Sensing layout method for fault diagnosis of piston connecting rod system

Abstract

The invention discloses a piston connecting rod system fault diagnosis-oriented sensing layout method in a diesel engine, which comprises the following steps of: acquiring data under each sensor pre-layout scheme, obtaining a feature vector of a sensor through a data preprocessing module, taking the feature vector as an LSTM neural network input, taking an actual fault and a normal working condition number as outputs, and obtaining a fault diagnosis result; training to obtain a fault diagnosis model, taking the test data set as the input of the fault diagnosis model, outputting a predicted working condition number, calculating to obtain the fault diagnosis accuracy, taking the fault diagnosis accuracy and a feature vector corresponding to each scheme as the input of a binary system litsea coreana algorithm sensing network optimization module, and taking the binary system code of the optimal sensing layout as the output; the sensing number, the information entropy and the fault diagnosis rate are used as evaluation indexes, redundant information is removed, the sensing number is reduced, and the accuracy of fault diagnosis is ensured.

Description

Technical Field

[0001] The present invention relates to the field of sensor network layout optimization and fault diagnosis, and in particular to a sensor network system and method for fault diagnosis of a piston connecting rod system in a diesel engine. Background Art

[0002] As the core component of the diesel engine's normal operation and efficient work, the piston-connecting rod system undertakes the key task of efficiently converting the piston's linear reciprocating motion into the crankshaft's rotational motion. At present, the operating status information of the piston-connecting rod system mainly relies on its sensor network. However, the number and location of sensors have a significant impact on the performance and reliability of the monitoring system. Improperly configured sensor networks will increase costs and reduce the accuracy of troubleshooting of the monitoring system.

[0003] The sensor network optimization layout disclosed in the document with the Chinese patent application publication number CN116341373A and the name "A sensor network optimization layout method" obtains the overall modal information of the structure by performing modal simulation analysis on the structural deformation, initializes the number of sensors and the layout position, and optimizes the layout by using the excellent local search ability of the particle swarm algorithm and the excellent global search ability of the fruit fly algorithm. The sensor network layout disclosed in the document with the Chinese patent application publication number CN109751113A and the name "Sensor network layout optimization method for mine microseismic monitoring and its application" constructs the optimization function of the sensor network layout of the unknown wave velocity system through the CRLB principle of optimal parameter estimation, and solves the optimization function by using the encoding operator, selection operator, crossover operator, mutation operator, elite retention strategy and convergence criterion in the improved genetic algorithm to obtain the optimal sensor network layout scheme. However, when the diesel engine is working, its piston connecting rod system is in a non-open state and cannot perform structural deformation simulation. Therefore, these methods are not suitable for piston connecting rod systems that perform linear reciprocating motion. Summary of the invention

[0004] The purpose of the present invention is to solve the problem that the existing sensor network layout is not suitable for a piston-connecting rod system performing linear reciprocating motion, and to propose a sensor network system and method for piston-connecting rod system fault diagnosis to achieve accurate diagnosis of piston-connecting rod faults.

[0005] The present invention provides a fault diagnosis method of a sensor network system for piston connecting rod system fault diagnosis using the following technical solutions:

[0006] Step 1): Pre-layout based on experience 2 n -1 sensor layout, n is the number of units under test, data of each sensor is collected under each sensor pre-layout scheme and divided into training data set and test data set, and the sensor feature vector is obtained through the data preprocessing module;

[0007] Step 2): Take the feature vector as the input of the LSTM neural network, and the actual fault and normal working condition numbers of the piston connecting rod system as the output, and train to obtain a fault diagnosis model;

[0008] Step 3): Use the test data set as the input of the fault diagnosis model, output the predicted operating condition number through the fault diagnosis model, and calculate 2 n - Fault diagnosis accuracy under 1 sensor layout;

[0009] Step 4): The fault diagnosis accuracy and the characteristic vector corresponding to each scheme are used as inputs of the binary Sky Eagle algorithm sensor network optimization module, and the binary code of the optimal sensor layout is used as output.

[0010] Furthermore, the specific method of step 4) is:

[0011] (1) Set the binary Skyhawk algorithm parameters, randomly assign the population size, and generate a set of random n-dimensional binary codes for each individual;

[0012] (2) Calculate the information entropy of the feature vector corresponding to each randomly generated binary code, the proportion of sensors arranged in the binary code, and the fitness value;

[0013] (3) Arrange all individuals from large to small according to their fitness values, and select the individual with the highest fitness value as the current best position;

[0014] (4) Update the population position. After the update, normalize the position of each individual in the population, perform binary coding mapping on the normalized position, and randomly generate a set of n-dimensional [0,1] vectors. In the normalized individual position, if the value of a certain dimension of the normalized individual position is less than the corresponding value of the corresponding dimension of the vector, then the value of the dimension is set to not install the sensor, otherwise install the sensor;

[0015] (5) Recalculate the updated fitness value for the individual position in the mapped population, compare the fitness value of the updated population with the current best fitness value, and if the fitness value in the updated generation is better than the current best fitness value, then use the individual position and corresponding fitness value of the updated best population as the current best fitness value and the individual position of the best population;

[0016] (6) Determine whether the maximum number of iterations has been reached. After the iteration is completed, the position of the Sky Eagle population is the binary code of the optimal sensor network.

[0017] Furthermore, when updating the population position, the maximum number of iterations is H. If the current number of iterations h≤2H / 3 and the random number rand≤0.5 in the range [0,1], the updated individual position is:

[0018]

[0019] If rand>0.5, the updated individual position is:

[0020] X2(h+1)=X best (h)×Levy(n)+X R (h)+(po)×rand,

[0021] X best (h) represents the best solution to the hth iteration, X M (h) represents the average position of the population at the hth iteration, Levy (Dim) is the Levy flight distribution function, X R (h) is the random solution taken in the range [1, Num] at the hth iteration, Num is the population size, and p and o represent the search spiral shape.

[0022] If the current number of iterations h>2H / 3, and the random number rand>0.5 in the range [0,1], the updated individual position is:

[0023] X3(h+1)=(X best (h)-X M (h))×a-rand+((UB-LB)×rand+LB)×a;

[0024] If rand≤0.5, the updated individual position is:

[0025] X4(h+1)=QF×X best (h)-(G1×X(h)×rand)-G2×Levy(n)+rand×G1,

[0026] α is the exploitation adjustment parameter of 0.1, QF represents the quality function used to balance the search strategy, G1 represents the various actions used by predators to track prey during the hunting process, and G2 represents the flight slope of following the prey.

[0027] The advantages of the present invention after adopting the above technical solution are:

[0028] 1) The sensor network system for piston-connecting rod system fault diagnosis of the present invention integrates modules such as data preprocessing and binary Sky Eagle algorithm sensor network layout optimization, which reduces the complexity of the central processing unit and improves the reliability and economy of piston-connecting rod system fault diagnosis.

[0029] 2) The adopted modules such as the Sky Eagle algorithm sensor network layout optimization solve the problems of long modeling time and complex modeling process for multi-coupled and nonlinear systems when arranging sensors based on mechanism analysis methods.

[0030] 3) The sensor layout method for piston-connecting rod system fault diagnosis of the present invention uses the number of sensors, information entropy, and fault diagnosis rate as evaluation indicators. While ensuring the fault diagnosis rate, it removes redundant information, reduces the number of sensors, reduces costs, improves monitoring accuracy and computing efficiency, and ensures accurate fault diagnosis. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 It is a structural schematic diagram of a piston connecting rod system;

[0032] Figure 2 for Figure 1 Schematic diagram of structural force analysis of the piston-connecting rod system;

[0033] Figure 3 It is a structural block diagram of a sensor layout method for piston-connecting rod system fault diagnosis of the present invention;

[0034] Figure 4 Binary coding diagram for sensor networks;

[0035] Figure 5 for Figure 3 Flowchart of the execution of the data preprocessing module;

[0036] Figure 6 for Figure 3 Flowchart of the execution of the sensor network optimization module of the binary Skyhawk algorithm. DETAILED DESCRIPTION

[0037] like Figure 1 The piston-connecting rod system shown in the figure mainly includes a piston 1, a connecting rod 2, a bearing 3 and a crankshaft 4. It is a key power transmission component in a diesel engine, and its working process includes an intake stroke, a compression stroke, a power stroke and an exhaust stroke. Among them, the connecting rod 2 is a connecting piece between the piston 1 and the crankshaft 4, converting the linear reciprocating motion of the piston 1 into the continuous rotation motion of the crankshaft 4. The bearing 3 is located at the rotation fulcrum of the crankshaft 4 to ensure the smooth operation of the crankshaft 4.

[0038] like Figure 2 The force diagram of the piston-connecting rod system is shown in the figure. Point O is the rotation center of the crankshaft 4, point B is the center of the piston pin, and the crank OA of the crankshaft 4 rotates. When the piston-connecting rod assembly reciprocates in the cylinder, the total force on the piston pin is F. Let D be the diameter of the cylinder liner, m j is the reciprocating inertia mass of the piston-connecting rod assembly, R is the crank OA radius, ω is the crank angular velocity, θ is the crank angle, λ is the ratio of crank radius to connecting rod length, ν h (t) is the lateral speed of the piston, a his the lateral acceleration of the piston, δ is the distance between the piston and the cylinder liner, then the total force F on the piston pin is decomposed into the horizontal force F along the horizontal direction as follows: h and the force F along the connecting rod c , F h That is the thrust of the piston on the cylinder liner:

[0039]

[0040] The lateral impact momentum M of the piston-connecting rod assembly on the cylinder liner j for:

[0041]

[0042] From the above formula, we can know that the impact momentum M of the piston-connecting rod assembly on the cylinder liner is j It is related to the speed and time t of the diesel engine. When the diesel engine is working, when the piston 1 reaches the top dead center, it is affected by the inertia force and changes its movement trend. At this time, due to the change in the movement direction of the piston 1 and the force of the piston connecting rod assembly, the left side of the piston 1 is prone to collision with the cylinder wall. Similarly, at the bottom dead center, the right side of the piston 1 is prone to collision with the cylinder wall. Therefore, based on the analysis of the minimum vibration transmission path, for the wear failure of the piston ring and cylinder liner of a single-cylinder diesel engine, at least one vibration sensor needs to be arranged on the cylinder head for monitoring.

[0043] When piston 1 fails, the thrust of connecting rod 2 on crankshaft 4 is F c Its direction acts on the journal center A through the center of connecting rod 2, and can be decomposed into two mutually perpendicular forces F n and F t , causing bearing 3 to start wearing. When bearing 3 goes through a certain wear stage, the effective average oil film thickness of the lubricating oil decreases, accompanied by an increase in the amount of lubricating oil leakage and increased heat dissipation. In addition, severe wear conditions may cause the main journal to directly contact the bearing shell in the poorly lubricated area, resulting in an impact effect. Therefore, for a single-cylinder diesel engine, based on the above-mentioned arrangement of vibration sensors, 2 more sensors are needed to monitor the vibration and temperature changes of bearing 3. Therefore, for an N-cylinder diesel engine, there must be at least 3N units to be tested.

[0044] like Figure 3 As shown, for a diesel engine with a total of n units under test, each unit under test can be connected to a corresponding sensor. If a sensor is set for each unit under test, data redundancy and system complexity will increase, thereby affecting the rapid and accurate acquisition of key information. Therefore, the present invention proposes a sensor layout method for piston connecting rod system fault diagnosis to achieve more efficient and accurate monitoring functions.

[0045] like Figure 4As shown in the figure, the sensor layout problem can be regarded as a combinatorial binary optimization problem. Whether each unit under test is equipped with a sensor can be represented by a binary character vector Vectors = (S1, S2, ..., S n ) means that the elements S1, S2, ..., S n Indicates the unit under test, S i = 0 means that the i-th unit under test is not equipped with a sensor; on the contrary, S i =1 means that the sensor is installed on the i-th unit under test, 1≤i≤n.

[0046] Specifically, taking an N-cylinder diesel engine as an example, based on Figure 1-2 According to the above empirical analysis, when monitoring the piston-connecting rod system, there are at least 3N units to be tested, where 3N≤n, 2 n -1 sensor layout scheme. According to the above experience analysis, each sensor is pre-arranged. Under each sensor pre-arrangement scheme, the data of each sensor is collected. After the data is calculated by the data preprocessing module, the binary Skyhawk algorithm sensor network optimization module is used to find the binary code of the optimal sensor layout. The binary code is transmitted to the central processing unit through the bus communication module, and the central processing unit performs reverse analysis and determines the optimal sensor layout, such as Figure 3 shown.

[0047] like Figure 5 The data preprocessing process is shown in Figure 2. n -1 sensor layout scheme, the data collected by each sensor is used as input, and the fault diagnosis accuracy and feature vector are used as output. n -1 sensor layout scheme, the working conditions of each sensor under each sensor layout scheme are collected respectively, and the number of faulty and normal working conditions is s in total, and the corresponding actual faulty and normal working condition numbers are 1, 2, …s. There are 200 samples under each working condition, and a total of 200×s m-dimensional samples are collected, where m is the number of sensors under the current layout. 80% of the data in the 200×s m-dimensional samples are used as training data sets, and 20% are used as test data sets. Then, the data collected by each sensor in the training data set and the test data set are decomposed into three layers using wavelet packet transform technology to generate 8 frequency bands, and the energy entropy of each frequency band signal after decomposition is calculated as the feature vector of the sensor, forming an 8×m-dimensional feature vector Z (z1, z2, ..., z 8×m ), extract the feature vector Z(z1,z2,...,z 8×m ). This eigenvector Z(z1,z2,...,z 8×m ) as the input of the LSTM neural network and binary Sky Eagle algorithm sensor network optimization module.

[0048] In the LSTM neural network, the initial learning rate is set to 0.01. After 60 training times, the initial learning rate is adjusted. The learning rate adjustment factor is 0.2, the regularization parameter is 0.001, and the maximum number of training times is 100. Finally, the 8×m-dimensional feature vector of the training set is used as the input of the LSTM neural network, and the actual fault and normal working condition numbers 1, 2, …s of the piston connecting rod system are used as output. After sufficient training, the fault diagnosis model is obtained. Then, the pre-divided test data set is used as input, and the predicted working condition number is output through the fault diagnosis model. Finally, formula (5) is used to calculate 2 n -1 Fault diagnosis accuracy under one sensor layout F(2 n -1):

[0049]

[0050] Where d represents the number of samples in which the output predicted operating condition number is different from the actual fault and normal operating condition numbers 1, 2, …s during the test data set verification.

[0051] The fault diagnosis accuracy F(2 n -1) Input into the SkyEagle algorithm sensor network optimization module.

[0052] like Figure 6 The figure shows the flowchart of the binary Skyhawk algorithm sensor network optimization module execution. n -1 sensor layout scheme is built into the binary Skyhawk algorithm sensor network optimization module, which uses the fault diagnosis accuracy F(2 n -1) and the eigenvector Z(z1,z2,...,z 8×m ) as input and the binary code of the optimal sensor layout as output. The specific process is as follows:

[0053] (1) Set the parameters of the binary Skyhawk algorithm: population size Num, maximum number of iterations H, population dimension equal to the number of units under test, both n, search upper limit UB, search lower limit LB, and randomly initialize the position of the Skyhawk individual.

[0054] (2) Randomly assign the population size and generate a set of random n-dimensional binary codes for each individual, that is, the position of each individual in the population. The position of each individual in the population is a set of n-dimensional vectors, which is an n-dimensional binary code.

[0055] (3) Using formula (6), calculate the information entropy G(2) of each randomly generated binary code corresponding to the feature vector n -1):

[0056]

[0057] Among them, for the eigenvector Z(z1,z2,...,z 8×m ), P(z e ) is the state z e The probability distribution of e ) is calculated using formula (7):

[0058]

[0059] z j are the eigenvalues ​​corresponding to different dimensions of the eigenvector, j = 1, 2, … 8 × m.

[0060] (4) Use formula (8) to calculate the proportion of sensors that generate random binary codes:

[0061] K(2 n -1)=m / n (8)

[0062] (5) The proportion of sensors arranged based on binary coding K(2 n -1), fault diagnosis accuracy F(2 n -1) and information entropy G(2 n -1), and use formula (9) to calculate the fitness value fit(2 n -1) to assess the performance of each individual in the contemporary population:

[0063] fit(2 n -1)=2-F(2 n -1)-G(2 n -1)+K(2 n -1) (9)

[0064] (6) All individuals are converted into n -1) Arrange from large to small, and select the individual with the highest fitness value as the current best position X fit , the corresponding fitness value is the current optimal fitness value fit.

[0065] (7) Update the population position. If the current number of iterations h≤2H / 3, and the random number rand≤0.5 in the range [0,1], enter the expanded exploration phase X1, which is equivalent to exploring the search space from a global perspective to find possible solution areas. In this way, the algorithm can quickly identify the area where the potential optimal solution is located in a larger range. The specific update formula (10) is:

[0066]

[0067] Where X1(h+1) represents the individual position updated by the first search method, i.e., the expanded exploration phase X1, in the h+1th iteration of the Skyhawk algorithm. best (h) represents the best solution to the hth iteration, X M (h) represents the average position of the population at the hth iteration, H represents the maximum number of iterations, and rand is a random number in [0,1].

[0068] (8) If the current number of iterations h≤2H / 3 and rand>0.5, the algorithm enters the reduced exploration phase X2. This step means that after finding a potential solution, the algorithm will conduct a more concentrated local search in the area to approach the optimal solution. The specific update formula (11) is:

[0069] X2(h+1)=X best (h)×Levy(n)+X R (h)+(po)×rand (11)

[0070] Where X2(h+1) represents the individual position updated by the second search method, i.e., the reduced exploration phase X2, in the h+1th iteration of the Skyhawk algorithm; Levy(Dim) is the Levy flight distribution function; n is the population dimension; X R (h) is the random solution taken in the range [1, Num] at the hth iteration; p and o represent the search spiral shape.

[0071] (9) When the number of iterations h>2H / 3 and rand>0.5, the algorithm enters the expansion development stage (X3). This means that during the search process, the algorithm will take a more detailed and slow approach to gradually approach the optimal solution. The specific update process is formula (12).

[0072] X3(h+1)=(X best (h)-X M (h))×a-rand+((UB-LB)×rand+LB)×a (12)

[0073] Among them, X3(h+1) represents the individual position updated by the third search method (X3) in the h+1th iteration of the Sky Eagle algorithm, and α is the mining adjustment parameter of 0.1;

[0074] (10) When the current number of iterations h>2H / 3 and rand≤0.5, the algorithm enters the narrowing development phase X4, and performs final fine-tuning and confirmation to ensure that the optimal solution is found. The specific update formula (13) is:

[0075] X4(h+1)=QF×X best(h)-(G1×X(h)×rand)-G2×Levy(n)+rand×G1 (13)

[0076] Among them, X4(h+1) represents the individual position updated by the fourth search method, i.e., the narrowing development stage X4, in the h+1th iteration of the Skyhawk algorithm; QF represents the quality function used to balance the search strategy; G1 represents the various actions used by predators to track prey during the hunting process; G2 represents the flight slope of following the prey.

[0077] (11) After the population update is completed, the position of each individual in the population is normalized using formula (14):

[0078]

[0079] Where, X b (h+1) represents the individual position updated by the b-th search method in the h+1th iteration of the Skyhawk algorithm (where b = 1, 2, 3, 4). The position after normalization of the updated individual position of the population is T(X b (h+1)), at this time, the position of each individual is an n-dimensional vector of (0-1).

[0080] (12) Using formula (15), the normalized position T(X b (h+1)) performs binary coding mapping, randomly generates a set of n-dimensional [0,1] vectors r, and b (h+1)), compare each individual position with the random vector r. If the value of a dimension of the normalized individual position is less than the corresponding value of the corresponding dimension of the vector r, the value of the dimension is set to 0 and the sensor is not installed; otherwise, it is set to 1 and the sensor is installed:

[0081]

[0082] Where, X ij is the position of the individuals in the population after mapping.

[0083] (13) Reuse steps (3-5) to calculate the updated fitness value for the individual positions in the mapped population. Therefore, the individual positions at this time have been mapped into a set of binary codes. Sort the fitness values ​​and find the individual positions and corresponding fitness values ​​of the updated best population. Compare the fitness value of the updated population with the current best fitness value fit. If the fitness value in the updated generation is better than the current best fitness value fit, the individual position and corresponding fitness value of the updated best population are used as the current best fitness value and the individual position of the best population.

[0084] (14) Determine whether the maximum number of iterations H has been reached. If so, stop the iteration; otherwise, return to steps (3)-(13).

[0085] (15) After the iteration is completed, the position of the eagle population is the binary code of the optimal sensor network.

[0086] Find the optimal sensor layout code and transmit the code information to the central processing unit through the bus communication module. The central processing unit will perform reverse analysis and determine the sensor layout to improve monitoring accuracy.

Claims

1. A sensor layout method for piston-connecting rod system fault diagnosis, characterized in that The following steps are involved: Step 1): Pre-layout based on experience 2 n -1 sensor layout, n is the number of units under test, data of each sensor is collected under each sensor pre-layout scheme and divided into training data set and test data set, and the sensor feature vector is obtained through the data preprocessing module; Step 2): Take the feature vector as the input of the LSTM neural network, and the actual fault and normal working condition numbers of the piston connecting rod system as the output, and train to obtain a fault diagnosis model; Step 3): Use the test data set as the input of the fault diagnosis model, output the predicted operating condition number through the fault diagnosis model, and calculate 2 n - Fault diagnosis accuracy under 1 sensor layout; Step 4): The fault diagnosis accuracy and the characteristic vector corresponding to each scheme are used as inputs of the binary Sky Eagle algorithm sensor network optimization module, and the binary code of the optimal sensor layout is used as output.

2. A sensor layout method for piston-connecting rod system fault diagnosis according to claim 1, characterized in that: In step 1), the working condition of each sensor under each sensor layout scheme is collected, and the number of fault and normal working conditions is s in total. The corresponding actual fault and normal working condition numbers are 1, 2, ... s. There are 200 samples under each working condition, and a total of 200×s m-dimensional samples are collected, where m is the number of sensors under the current layout. 80% of the data in the 200×s m-dimensional samples are used as training data sets, and 20% are used as test data sets; The data collected by each sensor in the training data set and the test data set are decomposed into three layers using wavelet packet transform to generate 8 frequency bands. The energy entropy of each frequency band signal after decomposition is calculated as the feature vector of the sensor to form an 8×m-dimensional feature vector.

3. The sensor layout method for piston-connecting rod system fault diagnosis according to claim 2 is characterized in that: In step 3), the fault diagnosis accuracy d represents the number of samples where the predicted operating condition number output during the test data set verification is different from the actual fault and normal operating condition numbers 1, 2, …s.

4. The sensor layout method for piston-connecting rod system fault diagnosis according to claim 1 is characterized in that: Step 4) The specific method is: (1) Set the binary Skyhawk algorithm parameters, randomly assign the population size, and generate a set of random n-dimensional binary codes for each individual; (2) Calculate the information entropy of the feature vector corresponding to each randomly generated binary code, the proportion of sensors arranged in the binary code, and the fitness value; (3) Arrange all individuals from large to small according to their fitness values, and select the individual with the highest fitness value as the current best position; (4) Update the population position. After the update, normalize the position of each individual in the population, perform binary coding mapping on the normalized position, and randomly generate a set of n-dimensional [0,1] vectors. In the normalized individual position, if the value of a certain dimension of the normalized individual position is less than the corresponding value of the corresponding dimension of the vector, then the value of the dimension is set to not install the sensor, otherwise install the sensor; (5) Recalculate the updated fitness value for the individual position in the mapped population, compare the fitness value of the updated population with the current best fitness value, and if the fitness value in the updated generation is better than the current best fitness value, then use the individual position and corresponding fitness value of the updated best population as the current best fitness value and the individual position of the best population; (6) Determine whether the maximum number of iterations has been reached. After the iteration is completed, the position of the Sky Eagle population is the binary code of the optimal sensor network.

5. The sensor layout method for piston-connecting rod system fault diagnosis according to claim 4 is characterized in that: In step (2), the information entropy P(z e ) is the state z e The probability distribution of z j are the eigenvalues ​​corresponding to different dimensions of the eigenvector, j = 1, 2, … 8 × m, where m is the number of sensors in the current layout.

6. The sensor layout method for piston-connecting rod system fault diagnosis according to claim 5 is characterized in that: Binary coding arrangement sensor number ratio K(2 n -1)=m / n.

7. The sensor layout method for piston-connecting rod system fault diagnosis according to claim 6 is characterized in that: Fitness value fit(2 n -1)=2-F(2 n -1)-G(2 n -1)+K(2 n -1), F(2 n -1) is the fault diagnosis accuracy.

8. The sensor layout method for piston-connecting rod system fault diagnosis according to claim 2 is characterized in that: In step (4), when updating the population position, the maximum number of iterations is H. If the current number of iterations h≤2H / 3 and the random number rand in the range [0,1]≤0.5, the updated individual position is: If rand>0.5, the updated individual position is: X2(h+1)=X best (h)×Levy(n)+X R (h)+(p-o)×rand X best (h) represents the best solution to the hth iteration, X M (h) represents the average position of the population at the hth iteration, Levy (Dim) is the Levy flight distribution function, X R (h) is the random solution taken in the range [1, Num] at the hth iteration, Num is the population size, and p and o represent the search spiral shape.

9. The sensor layout method for piston-connecting rod system fault diagnosis according to claim 2, characterized in that: In step (4), when updating the population position, the maximum number of iterations is H. If the current number of iterations h>2H / 3, and the random number rand in the range [0,1]>0.5, the updated individual position is: X3(h+1)=(X best (h)-X M (h))×α-rand+((UB-LB)×rand+LB)×α; If rand≤0.5, the updated individual position is: X4(h+1)=QF×X best (h)-(G1×X(h)×rand)-G2×Levy(n)+rand×G1, α is the exploitation adjustment parameter of 0.1, QF represents the quality function used to balance the search strategy, G1 represents the various actions used by predators to track prey during the hunting process, and G2 represents the flight slope of following the prey.

10. A sensor layout method for piston-connecting rod system fault diagnosis according to claim 8 or 9, characterized in that: After updating, use Normalize the position of each individual in the population, X b (h+1) represents the individual position updated by the b-th search method in the h+1-th iteration of the Skyhawk algorithm, b=1,2,3,4, and the position of each individual is an n-dimensional vector of (0-1).

Images