Gate signal analysis method based on EMD and COA-LSTM

Abstract

The invention relates to the technical field of hydropower station management, in particular to a gate signal analysis method based on EMD and COA-LSTM, and the method comprises the steps: collecting the vibration data of a gate acceleration sensor; constructing an EMD algorithm and an LSTM algorithm; carrying out empirical mode decomposition on the collected vibration data of the gate by adopting an EMD (Empirical Mode Decomposition) algorithm; a subwolf optimization algorithm is proposed, the subwolf optimization algorithm is combined with an EMD algorithm and an LSTM algorithm, the EMD algorithm and the LSTM algorithm are optimized, a COA-LSTM algorithm is formed, and a future gate vibration signal is accurately predicted. According to the method provided by the invention, the gate vibration signal is trained through the algorithm model, the corresponding prediction model is established, the prediction of the gate vibration frequency of the hydropower station is realized, and the efficiency and the precision of the prediction model are improved by optimizing the parameters in the model.

Description

Technical Field

[0001] This invention relates to the field of hydropower station management technology, specifically to a gate signal analysis method based on EMD and COA-LSTM. Background Technology

[0002] Floodgates at hydropower stations are a crucial component of water conservancy projects. By opening and closing these gates, they regulate water levels and release floodwaters. The high-performance operation of the gate control system bears the heavy responsibility of ensuring the rational use of water resources, flood control at the power station, and the safety of life and property upstream and downstream, as well as the safety of the water conservancy project itself. Therefore, floodgate systems require real-time fault detection capabilities to ensure safe operation of the gates.

[0003] Gate malfunctions are closely related to their vibration frequency. While gates typically vibrate to some extent during normal operation, the vibration frequency and pattern change under fault or abnormal conditions. Continuous monitoring of the gate's vibration frequency can quickly identify potential problems such as bearing wear, loose mechanical components, or hydraulic system failures. Therefore, understanding the gate's vibration frequency patterns and monitoring frequency changes is crucial for preventing and resolving gate malfunctions. Summary of the Invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by this invention is that the gate failure problem is closely related to the gate's vibration frequency. When the gate is operating normally, it usually generates a certain degree of vibration, but when there is a fault or abnormal condition, the vibration frequency and mode will change. By continuously monitoring the gate's vibration frequency, potential problems can be quickly detected, such as bearing wear, loose mechanical parts, or hydraulic system failure. The method of this invention is used to predict the gate's vibration frequency mode and monitor frequency changes, so as to prevent and solve the gate failure problem.

[0006] To address the aforementioned technical problems, this invention provides the following technical solution: collecting vibration data from a gate accelerometer; constructing an EMD algorithm and an LSTM algorithm; performing empirical mode decomposition on the collected gate vibration data using the EMD algorithm; proposing a coyote optimization algorithm, which is combined with the EMD and LSTM algorithms to optimize the EMD and LSTM algorithms, forming a COA-LSTM algorithm to accurately predict future gate vibration signals.

[0007] As a preferred embodiment of the gate signal analysis method based on EMD and COA-LSTM described in this invention, the gate vibration data includes gate vibration data obtained by analyzing gate stress, gate acceleration, gate displacement, and gate tilt angle.

[0008] The EMD algorithm includes performing empirical mode decomposition on the gate vibration data. The principle of the EMD algorithm is as follows:

[0009] Find the maxima and minima of the original signal X(t), and fit the extreme points using curve interpolation to obtain the upper envelope X of the signal. max (t) and lower envelope X min (t);

[0010] Calculate the average value of the routes for the upper and lower packages:

[0011]

[0012] Subtract the original signal X(t) from the average envelope m1(t) to obtain the remaining signal d1(t);

[0013] For a stationary signal, the stationary signal is the first mode function (IMF) of the original signal X(t);

[0014] For non-stationary signals, which are not monotonically increasing within a certain region, inflection points will appear. Inflection points can reflect the specific characteristics of the original signal X(t). If an inflection point is not selected, the obtained first-order mode function will be inaccurate, that is, the obtained d1(t) will not satisfy the two conditions of IMF, and further screening will be carried out.

[0015] As a preferred embodiment of the gate signal analysis method based on EMD and COA-LSTM described in this invention, the EMD algorithm further includes:

[0016] The average value of the upper and lower packet paths is calculated for the remaining signal d1(t) until the screening threshold value is less than the threshold value, thus obtaining the appropriate first-order mode component c1(t), i.e., the first IMF. The screening threshold value SD is calculated as follows:

[0017]

[0018] The difference between signal X(t) and c1(t) is calculated to obtain the first-order residual r1(t). r1(t) is then used to replace the original signal for iterative processing. After repeating this process n times, the nth-order mode function c1(t) can be obtained. n (t) and the final compliant residual r n The expression for the original signal X(t) after EMD decomposition is:

[0019]

[0020] Among them, c n (t) is the nth-order mode function, r n (t) represents the final residual that meets the standard, and n is a parameter.

[0021] As a preferred embodiment of the gate signal analysis method based on EMD and COA-LSTM described in this invention, the LSTM algorithm includes an RNN model used to characterize time-related sequences. The decomposed data is fed into the LSTM algorithm, which uses multiple gate functions to determine the information to be forgotten, the information to be remembered, and the information to be output, thereby selectively transmitting information.

[0022] LSTM mainly consists of three gate structures: forget gate, memory gate, and output gate. These gates interact to improve the information analysis capability of the LSTM model. The LSTM algorithm structure is set as a cellular structure, and the specific algorithm settings are as follows:

[0023] Let c k Let x represent the cell state at time k. k h represents the input data at time k. k This represents the hidden state output at time k;

[0024] The LSTM cell at time k will receive three types of input information: c k-1 h k-1 and x k x k Input information is first sent to the forget gate and the remember gate for processing. The forget gate is used to forget information, and the forget gate determines c. k-1 Memory gates are used to filter out which information will be forgotten. k-1 and x k Information that needs to be retained, forget gate f k and memory gate i k and the current input cell state g in the memory gate k The specific calculation formula is as follows:

[0025]

[0026] i k =σ(W i ·[h k-1 ,x k ]+b i )

[0027] g k =tanh(W g ·[h k-1 ,x k ]+b g )

[0028] Among them, W f Here is the weight matrix of the forget gate, [h t-1 ,xt The symbol ] represents concatenating two vectors into a longer vector, b. f It is the bias term of the forget gate, σ is the sigmoid function; W i Here is the weight matrix of the memory gates, [h k-1 ,x k The symbol ] represents concatenating two vectors into a longer vector, b. i It is the bias term of the memory gate; W g It is the weight matrix of the tanh layer, b g It is the bias term of the tanh layer;

[0029] After the information is processed through the forget gate and the memory gate, the cell state c at time k k This will be updated accordingly:

[0030] c k =f k *c k-1 +i k *g k

[0031] Output gate o k h k-1 x k and c k The information is integrated to generate an output signal:

[0032] o k =σ(W o [h k-1 ,x k ]+b o )

[0033] h k =o k *tanh(c k )

[0034] Among them, o k W is the output signal of the output gate. o For the output gate weight matrix, [h k-1 ,x k The symbol ] represents concatenating two vectors into a longer vector, b. o This is the output gate bias term.

[0035] As a preferred embodiment of the gate signal analysis method based on EMD and COA-LSTM described in this invention, the coyote optimization algorithm includes the Coyote Optimization Algorithm (COA), which is an intelligent biomimetic optimization algorithm. The COA algorithm maintains high population diversity while improving convergence efficiency. Using the COA algorithm, weight parameters are optimized and calculated. The specific process of the COA algorithm is as follows:

[0036] Set up a coyote pack Np Each pack contains N individual coyotes. c Dimension D and the termination condition nfevalMAX parameter,

[0037] A coyote pack is randomly initialized. The i-th coyote individual in the p-th pack at time t is defined as follows:

[0038]

[0039]

[0040] Among them, ub j lb j Let r represent the upper and lower bounds of the j-th dimension, respectively. j A real number in the range [0,1] that is randomly generated;

[0041] The formula for evaluating the adaptability of coyotes is as follows:

[0042]

[0043] Coyotes sometimes break away from or are driven out of their original pack, resulting in pack migration, the probability of which is defined as P. e :

[0044]

[0045] Find the alpha wolf in the current group. p,t Calculate the current cultural trends of the coyote pack. p,t :

[0046]

[0047]

[0048] in, Indicates when N c When the number is odd, it is the median of the j-th dimension variable of all coyotes in group p at time t.

[0049] As a preferred embodiment of the gate signal analysis method based on EMD and COA-LSTM described in this invention, the coyote optimization algorithm further includes simulating birth and death events in genetics: the coyote's age is denoted as... The birth of the new coyotes (pup) p,t It is written as a combination of the parents' social status and environmental influences:

[0050]

[0051] Where m1 and m2 are random coyotes from the p wolf pack, j1 and j2 are two random dimensions of the problem, and Rj with rand j All numbers are random numbers within the range [0,1] generated by uniform probability, with discrete probability P. s and the correlation probability P a Influencing cultural diversity among individuals within a coyote pack, P s and P a Defined as:

[0052]

[0053] Calculate the effects δ1 and δ2 of the alpha wolf and pack culture trends on individual renewal within the coyote pack at the current moment:

[0054]

[0055] Where cr1 and cr2 represent random coyotes in the current swarm;

[0056] The new coyote individuals are obtained by sequentially updating all individual coyotes within the coyote pack. The fitness of the new coyotes compared to the original coyotes is selected based on their relative fitness, and the best coyote is retained.

[0057]

[0058]

[0059] Where r1 and r2 are real numbers in the range [0,1] generated with uniform probability, representing the weight of individual coyotes affected by alpha wolves and group cultural trends;

[0060] Simulate the process of an individual growing over time to update the age of coyotes;

[0061] Determine the termination condition. If it is met, output the social state of the coyote with the best adaptability; otherwise, return to the coyote's adaptability and continue the algorithm steps.

[0062] As a preferred embodiment of the gate signal analysis method based on EMD and COA-LSTM described in this invention, the COA-LSTM algorithm includes: inputting the weight parameters optimized by the COA algorithm into the forget gate, memory gate and output gate of the LSTM algorithm as a weight matrix to form the COA-LSTM algorithm; inputting the gate vibration data decomposed by the EMD algorithm into the COA-LSTM algorithm to predict the future vibration signal of the gate.

[0063] Another objective of this invention is to provide a system for gate signal analysis based on EMD and COA-LSTM, which can predict future vibration signals of gates by constructing a gate signal analysis system, thereby preventing and solving gate failure problems.

[0064] As a preferred embodiment of the gate signal analysis system based on EMD and COA-LSTM described in this invention, the system includes an EMD algorithm module, an LSTM algorithm module, a COA algorithm module, and a COA-LSTM algorithm module. The EMD algorithm module stores the EMD algorithm and is used to perform empirical mode decomposition on the collected gate vibration data. The LSTM algorithm module stores the LSTM algorithm. The COA algorithm module stores the COA algorithm and is used to optimize the weight matrix parameters in the LSTM algorithm module. The COA-LSTM algorithm module is used to input the weight matrix parameters optimized by the COA algorithm into the LSTM algorithm to form the COA-LSTM algorithm, and then input the gate vibration data empirically decomposed by the EMD algorithm module to predict future gate vibration signals.

[0065] The present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, characterized in that the processor executes the computer program to implement the steps of the gate signal analysis method based on EMD and COA-LSTM.

[0066] The present invention also provides a computer-readable storage medium having a computer program stored thereon, characterized in that, when the computer program is executed by a processor, it implements the steps of the gate signal analysis method based on EMD and COA-LSTM.

[0067] The beneficial effects of the present invention are as follows: The method provided by the present invention trains the gate vibration signal through an algorithm model, establishes a corresponding prediction model, and realizes the prediction of the gate vibration frequency of hydropower stations. By optimizing the parameters in the model, the efficiency and accuracy of the prediction model are improved. Attached Figure Description

[0068] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein:

[0069] Figure 1 The present invention provides an overall flowchart of a gate signal analysis method based on EMD and COA-LSTM.

[0070] Figure 2 The diagram shows the LSTM cell structure of a gate signal analysis method based on EMD and COA-LSTM provided by this invention.

[0071] Figure 3 The diagram shows the EMD algorithm decomposition results of a gate signal analysis method based on EMD and COA-LSTM provided by this invention.

[0072] Figure 4 Error diagram of prediction results for a gate signal analysis method based on EMD and COA-LSTM provided by this invention. Detailed Implementation

[0073] To make the above-mentioned objects, features, and advantages of the present invention more readily understood, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0074] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0075] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0076] This invention is described in detail with reference to the schematic diagrams. When detailing the embodiments of this invention, for ease of explanation, the cross-sectional views illustrating the device structure may be partially enlarged, not adhering to the usual scale. Furthermore, the schematic diagrams are merely examples and should not be construed as limiting the scope of protection of this invention. In actual fabrication, the three-dimensional spatial dimensions of length, width, and depth should be included.

[0077] Furthermore, in the description of this invention, it should be noted that the terms "upper," "lower," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. These terms are used solely for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. In addition, the terms "first," "second," or "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0078] Unless otherwise explicitly specified and limited, the terms "installation," "connection," and "joining" in this invention should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; similarly, they can refer to mechanical connections, electrical connections, or direct connections, or indirect connections through an intermediate medium, or internal connections between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0079] Example 1

[0080] Reference Figure 1 and Figure 2 This is the first embodiment of the present invention, which provides a gate signal analysis method based on EMD and COA-LSTM.

[0081] S1: Collect gate data.

[0082] The gate vibration data includes data obtained by analyzing gate stress, gate acceleration, gate displacement, and gate tilt angle.

[0083] S2: Construct an EMD algorithm to perform modal decomposition on gate vibration data.

[0084] Specifically, the EMD algorithm includes performing empirical mode decomposition on the gate vibration data using the EMD algorithm. The principle of the EMD algorithm is as follows:

[0085] Step a: Find the maximum and minimum points of the original signal X(t), and fit the extreme points using curve interpolation to obtain the upper envelope X of the signal. max (t) and lower envelope X min (t).

[0086] Step b: Calculate the average value of the routes for the upper and lower packages:

[0087]

[0088] Step c: Subtract the original signal X(t) from the average envelope m1(t) to obtain the remaining signal d1(t);

[0089] For a stationary signal, the stationary signal is the first mode function (IMF) of the original signal X(t);

[0090] For non-stationary signals, which are not monotonically increasing in a certain region, inflection points will appear. Inflection points can reflect the specific characteristics of the original signal X(t). If an inflection point is not selected, the obtained first-order mode function will be inaccurate, that is, the obtained d1(t) does not meet the two conditions of IMF and further screening is required.

[0091] Step d: Process the remaining signal d1(t) using steps a to c until a screening threshold is reached (typically 0.2-0.3; processing stops only when the threshold is less than the threshold). This yields a suitable first-order modal component c1(t), i.e., the first IMF. The screening threshold SD is calculated as follows:

[0092]

[0093] Step e: Calculate the difference between signal X(t) and c1(t) to obtain the first-order residual r1(t). Replace the original signal with r1(t) and perform the processing from steps a to e. Repeat this process n times to obtain the nth-order mode function c. n (t) and the final compliant residual r n The expression for the original signal X(t) after EMD decomposition is:

[0094]

[0095] Among them, c n (t) is the nth-order mode function, r n (t) represents the final residual that meets the standard, and n is a parameter.

[0096] S3: Construct the LSTM algorithm.

[0097] It should be noted that the LSTM algorithm is a special RNN model that can well characterize time-related sequences. LSTM uses multiple gating functions to determine the information to be forgotten, the information to be remembered, and the information to be output, thus selectively transmitting information.

[0098] Reference Figure 2 , is a diagram of the LSTM cell structure.

[0099] Specifically, the LSTM algorithm includes an RNN model used to characterize time-related sequences. The decomposed data is fed into the LSTM algorithm, which uses multiple gating functions to determine the information to be forgotten, the information to be remembered, and the information to be output, thereby selectively transmitting information.

[0100] LSTM mainly consists of three gate structures: forget gate, memory gate, and output gate. These gates interact to improve the information analysis capability of the LSTM model. The LSTM algorithm structure is set as a cellular structure, and the specific algorithm settings are as follows:

[0101] Let c k Let x represent the cell state at time k. k h represents the input data at time k. k This represents the hidden state output at time k.

[0102] The LSTM cell at time k will receive three types of input information: c k-1 h k-1 and x k x k Input information is first sent to the forget gate and the remember gate for processing. The forget gate is used to forget information, and the forget gate determines c. k-1 Memory gates are used to filter out which information will be forgotten. k-1 and x k Information that needs to be retained, forget gate f k and memory gate i k and the current input cell state g in the memory gate k The specific calculation formula is as follows:

[0103] f k =σ(W f ·[h t-1 ,x t ]+b f )

[0104] i k =σ(W i ·[h k-1 ,x k ]+b i )

[0105] g k =tanh(W g ·[h k-1 ,x k ]+b g )

[0106] Among them, W f Here is the weight matrix of the forget gate, [h t-1 ,x t The symbol ] represents concatenating two vectors into a longer vector, b. f It is the bias term of the forget gate, σ is the sigmoid function; W i Here is the weight matrix of the memory gates, [h k-1 ,x k The symbol ] represents concatenating two vectors into a longer vector, b.i It is the bias term of the memory gate; W g It is the weight matrix of the tanh layer, b g It is the bias term of the tanh layer.

[0107] After the information is processed through the forget gate and the memory gate, the cell state c at time k k This will be updated accordingly:

[0108] c k =f k *c k-1 +i k *g k

[0109] Output gate o k h k-1 x k and c k The information is integrated to generate an output signal:

[0110] o k =σ(W o [h k-1 ,x k ]+b o )

[0111] h k =o k *tanh(c k )

[0112] Among them, o k W is the output signal of the output gate. o For the output gate weight matrix, [h k-1 ,x k The symbol ] represents concatenating two vectors into a longer vector, b. o This is the output gate bias term.

[0113] S4: Construct the Coyote COA optimization algorithm.

[0114] The Coyote Optimization Algorithm (COA) is an intelligent biomimetic optimization algorithm. The COA algorithm maintains high population diversity while improving convergence efficiency. Using the COA algorithm, weight parameters are optimized and calculated. The specific process of the COA algorithm is as follows:

[0115] Set up a coyote pack N p Each pack contains N individual coyotes. c Dimension D and the termination condition nfevalMAX parameter;

[0116] A coyote pack is randomly initialized. The i-th coyote individual in the p-th pack at time t is defined as follows:

[0117]

[0118]

[0119] Among them, ub j lb j Let r represent the upper and lower bounds of the j-th dimension, respectively. j It is a randomly generated real number in the range [0,1].

[0120] The formula for evaluating the adaptability of coyotes is as follows:

[0121]

[0122] Coyotes sometimes break away from or are driven out of their original pack, resulting in pack migration, the probability of which is defined as P. e :

[0123]

[0124] Find the alpha wolf in the current group. p,t Calculate the current cultural trends of coyote packs using clur p,t :

[0125]

[0126]

[0127] in, Indicates when N c When the number is odd, it is the median of the j-th dimension variable of all coyotes in group p at time t.

[0128] The coyote optimization algorithm also includes simulating birth and death events in genetics: the coyote's age is denoted as... The birth of the new coyotes (pup) p,t It is written as a combination of the parents' social status and environmental influences:

[0129]

[0130] Where m1 and m2 are random coyotes from the p wolf pack, j1 and j2 are two random dimensions of the problem, and R j with rand j All numbers are random numbers within the range [0,1] generated by uniform probability, with discrete probability P. s and the correlation probability P a Influencing cultural diversity among individuals within a coyote pack, P s and P a Defined as:

[0131]

[0132] Calculate the effects δ1 and δ2 of the alpha wolf and pack culture trends on individual renewal within the coyote pack at the current moment:

[0133]

[0134] Here, cr1 and cr2 represent random coyotes in the current swarm.

[0135] The new coyote individuals are obtained by sequentially updating all individual coyotes within the coyote pack. The fitness of the new coyotes compared to the original coyotes is selected based on their relative fitness, and the best coyote is retained.

[0136]

[0137]

[0138] Where r1 and r2 are real numbers in the range [0,1] generated with uniform probability, representing the weight of individual coyotes affected by alpha wolves and group cultural trends.

[0139] The process of simulating an individual's growth over time is used to update the age of coyotes.

[0140] Determine the termination condition. If it is met, output the social state of the coyote with the best adaptability; otherwise, return to the coyote's adaptability and continue the algorithm steps.

[0141] S5: Construct the COA-LSTM algorithm to predict future gate vibration signals.

[0142] Specifically, the weight parameters optimized by the COA algorithm are input into the forget gate, memory gate, and output gate of the LSTM algorithm as weight matrices to form the COA-LSTM algorithm. The gate vibration data decomposed by the EMD algorithm are then input into the COA-LSTM algorithm to predict the future vibration signal of the gate.

[0143] Example 2

[0144] Reference Figure 2 This is the second embodiment of the present invention, which provides a system for gate signal analysis based on EMD and COA-LSTM.

[0145] The system includes an EMD algorithm module, an LSTM algorithm module, a COA algorithm module, and a COA-LSTM algorithm module.

[0146] The EMD algorithm module stores the EMD algorithm, which is used to perform empirical mode decomposition on the collected gate vibration data.

[0147] The LSTM algorithm module stores the LSTM algorithm.

[0148] The COA algorithm module stores the COA algorithm, which is used to optimize the weight matrix parameters in the LSTM algorithm module.

[0149] The COA-LSTM algorithm module is used to input the weight matrix parameters selected by the COA algorithm into the LSTM algorithm to form the COA-LSTM algorithm. It inputs the gate vibration data that has been empirically decomposed by the EMD algorithm module to predict the future gate vibration signal.

[0150] Example 3

[0151] Reference Figure 3 and Figure 4 This is the third embodiment of the present invention, which provides an implementation process and a comparison of beneficial effects of a gate signal analysis method based on EMD and COA-LSTM.

[0152] Read the vibration frequency of the gate's acceleration sensor.

[0153] The obtained vibration data of the gate is first processed using the EMD algorithm for empirical mode decomposition, and then the decomposition results are as follows: Figure 3 As shown.

[0154] The data obtained from empirical mode decomposition are fed into the COA-LSTM algorithm to determine the vibration frequency.

[0155] Gate vibration frequency prediction Figure 4 As shown, the error after reconstruction is relatively small.

[0156] Referring to Table 1, a comparison of the prediction accuracy of the model method of this invention with existing model methods is presented.

[0157]

[0158] As can be seen from Table 1, compared with the existing LSTM algorithm model, the present invention has significantly improved prediction accuracy and shortened working cycle, thus improving work efficiency.

[0159] Example 4

[0160] The fourth embodiment of the present invention differs from the previous embodiment in that:

[0161] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0162] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0163] More specific examples of computer-readable media (a non-exhaustive list) include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0164] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0165] Furthermore, in order to provide a concise description of exemplary embodiments, not all features of actual embodiments (i.e., those features that are not relevant to the currently considered best mode for carrying out the invention, or those features that are not relevant to implementing the invention) may be omitted.

[0166] It should be understood that numerous specific implementation decisions can be made during the development of any practical implementation, such as in any engineering or design project. Such development efforts may be complex and time-consuming, but for those skilled in the art who benefit from this disclosure, the development effort will be a routine task in design, manufacturing, and production without requiring extensive experimentation.

[0167] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A gate signal analysis method based on EMD and COA-LSTM, characterized in that: include, Collect vibration data from the gate's acceleration sensor; Construct EMD and LSTM algorithms; The collected vibration data of the gate were subjected to empirical mode decomposition using the EMD algorithm. A coyote optimization algorithm is proposed and combined with the EMD and LSTM algorithms to optimize the EMD and LSTM algorithms, forming the COA-LSTM algorithm, which can accurately predict future gate vibration signals.

2. The gate signal analysis method based on EMD and COA-LSTM as described in claim 1, characterized in that: The gate vibration data includes data obtained by analyzing gate stress, gate acceleration, gate displacement, and gate tilt angle. The EMD algorithm includes performing empirical mode decomposition on the gate vibration data. The principle of the EMD algorithm is as follows: Find the maxima and minima of the original signal X(t), and fit the extreme points using curve interpolation to obtain the upper envelope X of the signal. max (t) and lower envelope X min (t); Calculate the average value of the routes for the upper and lower packages: Subtract the original signal X(t) from the average envelope m1(t) to obtain the remaining signal d1(t); For a stationary signal, the stationary signal is the first mode function (IMF) of the original signal X(t); For non-stationary signals, which are not monotonically increasing within a certain region, inflection points will appear. Inflection points can reflect the specific characteristics of the original signal X(t). If an inflection point is not selected, the obtained first-order mode function will be inaccurate, that is, the obtained d1(t) will not satisfy the two conditions of IMF, and further screening will be carried out.

3. The gate signal analysis method based on EMD and COA-LSTM as described in claim 2, characterized in that: The EMD algorithm also includes, The average value of the upper and lower packet paths is calculated for the remaining signal d1(t) until the screening threshold value is less than the threshold value, thus obtaining the appropriate first-order mode component x1(t), i.e., the first IMF. The screening threshold value SD is calculated as follows: The difference between signal X(t) and c1(t) is calculated to obtain the first-order residual r1(t). r1(t) is then used to replace the original signal for iterative processing. After repeating this process n times, the nth-order mode function c1(t) can be obtained. n (t) and the final compliant residual r n The expression for the original signal X(t) after EMD decomposition is: Among them, c n (t) is the nth-order mode function, r n (t) represents the final residual that meets the standard, and n is a parameter.

4. The gate signal analysis method based on EMD and COA-LSTM as described in claim 3, characterized in that: The LSTM algorithm includes an RNN model used to characterize time-related sequences. The decomposed data is fed into the LSTM algorithm. LSTM uses multiple gate functions to determine the information to be forgotten, the information to be remembered, and the information to be output, thereby selectively transmitting information. LSTM mainly consists of three gate structures: forget gate, memory gate, and output gate. These gates interact to improve the information analysis capability of the LSTM model. The LSTM algorithm structure is set as a cellular structure, and the specific algorithm settings are as follows: Let c k Let x represent the cell state at time k. k h represents the input data at time k. k This represents the hidden state output at time k; The LSTM cell at time k will receive three types of input information: c k-1 h k-1 and x k x k Input information is first sent to the forget gate and the remember gate for processing. The forget gate is used to forget information, and the forget gate determines c. k-1 Memory gates are used to filter out which information will be forgotten. k-1 and x k Information that needs to be retained, forget gate f k and memory gate i k and the current input cell state g in the memory gate k The specific calculation formula is as follows: f k =σ(W f ·[h t-1 ,x t ]+b f ) i k =σ(W i ·[h k-1 ,x k ]+b i ) rice k = tanh(W g ·[h k-1 x k ]+b g ) Among them, W f Here is the weight matrix of the forget gate, [h t-1 x t The symbol ] represents concatenating two vectors into a longer vector, b. f It is the bias term of the forget gate, σ is the sigmoid function; W i Here is the weight matrix of the memory gates, [h k-1 x k The symbol ] represents concatenating two vectors into a longer vector, b. i It is the bias term of the memory gate; W g It is the weight matrix of the tanh layer, b g It is the bias term of the tanh layer; After the information is processed through the forget gate and the memory gate, the cell state c at time k k This will be updated accordingly: c k =f k *c k-1 +i k *g k Output gate o k h k-1 x k and c k The information is integrated to generate an output signal: the k =σ(W o [h k-1 ,x k ]+b o ) h k = no k *fish(c k ) Among them, o k W is the output signal of the output gate. o For the output gate weight matrix, [h k-1 ,x k The symbol ] represents concatenating two vectors into a longer vector, b. o This is the output gate bias term.

5. The gate signal analysis method based on EMD and COA-LSTM as described in claim 4, characterized in that: The Coyote Optimization Algorithm (COA) is an intelligent biomimetic optimization algorithm. The COA algorithm maintains high population diversity while improving convergence efficiency. Using the COA algorithm, weight parameters are optimized and calculated. The specific process of the COA algorithm is as follows: Set up a coyote pack N p Each pack contains N individual coyotes. c Dimension D and the termination condition nfevalMAX parameter, A coyote pack is randomly initialized. The i-th coyote individual in the p-th pack at time t is defined as follows: Among them, ub j lb j Let r represent the upper and lower bounds of the j-th dimension, respectively. j A real number in the range [0,1] that is randomly generated; The formula for evaluating the adaptability of coyotes is as follows: Coyotes sometimes break away from or are driven out of their original pack, resulting in pack migration, the probability of which is defined as P. e : Find the alpha wolf in the current group. p,t Calculate the current cultural trends of the coyote pack. p,t : in, Indicates when N c When the number is odd, it is the median of the j-th dimension variable of all coyotes in group p at time t.

6. The gate signal analysis method based on EMD and COA-LSTM as described in claim 5, characterized in that: The coyote optimization algorithm also includes simulating birth and death events in genetics: the coyote's age is denoted as... The birth of the new coyotes (pup) p,t It is written as a combination of the parents' social status and environmental influences: Where m1 and m2 are random coyotes from the p wolf pack, j1 and j2 are two random dimensions of the problem, and R j with rand j All numbers are random numbers within the range [0,1] generated by uniform probability, with discrete probability P. s and the correlation probability P a Influencing cultural diversity among individuals within a coyote pack, P s and P a Defined as: Calculate the effects δ1 and δ2 of the alpha wolf and pack culture trends on individual renewal within the coyote pack at the current moment: Where cr1 and cr2 represent random coyotes in the current swarm; The new coyote individuals are obtained by sequentially updating all individual coyotes within the coyote pack. The fitness of the new coyotes compared to the original coyotes is selected based on their relative fitness, and the best coyote is retained. Where r1 and r2 are real numbers in the range [0,1] generated with uniform probability, representing the weight of individual coyotes affected by alpha wolves and group cultural trends; Simulate the process of an individual growing over time to update the age of coyotes; Determine the termination condition. If it is met, output the social state of the coyote with the best adaptability; otherwise, return to the coyote's adaptability and continue the algorithm steps.

7. The gate signal analysis method based on EMD and COA-LSTM as described in claim 6, characterized in that: The COA-LSTM algorithm includes: inputting the weight parameters optimized by the COA algorithm into the forget gate, memory gate and output gate of the LSTM algorithm as a weight matrix to form the COA-LSTM algorithm; inputting the gate vibration data decomposed by the EMD algorithm into the COA-LSTM algorithm to predict the future vibration signal of the gate.

8. A system employing a gate signal analysis method based on EMD and COA-LSTM as described in any one of claims 1 to 7, characterized in that: The system includes an EMD algorithm module, an LSTM algorithm module, a COA algorithm module, and a COA-LSTM algorithm module; The EMD algorithm module stores the EMD algorithm, which is used to perform empirical mode decomposition on the collected gate vibration data. The LSTM algorithm module stores the LSTM algorithm; The COA algorithm module stores the COA algorithm, which is used to optimize the weight matrix parameters in the LSTM algorithm module. The COA-LSTM algorithm module is used to input the weight matrix parameters selected by the COA algorithm into the LSTM algorithm to form the COA-LSTM algorithm. It inputs the gate vibration data that has been empirically decomposed by the EMD algorithm module to predict the future gate vibration signal.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the gate signal analysis method based on EMD and COA-LSTM as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the gate signal analysis method based on EMD and COA-LSTM as described in any one of claims 1 to 7.

Images