed by formula 4:

[0071] (Formula 4)

[0072] exist When the demodulated signal and The bandwidth is the smallest, then the signal decomposition problem can be expressed by the following formula:

[0073] (Formula 5)

[0074] (Formula 6)

[0075] Among them, formula 5 represents the demodulated signal and The bandwidth is the smallest. It can be expressed as the L2 norm, express The second-order derivative of express The second-order derivative of , Formula 6 is the constraint condition, and the constraint problem can be solved by the augmented Lagrange multiplier method. The purpose of Formula 5 and Formula 6 is to find a set of demodulated signals and , so that the sum of squares of the L2 norm of its second-order derivative is minimized and the constraint condition of formula 6 is satisfied, thereby realizing the decomposition of the response data and obtaining at least one single component data.

[0076] See also Figure 3a , Figure 3a FIG. 1 shows a schematic diagram of an engineering structure provided by an embodiment of the present invention. Figure 3a As shown, the engineering structure can be a metal frame structure, and the engineering structure can be composed of 5 layers of metal frames, including the first layer, the second layer, the third layer, the fourth layer and the fifth layer. The substructure includes two layers, that is, the acceleration sensor 31 and the acceleration sensor 32 measure two layers at the same time, and the 5 layers are divided into four substructures [first layer, second layer], [second layer, third layer], [third layer, fourth layer], [fourth layer, fifth layer]. The two sensors measure the floors in the substructure in order from low to high in order of the height of the floors, so as to obtain at least one set of response data for each measurement point in the engineering structure. The measurement order is as follows: Figure 3a (a), (b), (c), and (d) shown in the figure.

[0077] When modal identification of an engineering structure is required, environmental excitation, such as white noise, can be input into the engineering structure, and then an acceleration sensor is used to perform measurements in order from low to high to collect response data of the engineering structure under environmental excitation.

[0078] See also Figure 3b , Figure 3b FIG. 1 shows an example schematic diagram of response data in an engineering structure provided by an embodiment of the present application. Figure 3b As shown, curve a can be expressed as the response data of the fifth-layer metal frame in the engineering structure.

[0079] In practical applications, a synchronous substructure measurement group is constructed by two or more sensors. The measurement group is moved step by step, and the structural response data of different layers of metal frames can be collected in sequence by using the measurement groups. For example, two sensors are used to collect the structural response data of the first and second layers of the metal frame, and after the collection is completed, the sensors are moved to the second and third layers of the metal frame, and so on. Sensors can also be arranged in each layer of the metal frame, so that the structural response data of each layer of the metal frame can be collected simultaneously. The specific method can be customized by the user.

[0080] See also Figure 3c , Figure 3c FIG. 4 shows an example schematic diagram of a decomposition response data provided by an embodiment of the present invention. Figure 3cAs shown, curve a can be represented as the response data of the fifth-layer metal frame in the engineering structure, curve b1 can be single component data 1 (i.e., the component is the first order) obtained by decomposing the response data of the fifth-layer metal frame in the engineering structure using NCMD, and curve c1 can be single component data 2 (i.e., the component is the second order) obtained by decomposing the response data of the fifth-layer metal frame in the engineering structure using NCMD, and so on. Curve d1 and curve e1 can be respectively represented as single component data 3 (i.e., the component is the third order) and single component data 4 (i.e., the component is the fourth order) obtained by decomposing the response data of the fifth-layer metal frame in the engineering structure using NCMD.

[0081] Furthermore, Figure 3d An example schematic diagram of a time-frequency curve of decomposed response data provided by an embodiment of the present invention is shown, wherein curve b2 is the time-frequency curve of single component data 1, curve c2 is the time-frequency curve of single component data 2, curve d2 is the time-frequency curve of single component data 3, and curve e2 is the time-frequency curve of single component data 4. Figure 3e An example schematic diagram of a wavelet spectrum of decomposed response data provided by an embodiment of the present invention is shown, wherein curve b3 is the wavelet spectrum of single component data 1, curve c3 is the wavelet spectrum of single component data 2, curve d3 is the wavelet spectrum of single component data 3, and curve e3 is the wavelet spectrum of single component data 4. Figure 3f An example schematic diagram of a spectrum curve of decomposed response data provided by an embodiment of the present invention is shown, wherein curve b4 is the spectrum curve of single component data 1, curve c4 is the spectrum curve of single component data 2, curve d4 is the spectrum curve of single component data 3, and curve e4 is the spectrum curve of single component data 4.

[0082] 203. For each target single component data, obtain spectrum data based on the target single component data.

[0083] Optionally, Fourier transform is performed on the target single-component data to obtain frequency spectrum data of the target single-component signal.

[0084] It can be understood that the amplitude corresponding to each peak in the spectrum data is the amplitude of the vibration mode corresponding to the order of the engineering structure. In the same substructure, there are two or more synchronized sensors. The spectrum amplitudes of the target single component data of each order at different measurement points are obtained through Fourier transform, and the relative ratios of the amplitudes corresponding to the vibration modes of each order at different measurement points in the substructure are obtained. In this way, the relative ratios of the amplitudes of the vibration modes of each order at each measurement point can be obtained.

[0085] 204. For each substructure, the amplitude of each order of vibration mode of each substructure is obtained according to the ratio of the amplitude of the same order spectrum of each measurement point in the spectrum data.

[0086] 205. According to the amplitudes of the vibration modes of each substructure, the vibration mode amplitudes of the common measurement points of each substructure are adjusted to be consistent, so as to obtain the amplitudes of the vibration modes of all the measurement points of the engineering structure.

[0087] Optionally, the response data of the engineering structure is decomposed into multi-order single-component signals, and the frequency corresponding to the spectrum peak of the single-component signal is the natural frequency of the engineering structure.

[0088] According to the amplitudes of each order of vibration mode at all measurement points of the engineering structure, the direction of the vibration mode is determined based on the minimization of the modal confidence criterion.

[0089] 207. According to the direction of vibration mode, the modal vibration mode of the engineering structure is obtained.

[0090] The method provided in the embodiment of the present application takes into account the requirements of the prior art for synchronous acquisition of the response data of the structure. It not only requires the deployment of a large number of sensors at the same time, but also must ensure that these sensors can synchronously measure the response data of the structure, which increases the identification cost; it also increases the requirements for the data acquisition accuracy of the sensor, which causes the problem of excessive difficulty in acquiring the response data. Therefore, at least one set of response data for each measuring point in the engineering structure acquired by the present application may be response data collected at different times, so as to greatly reduce the amount of synchronous deployment of sensors. The requirements for data synchronization are greatly reduced, thereby reducing the difficulty of acquiring response data. Combined with the nonlinear frequency modulation mode decomposition algorithm, the response data is subjected to signal demodulation processing with the goal of minimizing the objective function, so as to decompose the response data into at least first-order target single component data. In this way, by extracting single-component data of the response data and establishing the relationship between the vibration mode amplitude and the target single-component signal spectrum amplitude, the vibration mode amplitude of the common measurement point of each substructure is adjusted to be consistent. Therefore, the acquisition time synchronization of the response data of all measurement points is no longer restricted. It only needs to ensure the synchronization of data acquisition time between different measurement points of the same substructure. Even if two sensors are used to obtain response data of different measurement points of multiple substructures in steps and multiple times, the accuracy of the recognition of the modal vibration mode of the engineering structure can still be guaranteed, so that the obtained modal vibration mode can accurately reflect the health status of the engineering structure.

[0091] In addition, usually, noise will be dispersed into multiple frequency components. In multi-component data, cross-interference will occur between different components, which will reduce the signal-to-noise ratio of the signal and affect the accurate extraction and analysis of the signal. The single-component data in this application only contains modal information of a specific order, so that the impact of noise will be relatively small. In this way, by performing spectral analysis on each target single-component data, the frequency response of each component can be more finely characterized, thereby effectively reducing the errors caused by signal aliasing or interference. This refined analysis method enables the method to more accurately extract the modal vibration shape of the engineering structure during the modal identification process, which not only helps to improve the accuracy of modal vibration shape identification, but also can greatly reduce the amount of calculation and calculation time, and improve the overall analysis efficiency.

[0092] In one embodiment of the present application, in order to improve the accuracy of the nonlinear frequency modulation modal decomposition algorithm for decomposing the response data, the embodiment of the present application also provides a non-parameterized heuristic optimization algorithm to optimize the adjustment parameters used in the nonlinear frequency modulation modal decomposition algorithm. Therefore, the method also includes: determining multiple groups of initial adjustment parameters in the nonlinear frequency modulation modal decomposition algorithm; and optimizing the multiple groups of initial adjustment parameters using the non-parameterized heuristic optimization algorithm to obtain target adjustment parameters; the target adjustment parameters are used to obtain at least first-order target single component data.

[0093] Optionally, when the current iteration number is 1, multiple groups of initial adjustment parameters are randomly generated.

[0094] Optionally, in the nonlinear frequency modulation mode decomposition algorithm, the adjustment parameters involved include two types: mode decomposition quantity and penalty parameter.

[0095] In some embodiments, the adjustment parameters may include the modal decomposition number k in the nonlinear frequency modulation modal decomposition algorithm, the penalty coefficient α for controlling the filter bandwidth of the nonlinear frequency modulation modal decomposition algorithm, and the penalty coefficient β for controlling the incremental smoothness of the instantaneous frequency in the iterative process. The smaller the penalty coefficient α, the narrower the bandwidth of the demodulated signal, and the smaller the penalty coefficient β, the smoother the incremental instantaneous frequency.

[0096] As an example, the target adjustment parameter may be an adjustment parameter when the accuracy of decomposing the response data into at least one single component data based on the nonlinear frequency modulation modal decomposition algorithm is optimal, and the initial adjustment parameter may be an adjustment parameter when the nonlinear frequency modulation modal decomposition algorithm is capable of decomposing the structural response data into at least one single component data.

[0097] When modal identification of an engineering structure is required, a nonlinear frequency modulation modal decomposition algorithm for decomposing the response data is determined, and initial adjustment parameters related to the demodulation processing in the nonlinear frequency modulation modal decomposition algorithm are determined. Then, a parameter-free heuristic optimization algorithm, namely the Dynamic Swarm Hunting Optimizer (DSHO), can be used to optimize the initial adjustment parameters, and sample entropy is used as the objective function of the initial adjustment parameters to obtain the target adjustment parameters.

[0098] In one embodiment of the present application, referring to Figure 4 , using a non-parameterized heuristic optimization algorithm to optimize multiple sets of initial adjustment parameters to obtain target adjustment parameters, including:

[0099] 401. In the current iteration number, determine the fitness corresponding to each set of initial adjustment parameters based on the objective function.

[0100] The objective function may be sample entropy (Sample Entropy, SampEn).

[0101] In practical applications, fitness can be used to characterize the similarity or matching degree between the reference signal or the ideal signal of the initial single component data obtained after the nonlinear frequency modulation mode decomposition algorithm decomposes the response data using the initial adjustment parameters. Taking sample entropy as the objective function, fitness is specifically used to characterize the signal complexity of the initial single component data.

[0102] It can be understood that the smaller the fitness is, the smaller the signal complexity of the single component data is, which further indicates that the decomposition effect of the nonlinear frequency modulation mode decomposition algorithm on the response data is better.

[0103] The current iteration number refers to the number of current iterations in the iteration process.

[0104] Optional, see Figure 5 , based on the objective function, determining the fitness corresponding to each set of initial adjustment parameters includes the following steps:

[0105] 4011. For each group of initial adjustment parameters, perform signal demodulation processing on the response data according to a nonlinear frequency modulation mode decomposition algorithm using the initial adjustment parameters to obtain at least one initial single component data.

[0106] 4012. Determine the signal complexity of each initial single component data using the objective function.

[0107] In one possible implementation, the sample entropy is used as the objective function, and the signal complexity of each initial single component data can be determined by using the objective function in the following manner:

[0108] Represent the single-component data as: , where N represents the length of the single-component data. After that, divide each single-component data sequence into different subsequences, each subsequence having a length of m, then the subsequence can be understood as in starting from a sequence of m consecutive values, that is

[0109] For a given similarity tolerance r, define two subsequences and (where I ≠ J) to be similar if the maximum absolute difference between their corresponding elements does not exceed r, that is holds for all 0 ≤ k < m.

[0110] For each I (1 ≤ I ≤ N - m), calculate how many J (1 ≤ J ≤ N - m, J ≠ I) such that and are similar, and denote this quantity as .

[0111] Calculate the average value of all , , denoted as:

[0112] (Equation Seven)

[0113] For m + 1, repeat the above steps to calculate and .

[0114] denotes how many J (1 ≤ J ≤ N - m - 1, J ≠ I) such that and are similar. denotes the average value of.

[0115] After that, calculate the objective function SampEn(m, r, N) according to the following formula:

[0116] (Equation Eight)

[0117] It can be seen from this that the value of the objective function calculated based on the above initial single-component data is the fitness of the initial single-component data.

[0118] 4013. Determine the minimum signal complexity among at least one initial single-component data as the fitness corresponding to the initial adjustment parameter.

[0119] 402. Construct an initial position vector for each set of initial adjustment parameters, and determine a core position vector with the minimum fitness among multiple initial position vectors.

[0120] In some embodiments, coordinates for describing the adjustment parameters may be defined based on the dimensions of the adjustment parameters, and a position vector may be randomly generated within a preset search domain.

[0121] For example, the adjustment parameters may include the number of modal decompositions Q, the penalty coefficient α for controlling the filter bandwidth of the nonlinear frequency modulation modal decomposition algorithm, and the penalty coefficient β for controlling the incremental smoothness of the instantaneous frequency of the iterative process. Based on the dimension of the adjustment parameters, the position vector used to describe the adjustment parameters can be defined as (Q, α, β), and then the value of each adjustment parameter can be randomly selected within the preset search domain, that is, the value of each adjustment parameter can be randomly determined to generate multiple initial position vectors.

[0122] Specifically, the position vector of the adjustment parameter can be determined by the following formula:

[0123] (Formula 9)

[0124] in, You can preset the upper limit of the search domain. You can preset the lower limit of the search domain. Can be a random number between (0,1), It can be expressed as the position vector of the wth adjustment parameter among all the parameters to be optimized, and w can be expressed as any one of the position vectors of all the adjustment parameters.

[0125] 403. Update each initial position vector based on the core position vector to obtain a candidate position vector corresponding to each initial position vector; if the current iteration number does not meet the iteration number condition, use multiple candidate position vectors to update multiple initial position vectors until the current iteration number meets the iteration number condition, and then determine the parameter in the position vector with the smallest fitness among the multiple candidate position vectors as the target adjustment parameter.

[0126] It can be understood that after the multiple initial position vectors are updated using the multiple candidate position vectors, steps 402 and 403 are re-executed until the current iteration number meets the iteration number condition, and then the iteration can be stopped.

[0127] Among them, each initial position vector is updated based on the core position vector to obtain a candidate position vector corresponding to each initial position vector, including: for each initial position vector, with the actual position difference being less than the preset position difference as the goal, the initial position vector is updated according to the core position vector, the weight value and the balance index to obtain the candidate position vector; wherein the position difference is used to characterize the position deviation rate of the candidate position vector compared to the core position vector, and the balance index is used to characterize the movable range of the initial position vector.

[0128] In one embodiment of the present application, referring to Figure 6 , with the goal of making the actual position difference smaller than the preset position difference, the initial position vector is updated according to the core position vector, weight value and balance index to obtain the candidate position vector, including:

[0129] 601. Determine an initial value of a balance index within a first index interval and a weight value of each initial position vector among a plurality of initial position vectors.

[0130] Optionally, the balance index is related to the current iteration number, and the balance index is used to determine the direction of the iterative update of the position vector. The direction of the iterative update of the position vector may include determining the update direction of the target adjustment parameter within all preset search domains, and determining the update direction of the target adjustment parameter based on the area corresponding to all position vectors in the current iteration. The area corresponding to all position vectors in the current iteration may be an area determined based on the global coordinates of each position vector in the current iteration within the preset search domain.

[0131] In some embodiments, the weight value of the initial position vector among the multiple initial position vectors is determined based on the fitness of each initial position vector and the total fitness value of the multiple initial position vectors.

[0132] Specifically, the weight value may represent the positional relationship between each position vector in the preset search domain. The larger the weight value is, the farther the relative position between the position vector and other position vectors in the preset search domain is.

[0133] During each iteration, the current iteration number can be determined, and then the balance indicator can be determined according to the current iteration number.

[0134] Specifically, the balance index is determined using the following formula:

[0135] (Formula 10)

[0136] in, can be expressed as the current iteration number, It can be expressed as the maximum number of iterations, Can be a random number between (0,1), It can be the balance indicator corresponding to the current iteration number.

[0137] Reference Figure 7a and Figure 7b , Figure 7a FIG. 1 is a schematic diagram showing an example of a balance indicator in an iterative process provided by an embodiment of the present invention. Figure 7b FIG. 2 is a schematic diagram showing an example of a fitness convergence curve in an iterative process provided by an embodiment of the present invention. Figure 7a As shown, the horizontal axis is the number of iterations, and the vertical axis is the value of the balance index, such as Figure 7b As shown in , the horizontal axis is the number of iterations, and the vertical axis is the value of fitness. Figure 7a It can be seen that the balance index is a changing value in the interval (0,2), and as time goes by, that is, the balance index shows a decreasing trend with the increase of the number of iterations, so the balance index can be used It can ensure that as the iteration proceeds, the optimization process can converge, thereby obtaining the minimum value of the objective function (i.e. fitness). Balanced indicator The size of determines whether the direction of later iterations is to explore more extensively within the preset search domain or to conduct in-depth development in the current field. In the early iteration of the optimization process, the algorithm has a higher level of exploration ability, which can escape the local optimal solution and find the global optimal solution. When the optimization result is close to the optimal solution (that is, the position vector corresponding to the target adjustment parameter) in the later stage of the optimization process, the algorithm enters the key development and focuses on searching for the optimal solution that may appear nearby. Therefore, the balance index It is an important indicator to weigh the extensive exploration and focused development of the algorithm, which is only related to the current number of iterations.

[0138] In each iteration, the relative position between each position vector and other position vectors can be determined based on the fitness of all position vectors, that is, the weight value of the position vector in all position vectors can be determined. Specifically, the weight value can be determined by the following formula:

[0139] (Formula 11)

[0140] in, represents the weight value of the w-th position vector, represents the fitness of the w-th position vector, and Num represents the maximum population size for iterative updates, that is, the number of position vectors.

[0141] 602. Using the core position vector, the initial value and the weight value, the initial position vector is updated in a preset search domain to obtain a first position vector corresponding to each initial position vector.

[0142] Optionally, a position difference vector between each initial position vector and the core position vector is determined, and the initial position vector is updated in a preset search domain according to the balance index, the position weight value and the position difference vector to obtain a first position vector.

[0143] The position difference vector is used to represent the relative position between each initial position vector and the core position vector in the preset search domain.

[0144] Specifically, the position difference vector can be determined using the following formula:

[0145] (Formula 12)

[0146] in, can be expressed as the w-th position difference vector, It can be expressed as the coordinates of the core position vector in the preset search domain. It can be the coordinate of the w-th initial position vector in the search domain.

[0147] After the position difference vector is obtained, when the balance index is within the first index interval, for each initial position vector, the initial position vector can be updated within the preset search domain according to the balance index, the weight value and the position difference vector to obtain the first position vector. Specifically, the initial position vector can be updated by the following formula:

[0148] (Formula 13)

[0149] in, It can be expressed as the weight value of the w-th initial position vector, It can be expressed as the balance index corresponding to the current iteration number, can be the position difference vector of the w-th initial position vector, It can be expressed as the first position vector obtained after updating the w-th initial position vector.

[0150] It should be noted that updating the initial position vector in the preset search domain is to update all the initial position vectors including the core position vector. It is understandable that the core position vector also corresponds to the first position vector. That is to say, when the initial position vector is updated using the first position vector, the core position vector will be re-determined in the updated initial position vector.

[0151] 603. For each initial position vector, determine a candidate value of the balance index, and use the candidate value to update the initial value and use the first position vector to update the initial position vector until the candidate value is within the second index interval, and determine the first position vector as the second position vector; wherein the candidate value is smaller than the initial value, and the first index interval is adjacent to and larger than the second index interval.

[0152] Exemplarily, the first index interval is [1, 2], and the second index interval is [0, 1).

[0153] It can be understood that using candidate values ​​to update the initial value and using the first position vector to update the initial position vector means that the entire large iterative process of the algorithm also includes a sub-iterative process, and the sub-iterative process includes a first sub-iterative process and a second sub-iterative process. The first sub-iterative process corresponds to an iterative process in which the balance index is within the first index interval, and the second sub-iterative process corresponds to an iterative process in which the balance index is within the second index interval.

[0154] Optionally, the candidate value being smaller than the initial value indicates that the balance indicator is assigned in a gradually decreasing manner.

[0155] 604. If the position deviation rate of each second position vector compared to the core position vector is within the preset deviation range, it is determined that the actual position difference is less than the preset position difference, and the core position vector is updated to a position vector corresponding to the core position vector in the plurality of second position vectors. For each second position vector, the second position vector is updated using the core position vector and the weight value to obtain a candidate position vector.

[0156] Optionally, if the position deviation rate of the second position vector compared to the core position vector is within a preset deviation range, the parameter average of the second position vector is determined, and the second position vector is updated according to the parameter average, balance index, weight value and core position vector to obtain a candidate position vector.

[0157] The parameter average is the coordinate average of the second position vector.

[0158] Specifically, the parameter average value may be the coordinate average value of the 1st second position vector to the ith second position vector.

[0159] Optionally, the position deviation rate may represent the ratio of the relative position between each second position vector and the core position vector in the preset search domain to the second position vector. If the position deviation rate is less than the preset deviation range (such as satisfying the position deviation rate ), it can be explained that in the current iteration number, the relative position between the second position vector and the core position vector is relatively close, that is, the area where the global optimal solution is located is narrow, and the area where the global optimal solution is located can be further updated.

[0160] As an example, parameter sharing information corresponding to the second position vector is determined based on the parameter average value and the position difference vector corresponding to the second position vector, and each second position vector is updated based on the parameter sharing information and the core position vector to obtain a candidate position vector corresponding to each second position vector (in short, the initial adjustment parameters are updated according to the parameter sharing information).

[0161] Specifically, parameter sharing information It can be determined by the following formula:

[0162] (Formula 14)

[0163] in, It can be the parameter average value of the 1st second position vector to the yth second position vector, It can be the position difference vector between the y-th second position vector and the core position vector, It can be the weight of the average parameter value of the first second position vector to the yth second position vector in the parameter sharing information, and the value is any random number in (0,1). It should be noted that the core position vector here is the core position vector among multiple second position vectors, that is, the position vector corresponding to the original core position vector among multiple second position vectors. y corresponds to w one by one.

[0164] After the parameter sharing information of each second position vector is obtained, the second position vector may be updated according to the parameter sharing information of each second position vector and the core position vector to obtain a candidate position vector.

[0165] Specifically, the second position vector can be updated using the following formula: :

[0166] (Formula 15)

[0167] 605. If there are one or more second position vectors whose position deviation rate compared to the core position vector is outside the preset deviation range, then for each second position vector, the second position vector is iteratively updated using the core position vector, the weight value and the candidate value until the position deviation rate of each second position vector compared to the core position vector is within the preset deviation range, the core position vector is updated to a position vector corresponding to the core position vector among the multiple second position vectors, and for each second position vector, the second position vector is updated using the core position vector and the weight value to obtain a candidate position vector.

[0168] It can be understood that the core position vector in the iterative update of the second position vector using the core position vector, the weight value and the candidate value in step 605 refers to the position vector corresponding to the core position vector in the plurality of second position vectors.

[0169] Optionally, after obtaining the parameter average value, for each second position vector, the second position vector may be updated according to the parameter average value, a balance index (where the value of the balance index is a candidate value), a weight value, and a core position vector.

[0170] Specifically, the second position vector may be updated using the following formula:

[0171] (Formula 16)

[0172] in, It can be expressed as the parameter average value from the 1st second position vector to the yth second position vector, It can be expressed as the position vector obtained after updating the y-th second position vector (that is, the updated y-th second position vector).

[0173] It can be understood that the second position vector is iteratively updated using Formula 16 until the position deviation rate of each second position vector compared to the core position vector is within the preset deviation range, the iteration stops, and the second position vector obtained by the last iteration is updated using Formulas 14 and 15 to obtain the candidate position vector corresponding to the second position vector.

[0174] In one embodiment of the present application, updating multiple initial position vectors using multiple candidate position vectors includes: determining the fitness of each candidate position vector based on an objective function; if there is a candidate position vector with a fitness less than the fitness of the candidate position vector corresponding to the core position vector among other candidate position vectors, updating the multiple candidate position vectors to multiple initial position vectors; the other candidate position vectors include position vectors among the multiple candidate position vectors except the candidate position vector corresponding to the core position vector.

[0175] If the fitness of the candidate position vectors corresponding to the core position vector is greater than or equal to the fitness of other candidate position vectors, each candidate position vector is updated based on the Levy flight algorithm until there is a candidate position vector whose fitness is less than the fitness of the candidate position vector corresponding to the core position vector, and the multiple candidate position vectors are updated to multiple initial position vectors.

[0176] Optionally, for each candidate position vector, a candidate adjustment parameter corresponding to the candidate position vector is determined, and the response data is subjected to signal demodulation processing according to a nonlinear frequency modulation modal decomposition algorithm that adopts the candidate adjustment parameter to obtain at least one candidate single component data; the signal complexity of each candidate single component data is determined using an objective function; and the minimum signal complexity among at least one candidate single component data is determined as the fitness of the candidate position vector.

[0177] In some embodiments, each candidate position vector is updated based on the following formula 17 and formula 18 until there is a candidate position vector whose fitness is less than the fitness of the candidate position vector corresponding to the core position vector among other candidate position vectors, and the candidate position vector is updated to the initial position vector.

[0178] (Formula 17)

[0179] in, can be expressed as the zth candidate position vector, It can be expressed as the yth updated candidate position vector, where z corresponds to y one by one. Specifically, the Levy flight algorithm can be expressed by the following formula:

[0180] (Formula 18)

[0181] in, and can be a random number between (0,1). It can be expressed as the gamma function, where The value can be 1.5.

[0182] For example, in order to better understand the optimization steps of optimizing multiple groups of initial adjustment parameters to obtain target adjustment parameters using the non-parameterized heuristic optimization algorithm, that is, the dynamic swarm hunting optimization algorithm (DSHO) in the embodiment of the present application, refer to Figure 8 ,include:

[0183] S81, randomly generating multiple groups of initial adjustment parameters in the nonlinear frequency modulation mode decomposition algorithm, and determining an initial position vector corresponding to each group of initial adjustment parameters.

[0184] S82, determining the fitness of each group of initial position vectors, and determining a core position vector from all the initial position vectors according to the fitness.

[0185] S83, determine the balance index corresponding to the current iteration number .

[0186] S84, judging balance index Is it greater than 1? If so, execute step S85; if not, execute step S86.

[0187] S85, using the first position vector to update the candidate position vector, and executing step S810.

[0188] S86, determine the position deviation rate E, and judge whether the position deviation rate E is greater than 0.5, if so, execute step S87, if not, execute step S88.

[0189] S87, using the second position vector to update the candidate position vector, and executing step S810.

[0190] S88, determining parameter sharing information , execute step S89.

[0191] S89, update the second position vector, and use the updated second position vector to update the candidate position vector, and execute step S810.

[0192] S810, determining whether there is fitness in other candidate position vectors Is it smaller than the fitness of the candidate position vector corresponding to the core position vector? If so, execute step S811, if not, execute step S812.

[0193] S811 , updating the core position vector based on the position vectors whose fitness is smaller than the core position vector among all the candidate position vectors.

[0194] S812, update each candidate position vector based on the Levy flight algorithm, and execute step S813.

[0195] S813, determine whether there is fitness in other candidate position vectors Is it smaller than the fitness of the candidate position vector corresponding to the core position vector? , execute step S811, if not, execute step S814.

[0196] S814, determine whether the current iteration number meets the iteration number condition, if so, execute step S815, if not, execute step S83.

[0197] S815, determining the parameter in the position vector with the smallest fitness among the multiple candidate position vectors of the current iteration number as the target adjustment parameter.

[0198] It should be noted that Figure 8 Only a brief description is given of the above steps S81 - S815. Please refer to the contents of the above steps S81 - S815 for details.

[0199] In practical applications, the optimal solution corresponding to the current iteration number may be the candidate position vector in the current iteration number.

[0200] In a specific implementation, the above method can be used to determine the global optimal solution from the search domain of the adjustment parameters, and the balance index can gradually converge to the area where the global optimal solution is located. As the iteration continues, the balance index shows a trend of gradually shrinking, thereby determining the target adjustment parameters for optimizing the initial modal decomposition algorithm, thereby improving the accuracy of the target adjustment parameters. In addition, since the parameter-free heuristic optimization algorithm does not involve other adjustment parameters, when the parameter-free heuristic optimization algorithm is used to optimize the adjustment parameters in the nonlinear frequency modulation modal decomposition algorithm, more adjustment parameters will not be introduced, thereby reducing the influence of too many hyperparameters on the adjustment parameters in the optimization of the nonlinear frequency modulation modal decomposition algorithm.

[0201] In one embodiment of the present application, with the goal of adjusting the amplitudes of each order vibration mode at the common measurement points of each substructure to be consistent, the amplitudes of each order vibration mode in each substructure are adjusted to obtain the amplitudes of each order vibration mode at all measurement points; based on the amplitudes of each order vibration mode at all measurement points, the direction of the vibration mode is determined based on minimization of the modal confidence criterion; based on the direction of the vibration mode, the modal vibration mode of the engineering structure is obtained.

[0202] Optionally, spectrum data is obtained by performing Fourier transform on each order single component signal, and spectrum data of single component signals in each substructure is calculated by performing Fourier transform on each order single component signal, thereby obtaining the frequency and amplitude corresponding to the spectrum peak in the spectrum data, and further, according to the ratio of the amplitude of each order vibration mode of the engineering structure to the amplitude of each order spectrum, by adjusting the vibration mode amplitude of the common measurement point of each substructure to be consistent, the amplitude of each order vibration mode at each measurement point and the frequency corresponding to each order single component signal are obtained in turn, thereby obtaining the relative ratio of the amplitude of each order vibration mode of the engineering structure at each measurement point; and then, according to the relative ratio of each order vibration mode at each measurement point, the direction of each order vibration mode of the engineering structure is obtained using the modal confidence criterion, that is, first determine that the direction of the first-order vibration mode is all positive, traverse the positive and negative directions of each measurement point of the second-order vibration mode in turn, calculate the orthogonality of the first-order vibration mode to all possible second-order vibration mode directions, and determine the direction of the second-order vibration mode based on the principle of minimum modal confidence criterion. Similarly, the vibration modes of other orders can be obtained, thereby obtaining the modal parameters of the overall engineering structure, such as frequency and vibration mode. It can be understood that the frequency corresponding to each peak in the frequency data (i.e., Fourier spectrum) is the natural frequency of the structure (the spectrum peak ratio of each layer in each order is the vibration amplitude).

[0203] Optionally, use the MAC formula to calculate the MAC value between two modal vibration vectors. The calculation formula for the MAC value is:

[0204] (Formula 19)

[0205] in, and represent the i-th and j-th vibration mode vectors respectively.

[0206] It should be noted that the orthogonality between two modes is evaluated based on the size of the MAC value. The closer the MAC value is to 1, the more correlated the two modes are; the smaller the MAC value (closer to 0), the less correlated the two modes are.

[0207] The method provided in the present application is free from noise interference and does not need to eliminate false modes, thereby obtaining a more accurate modal vibration shape of the engineering structure.

[0208] As an example, the substructure modal identification method that can be measured in multiple steps provided in the embodiment of the present application is used to identify Figure 3a The modal identification of the response data of the five-layer metal frame in the final frequency results is carried out, and the stochastic subspace identification (SSI) algorithm is used to identify the Figure 3a The modal identification of the response data of the five-layer metal frame in the figure finally yields the frequency results of each order, see Table 1.

[0209]

[0210] Based on Table 1, it can be seen that the substructure modal identification method that can be measured in steps and multiple times in the embodiment of the present application is not much different from the random subspace algorithm in terms of identification accuracy, and can meet the accuracy requirements of engineering structures. However, this method does not require all measurement points to be time synchronized in the collection of response data, and does not require the elimination of false modes and the determination of system orders in the modal identification process, so that the modal identification process reduces human intervention, avoids the subjectivity of the identification structure, and improves the accuracy of modal vibration type identification.

[0211] Further, see Fig. 9 , Fig. 9 In order to utilize the substructure modal identification method capable of multiple step-by-step measurements provided by the embodiment of the present application Figure 3a Schematic diagram of an example of the 4th-order vibration mode finally obtained by modal identification of the response data of the five-layer metal frame in FIG. 1 , wherein curve b5 is the first-order vibration mode curve, curve c5 is the second-order vibration mode curve, curve d5 is the third-order vibration mode curve, and curve e5 is the fourth-order vibration mode curve. Fig.10 Schematic diagram of an example of 4th-order orthogonality verification. Fig.10 The horizontal axis in represents the modal order.

[0212] It should be noted that although the operations of the method of the present application are described in a particular order in the drawings, this does not require or imply that these operations must be performed in this particular order or that all the operations shown must be performed to achieve the desired results.

[0213] Fig.11 A block diagram of a substructure modal identification device capable of performing multiple measurements in steps according to an embodiment of the present application.

[0214] like Fig.11 As shown, the substructure modal identification device capable of multiple step-by-step measurements includes: an acquisition module 1101, a decomposition module 1102, a transformation module 1103, an identification module 1104, and a parameter optimization module 1105. Among them,

[0215] The acquisition module 1101 is used to acquire the response data of each measurement point in each substructure of the engineering structure; the response data of all measurement points of the same substructure are response data collected synchronously at the same time; the response data of measurement points of different substructures are response data collected multiple times at different times in steps; each substructure has at least one common measurement point with other substructures.

[0216] The decomposition module 1102 is used to perform signal demodulation processing on each response data using a nonlinear frequency modulation mode decomposition algorithm with the objective of minimizing the objective function to obtain at least first-order target single component data; the objective function is used to evaluate the signal complexity of at least one single component data.

[0217] The transformation module 1103 is used to obtain spectrum data based on the target single component data for each target single component data.

[0218] The identification module 1104 is used to obtain the amplitude of each order of vibration mode of each substructure according to the ratio of the amplitude of the same order spectrum of each measurement point in the spectrum data for each substructure.

[0219] The identification module 1104 is also used to adjust the vibration mode amplitudes of the common measurement points of each substructure to be consistent according to the vibration mode amplitudes of each substructure, so as to obtain the vibration mode amplitudes of each order of all measurement points of the engineering structure; determine the direction of the vibration mode based on the minimization of the modal confidence criterion according to the vibration mode amplitudes of each order of all measurement points of the engineering structure; and obtain the modal vibration mode of the engineering structure according to the direction of the vibration mode.

[0220] In some embodiments, the parameter optimization module 1105 is used to determine multiple groups of initial adjustment parameters in the nonlinear frequency modulation mode decomposition algorithm; optimize the multiple groups of initial adjustment parameters using a non-parameterized heuristic optimization algorithm to obtain target adjustment parameters; the target adjustment parameters are used to obtain at least first-order target single component data.

[0221] In some embodiments, the parameter optimization module 1105 is specifically used to:

[0222] In the current iteration number, the fitness corresponding to each set of initial adjustment parameters is determined based on the objective function.

[0223] An initial position vector for each set of initial adjustment parameters is constructed, and a core position vector with the minimum fitness is determined among multiple initial position vectors.

[0224] Based on the core position vector, each initial position vector is updated to obtain a candidate position vector corresponding to each initial position vector; if the current iteration number does not meet the iteration number condition, multiple initial position vectors are updated using multiple candidate position vectors until the current iteration number meets the iteration number condition, and the parameter in the position vector with the smallest fitness among the multiple candidate position vectors is determined as the target adjustment parameter.

[0225] In some embodiments, the parameter optimization module 1105 is specifically used to update each initial position vector according to the core position vector, the weight value and the balance index with the goal of making the actual position difference smaller than the preset position difference, so as to obtain a candidate position vector corresponding to each initial position vector; wherein the position difference is used to characterize the position deviation rate of the candidate position vector compared to the core position vector, and the balance index is used to characterize the movable range of the initial position vector.

[0226] In some embodiments, the parameter optimization module 1105 is specifically used to:

[0227] An initial value of the balance index within the first index interval and a weight value of each initial position vector among the multiple initial position vectors are determined.

[0228] Each initial position vector is updated in a preset search domain using the core position vector, the initial value and the weight value to obtain a first position vector corresponding to each initial position vector.

[0229] For each initial position vector, a candidate value of the balance index is determined, and the initial value is updated using the candidate value and the initial position vector is updated using the first position vector until the candidate value is within the second index interval, and the first position vector is determined as the second position vector; wherein the candidate value is smaller than the initial value, and the first index interval is adjacent to and larger than the second index interval.

[0230] If the position deviation rate of each second position vector compared to the core position vector is within the preset deviation range, it is determined that the actual position difference is less than the preset position difference, and the position vector corresponding to the core position vector in the multiple second position vectors is updated to the core position vector, and for each second position vector, the second position vector is updated using the core position vector and the weight value to obtain a candidate position vector.

[0231] If there are one or more second position vectors whose position deviation rate compared to the core position vector is outside the preset deviation range, then for each second position vector, the second position vector is iteratively updated using the core position vector, weight value and candidate value until the position deviation rate of each second position vector compared to the core position vector is within the preset deviation range, then the position vector corresponding to the core position vector in the multiple second position vectors is updated to the core position vector, and for each second position vector, the second position vector is updated using the core position vector and weight value to obtain the candidate position vector.

[0232] In some embodiments, the parameter optimization module 1105 is specifically used to:

[0233] The fitness of each candidate position vector is determined based on the objective function.

[0234] If there is a candidate position vector with a fitness smaller than that of the candidate position vector corresponding to the core position vector among other candidate position vectors, the multiple candidate position vectors are updated to multiple initial position vectors; the other candidate position vectors include position vectors among the multiple candidate position vectors except the candidate position vector corresponding to the core position vector.

[0235] If the fitness of the candidate position vectors corresponding to the core position vector is greater than or equal to the fitness of other candidate position vectors, each candidate position vector is updated based on the Levy flight algorithm until there is a candidate position vector whose fitness is less than the fitness of the candidate position vector corresponding to the core position vector, and the multiple candidate position vectors are updated to multiple initial position vectors.

[0236] In some embodiments, the parameter optimization module 1105 is specifically used to:

[0237] For each group of initial adjustment parameters, a signal demodulation process is performed on the response data according to a nonlinear frequency modulation mode decomposition algorithm using the initial adjustment parameters to obtain at least first-order initial single-component data.

[0238] The signal complexity of each order of initial single component data is determined using the objective function.

[0239] The minimum signal complexity in at least one order of initial single component data is determined as the fitness corresponding to the initial adjustment parameter.

[0240] It should be noted that the information interaction, execution process and other contents between the above-mentioned devices are based on the same concept as the method embodiment of the present application. Their specific functions and technical effects can be found in the method embodiment part and will not be repeated here.

[0241] The technicians in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

[0242] Reference below Fig.12 , Fig.12 A schematic diagram of the structure of a terminal device suitable for implementing an embodiment of the present application is shown. Fig.12 As shown, the computer system 1200 includes a central processing unit (CPU) 1201, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 1202 or the program loaded from the storage part 1208 to the random access memory (RAM) 1203. In the RAM 1203, various programs and data required for the operation instructions of the system are also stored. The CPU 1201, the ROM 1202, and the RAM 1203 are connected to each other through a bus 1204. An input / output (I / O) interface 1205 is also connected to the bus 1204.

[0243] The following components are connected to the I / O interface 1205: an input section 1206 including a keyboard, a mouse, etc.; an output section 1207 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 1208 including a hard disk, etc.; and a communication section 1209 including a network interface card such as a LAN card, a modem, etc. The communication section 1209 performs communication processing via a network such as the Internet. A drive 1210 is also connected to the I / O interface 1205 as needed. A removable medium 1211, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 1210 as needed, so that a computer program read therefrom is installed into the storage section 1208 as needed.

[0244] In particular, according to an embodiment of the present application, the above reference flow chart Figure 2The described process can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through the communication part 1209, and / or installed from a removable medium 1211. When the computer program is executed by the central processing unit (CPU) 1201, the above-mentioned functions defined in the system of the present application are executed.

[0245] It should be noted that the computer-readable medium shown in the present application may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present application, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium such as a computer-readable storage medium that can send, propagate or transmit a program for use by or in conjunction with an instruction execution system, device or device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0246] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, functions and operating instructions of the system, method and computer program product according to various embodiments of the present application. In this regard, each box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the aforementioned module, program segment or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some implementations as replacements, the functions marked in the box can also occur in a sequence different from that marked in the accompanying drawings. For example, the boxes represented by two connections can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operating instruction, or can be implemented with a combination of dedicated hardware and computer instructions.

[0247] The units or modules involved in the embodiments described in the present application may be implemented by software or hardware. The units or modules described may also be provided in a processor, for example, may be described as: a processor including an illegal person detection unit, a multimodal detection unit, and a recognition unit. The names of these units or modules do not, in some cases, constitute limitations on the units or modules themselves.

[0248] As another aspect, the present application further provides a computer-readable storage medium, which may be included in the terminal device described in the above embodiment, or may exist independently without being assembled into the terminal device. The above computer-readable storage medium stores one or more programs, and when the above programs are used by one or more processors to execute the substructure modal identification method capable of multiple step-by-step measurements described in the present application. For example, it may be executed Figure 2 The various steps of the substructure modal identification method that can be measured step by step multiple times are shown.

[0249] The present application embodiment provides a computer program product, which includes instructions. When the instructions are executed, the method described in the present application embodiment is executed. For example, Figure 2 The various steps of the substructure modal identification method that can be measured step by step multiple times are shown.

[0250] The above description is only a preferred embodiment of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in the present application is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the aforementioned disclosed concept. For example, the above features are replaced with the technical features with similar functions disclosed in the present application (but not limited to) by each other to form a technical solution.

[0251] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A substructure modal identification method capable of multiple step-by-step measurements, characterized in that: The method comprises: Obtain the response data of each measurement point in each substructure of the engineering structure; the response data of all measurement points of the same substructure are response data collected synchronously at the same time; the response data of measurement points of different substructures are response data collected multiple times at different times; each substructure has at least one common measurement point with other substructures; For each response data, a nonlinear frequency modulation mode decomposition algorithm is used to perform signal demodulation processing on the response data with the objective of minimizing the objective function, so as to obtain at least first-order target single component data; the objective function is used to evaluate the signal complexity of the at least one single component data; For each of the target single component data, obtaining spectrum data based on the target single component data; For each substructure, the amplitude of each order of vibration mode of each substructure is obtained according to the ratio of the amplitude of the same order spectrum of each measurement point in the spectrum data; According to the vibration mode amplitudes of each substructure, the vibration mode amplitudes of the common measurement points of each substructure are adjusted to be consistent, so as to obtain the vibration mode amplitudes of all measurement points of the engineering structure; According to the amplitudes of each order vibration mode at all measurement points of the engineering structure, the direction of the vibration mode is determined based on the minimization of the modal confidence criterion; According to the direction of the vibration mode, the modal vibration mode of the engineering structure is obtained.

2. The substructure modal identification method capable of multiple step-by-step measurements according to claim 1 is characterized in that: The method further comprises: Determining multiple groups of initial adjustment parameters in the nonlinear frequency modulation mode decomposition algorithm; The multiple groups of initial adjustment parameters are optimized using a parameter-free heuristic optimization algorithm to obtain target adjustment parameters; the target adjustment parameters are used to obtain the at least first-order target single component data.

3. The substructure modal identification method capable of multiple step-by-step measurements according to claim 2, characterized in that: The method of optimizing the multiple groups of initial adjustment parameters using a non-parameterized heuristic optimization algorithm to obtain target adjustment parameters includes: In the current iteration number, determining the fitness corresponding to each set of initial adjustment parameters based on the objective function; Constructing an initial position vector for each set of initial adjustment parameters, and determining a core position vector with the minimum fitness among multiple initial position vectors; Based on the core position vector, each initial position vector is updated to obtain a candidate position vector corresponding to each initial position vector; if the current number of iterations does not meet the iteration number condition, the multiple initial position vectors are updated using multiple candidate position vectors until the current number of iterations meets the iteration number condition, and the parameters in the position vector with the smallest fitness among the multiple candidate position vectors are determined as the target adjustment parameters.

4. The substructure modal identification method capable of multiple step-by-step measurements according to claim 3 is characterized in that: The updating each initial position vector based on the core position vector to obtain a candidate position vector corresponding to each initial position vector includes: With the goal of making the actual position difference smaller than the preset position difference, each initial position vector is updated according to the core position vector, the weight value and the balance index to obtain a candidate position vector corresponding to each initial position vector; wherein the position difference is used to characterize the position deviation rate of the candidate position vector compared to the core position vector, and the balance index is used to characterize the movable range of the initial position vector.

5. The substructure modal identification method capable of multiple step-by-step measurements according to claim 4, characterized in that: The step of taking the actual position difference less than the preset position difference as a goal, updating each of the initial position vectors according to the core position vector, the weight value, and the balance index, and obtaining a candidate position vector corresponding to each of the initial position vectors includes: Determining an initial value of the balance index within a first index interval and a weight value of each initial position vector among the multiple initial position vectors; Using the core position vector, the initial value and the weight value, each of the initial position vectors is updated in a preset search domain to obtain a first position vector corresponding to each of the initial position vectors; For each initial position vector, determine a candidate value of the balance index, and use the candidate value to update the initial value and use the first position vector to update the initial position vector until the candidate value is within a second index interval, and determine the first position vector as the second position vector; wherein the candidate value is smaller than the initial value, and the first index interval is adjacent to and larger than the second index interval; If the position deviation rate of each of the second position vectors compared to the core position vector is within a preset deviation range, it is determined that the actual position difference is less than the preset position difference, the core position vector is updated to a position vector corresponding to the core position vector among the plurality of second position vectors, and for each of the second position vectors, the second position vector is updated using the core position vector and the weight value to obtain the candidate position vector; If there are one or more second position vectors whose position deviation rate compared to the core position vector is outside the preset deviation range, then for each of the second position vectors, the second position vector is iteratively updated using the core position vector, the weight value and the candidate value until the position deviation rate of each of the second position vectors compared to the core position vector is within the preset deviation range, the core position vector is updated to the position vector corresponding to the core position vector among the multiple second position vectors, and for each of the second position vectors, the second position vector is updated using the core position vector and the weight value to obtain the candidate position vector.

6. The substructure modal identification method capable of multiple step-by-step measurements according to claim 3, characterized in that: The method of updating the plurality of initial position vectors by using the plurality of candidate position vectors comprises: Determining the fitness of each candidate position vector based on the objective function; If there is a fitness of the candidate position vector corresponding to the core position vector among the other candidate position vectors which is smaller than the fitness of the candidate position vector corresponding to the core position vector, the multiple candidate position vectors are updated to the multiple initial position vectors; the other candidate position vectors include position vectors among the multiple candidate position vectors except the candidate position vector corresponding to the core position vector; If the fitness of the candidate position vectors corresponding to the core position vector are greater than or equal to the fitness of other candidate position vectors, each candidate position vector is updated based on the Levy flight algorithm until there is a candidate position vector whose fitness is less than the fitness of the candidate position vector corresponding to the core position vector, and the multiple candidate position vectors are updated to the multiple initial position vectors.

7. The substructure modal identification method capable of multiple step-by-step measurements according to claim 3, characterized in that: The determining the fitness corresponding to each group of initial adjustment parameters based on the objective function includes: For each group of initial adjustment parameters, performing signal demodulation processing on the response data according to a nonlinear frequency modulation mode decomposition algorithm using the initial adjustment parameters to obtain at least one initial single component data; Determine the signal complexity of each initial single component data using the objective function; The minimum signal complexity in at least one initial single component data is determined as the fitness corresponding to the initial adjustment parameter.

8. A substructure modal identification device capable of multiple step-by-step measurements, characterized in that: include: An acquisition module, used to acquire response data of each measurement point in each substructure of the engineering structure; The response data of all measurement points of the same substructure are the response data collected synchronously at the same time; The response data of the measurement points of different substructures are the response data collected multiple times at different times and in different steps; Each substructure has at least one common measurement point with other substructure; A decomposition module is used to perform signal demodulation processing on each response data by using a nonlinear frequency modulation mode decomposition algorithm with the objective of minimizing the objective function, so as to obtain at least first-order target single component data; The objective function is used to evaluate the signal complexity of the at least one single component data; A transformation module, configured to obtain spectrum data based on each target single component data; An identification module, for obtaining, for each substructure, the amplitude of each order of vibration mode of each substructure according to the ratio of the amplitude of the same order spectrum of each measurement point in the spectrum data; The identification module is further used to adjust the vibration mode amplitudes of the common measurement points of each substructure to be consistent according to the vibration mode amplitudes of each order of each substructure, so as to obtain the vibration mode amplitudes of each order of all measurement points of the engineering structure; to determine the direction of the vibration mode based on minimization of the modal confidence criterion according to the vibration mode amplitudes of each order of all measurement points of the engineering structure; and to obtain the modal vibration mode of the engineering structure according to the direction of the vibration mode.

9. A terminal device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 7 is implemented.

Images