Rigidity optimization method for coaxial counter-rotating multi-bearing rotor system
Through particle swarm optimization algorithm and finite element analysis software, the problems of vibration coupling and intermediary bearing stiffness changes in the counter-rotating double rotor system are solved, the model accuracy and dynamic balance effect are improved, and multi-angle optimization of the counter-rotating double rotor structure is achieved. Ensure the regularity of bearing stiffness changes with rotation speed.
Patent Information
- Application Number
- CN202410961090.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-17
- Publication Date
- 2024-11-12
AI Technical Summary
There are problems in the dual rotor system with internal and external rotor vibration coupling, intermediary bearing stiffness changes with speed and time, and insufficient utilization of historical dynamic balance data, resulting in low virtual dynamic balance accuracy, affecting the balance effect.
By writing finite element analysis software, a finite element model for a rotor system is established, and a particle swarm optimization algorithm is used to optimize the dynamic parameters based on the measured first-order critical speed and vibration data to achieve the variability of the intermediary bearing stiffness, and through history Unbalanced data inversion improves the accuracy of the virtual model.
The accuracy and dynamic balance effect of the double rotor system model are improved, making the simulation model closer to reality at different speeds, achieving multi-angle optimization of the double rotor structure of the double rotor structure, ensuring the regularity of the bearing stiffness with the rotation speed.
Smart Images

Figure CN118940616A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to a method for realizing stiffness optimization of a coaxial counter-rotating multi-bearing rotor system, and specifically belongs to the field of dual-rotor structure optimization algorithms. Background Art
[0002] Compared with the multi-stage fan structure rotating in the same direction, the coaxial counter-rotating dual rotor structure can achieve a higher pressure ratio of the second-stage fan, about 1.4 to 1.6, due to the relative rotation of the fan. And under the same size and pressure ratio, the fan speed of the counter-rotating structure is much lower than that of the general multi-stage fan in the same direction. Therefore, the coaxial counter-rotating dual rotor structure is very effective in solving the problem of excessive fan diameter. At the same time, the counter-rotating dual rotors can make the vortices perpendicular to the blade diameter generated by the blade tip offset each other, greatly reducing energy loss. In addition, the structure can also make the torque generated by the rotation of the two propellers offset each other, increasing stability while improving efficiency. The above advantages make coaxial counter-rotating aircraft engines favored by many countries. .
[0003] Although the coaxial counter-rotating twin rotor structure has many advantages, due to its more complex structure, there are greater vibration and noise problems during operation. Among them, unbalanced vibration is an important source of noise in the counter-rotating twin rotor structure. In addition to generating huge noise, unbalanced vibration may also cause engine performance to decrease and life to decrease. In severe cases, it may even cause blade collision, damage or even falling off, bearing damage and other faults. The virtual dynamic balancing method is based on the finite element method. It adds virtual excitation to the finite element model to analyze the rotor characteristics and replace the actual test weight process. If the model is accurate, dynamic balancing without test weight can be achieved. This method combines rotor dynamics theory and finite elements, converts the actual operation of the rotor to the model, and replaces the actual weighting process in the balance by applying virtual quantities to the model for research and calculation, which greatly saves the startup cost and improves the efficiency of dynamic balancing. However, virtual dynamic balancing often has the problems of low model accuracy and insufficient parameter accuracy, which seriously affects the balancing effect of virtual dynamic balancing. At the same time, for the twin rotor structure, the high-speed and light-loaded aircraft engine rotor system uses cylindrical roller bearings with inner and outer rings rotating in the same direction as the intermediate bearings between the high-pressure rotor and the low-pressure rotor. Usually, the inner ring is connected to the low-pressure rotor, and the outer ring is connected to the high-pressure rotor. The speed of the inner ring is lower than that of the outer ring, forming a dual-rotor-rolling bearing coupling system with relatively complex dynamic characteristics. Therefore, it is very important to improve the accuracy of the virtual model of the coaxial counter-rotating dual rotors.
[0004] In 2015, Patent No. CN104535262A disclosed a virtual dynamic balancing method for the whole machine of turbine machinery N+1 support shaft system without test weight, which belongs to the field of turbine machinery dynamics and dynamic balancing technology. This invention aims at the problems of strong coupling of rotor vibration characteristics of N+1 support shaft system, greater mutual correlation of rotor vibration, and strong and complex shaft system vibration. It proposes a virtual dynamic balancing method that can reduce the number of unit starts and stops, shorten dynamic balancing time, improve dynamic balancing efficiency, and reduce dynamic balancing costs. Compared with this patent, the difference is that this patent can be applied to the dynamic balancing of counter-rotating twin rotors, and takes into account the changes of bearing stiffness with time and speed, improves the accuracy of the finite element model, and thus improves the accuracy of the twin rotor model.
[0005] In 2020, patent number CN111639386A disclosed a keyless phase virtual dynamic balancing method for aircraft engines, which belongs to the field of active vibration control and artificial intelligence technology. This invention is aimed at some rotating machinery that cannot measure the speed signal but urgently needs dynamic balancing. It can establish a finite element model similar to it, and perform steady-state analysis on it after applying a virtual imbalance. According to the measured rotor vibration amplitude, the nonlinear particle swarm optimization algorithm is used to optimize the position of the imbalance using the written finite element rotor dynamics program. This method does not require phase information, and can accurately obtain the size of the counterweight through a single original vibration measurement at any speed. Then, through a trial weight test at any phase or weight, the two possible positions of the counterweight can be calculated. Compared with this patent, the difference of this patent is that this patent realizes a dynamic analysis method that can be quickly called through a self-written finite element program. While ensuring accuracy, the efficiency of dual-rotor model optimization is improved.
[0006] In 2022, patent number CN115640717A disclosed a method for optimizing the modal tolerance of a dual-rotor system with an intermediate bearing, which belongs to the field of aeroengines. This invention is to overcome the problems of the "critical speed margin" criterion in the prior art that the applicability of the "critical speed margin" criterion is getting smaller and smaller, the modal tolerance optimization design method is imperfect, and the modal tolerance optimization design method for the dual-rotor system is not targeted. Without considering the critical speed and its margin range, all modes of high-pressure excitation and low-pressure excitation of the dual-rotor system with an intermediate bearing in the full speed range are optimized, and all modes are evaluated using the modal tolerance evaluation function constructed by the invention. The maximum value of the optimization objective function constructed by the invention is pursued by the genetic algorithm, and the final optimization result can be output only after the modal tolerance evaluation function of each order is greater than 0.8. This makes each order mode meet the requirements of tolerance resonance after the modal tolerance optimization design of the dual-rotor system with an intermediate bearing, and can operate at the critical speed, so that the dual-rotor system can operate at any speed without following the "critical speed margin" criterion, which improves the high maneuverability of the aeroengine and provides reference and ideas for the design of the next generation of aeroengines. Compared with this patent, the difference is that: this patent uses a particle swarm algorithm to optimize the stiffness of the counter-rotating dual-rotor system under different speed combinations based on the measured dual-rotor critical speed and vibration response, thereby achieving variable stiffness of the intermediate bearing under different speeds and improving the model accuracy of the finite element model under different speed combinations.
[0007] In 2023, patent number CN116539224A disclosed a method for dynamic balancing of a dual-rotor system, systematically established a modal dynamic balancing method for the N1+N2 plane of an aircraft engine dual-rotor system, and performed orthogonal correction mass group calculations in the mode when high pressure is the main excitation and the mode when low pressure rotor is the main excitation. The constant speed ratio condition of balance is controlled by a dual parameter input method during the actual operation of the engine dual-rotor system. When performing dynamic balancing, the mode of high pressure excitation and the mode of low pressure excitation are balanced separately; when performing modal dynamic balancing of the aircraft engine dual-rotor system, the influence coefficient method is combined with the modal dynamic balancing method, and the trial regroup and balance group of each balance plane of the dual-rotor system are calculated, which combines the advantages of modal dynamic balancing not affecting modal vibrations of other orders and the simple algorithm of the influence coefficient method, making the dynamic balancing effect more obvious. This invention has a control and suppression effect on the excessive vibration of a dual-rotor aircraft engine, and provides a reliable and effective guarantee for the safe operation of the aircraft engine. Compared with this patent, the difference is that: this patent combines the finite element method and optimization algorithm, adds optimization correction to the model parameters on the basis of virtual dynamic balancing, takes into account the vibration coupling transmitted by the inner and outer rotors through the intermediate bearing, and improves the model optimization effect of the counter-rotating dual rotor system.
[0008] In 2023, patent number CN117540172A disclosed a rotor fault diagnosis method based on a time-series self-supervised domain adaptive network. Self-supervised auxiliary data is constructed through the original acceleration data of the rotor under multiple working conditions, and the division of the original data set and the self-supervised auxiliary data set is realized at the same time; model training and model parameter optimization are carried out by constructing auxiliary learning tasks and self-supervised auxiliary data. In the process of time-series self-supervised learning, the unsupervised domain adaptive method is adopted to cooperate with time-series self-supervised learning to realize time-series self-supervised domain adaptive learning; in the downstream fault identification task, a fault identifier is constructed and the model parameters are fine-tuned using multi-domain original training data, and then the final model is obtained and the rotor fault diagnosis under the target working condition is realized. Compared with this patent, the difference is that: this patent is applied to counter-rotating dual rotors, taking into account the complex transmission of the inner and outer rotor vibrations caused by the existence of the intermediate bearing of the dual rotor, and targeted variable stiffness optimization is carried out for the stiffness changes of the intermediate bearing at different speeds.
[0009] In 2024, patent number CN117647936A disclosed a multi-objective optimal control method for an aircraft engine of a linear variable parameter system. The invention sets the objective function and constraints according to the operation process and results of the aircraft engine, and establishes a multi-objective optimal control model; through the control parameterization method, the infinite-dimensional optimal control problem is equivalently converted into a finite-dimensional optimal parameter selection problem; the converted multi-objective optimization problem is solved by using an improved multi-objective evolutionary algorithm to obtain a Pareto solution set. It can avoid falling into a local optimal solution during the population evolution process, thereby obtaining a more ideal numerical result and providing a reference for the final decision; the solution set obtained based on the numerical experiment can be regarded as the oil input law of the aircraft engine. Good oil input can effectively improve the combustion efficiency of the aircraft engine, thereby reducing fuel loss and increasing thrust output, thereby improving the performance and efficiency of the aircraft engine. Compared with this patent, the difference is that this patent can be applied to a coaxial counter-rotating dual-rotor structure, and a particle swarm algorithm is used to simultaneously optimize multiple optimization objectives to obtain a more accurate finite element model. Summary of the invention
[0010] The purpose of the present invention is to overcome the problems of the counter-rotating dual-rotor system, such as the coupling of the vibration of the inner and outer rotors, the change of the intermediate bearing stiffness with the speed and time and the anisotropy, and the insufficient use of historical dynamic balancing data. The finite element model corresponding to the counter-rotating dual-rotor system is established by using the self-written and callable finite element analysis software. Through the particle swarm optimization algorithm, the dynamic parameters of the counter-rotating dual-rotor finite element model are optimized multiple times according to the measured first-order critical speed and the measured vibration data of some nodes, so as to realize the variable stiffness of the intermediate bearing at different speeds. By performing data inversion on the historical unbalanced data, the dynamic stiffness characteristics of the counter-rotating dual-rotor system are obtained, and compared with the results after the optimization of the finite element model, so as to realize the full use of the historical dynamic balancing data and improve the accuracy of the virtual model.
[0011] To achieve the above purpose, the technical solution adopted by the present invention is as follows:
[0012] A stiffness optimization algorithm for a coaxial counter-rotating multi-support rotor system comprises the following steps:
[0013] Step 1: Establish a finite element model of the counter-rotating twin rotors. The actual counter-rotating twin rotors are discretized using the finite element method to become a discretized model with limited degrees of freedom. The dynamics of the counter-rotating twin rotors are simulated by the written finite element program, and the steady-state unbalanced response analysis of the inner and outer rotors is performed on the single rotors.
[0014] Step 2: Perform modal analysis on the discretized contra-rotating twin-rotor model. Perform modal analysis on the contra-rotating twin-rotor finite element model to obtain the Campbell diagram corresponding to the speed ratio, thereby obtaining the first several critical speeds and corresponding vibration modes of the model.
[0015] Step 3: After completing the modal analysis, measure the critical speed of the actual counter-rotating twin rotors. Measure the Bode plot of the speed change with the specified speed ratio on the actual counter-rotating twin rotors and extract the critical speed. The critical speed is the speed corresponding to the peak of the amplitude and the sharp change of the phase in the Bode plot.
[0016] Step 4: Measure the vibration response of the inner and outer shafts of the counter-rotating dual rotors under a given test weight. Measure the response of three measuring points of the inner and outer rotors at the same speed or measure the response of one measuring point of the inner and outer rotors at three different speeds.
[0017] Step 5: After completing the measurement of all required data of the actual counter-rotating dual rotors, based on the simulation model, the measured critical speed and vibration response are substituted into the particle swarm algorithm for optimization and multiple optimization is performed. Steps 5.1 to 5.7 below are the specific operations of multiple optimization.
[0018] Step 5.1: Define the particle swarm fitness function Among them, Function is the fitness, x 1 ~xn are multiple optimization parameters. For this method, x 1 ~x n They are respectively taken as: the absolute value of the ratio of the difference between the critical speed calculation result and the measurement result to the measurement result, and the absolute value of the ratio of the difference between the calculation result and the measurement result of the response of each measuring point at the selected speed to the measurement result.
[0019] Step 5.2, start particle swarm optimization, first initialize parameters, set parameters such as number of iterations, learning factor and population size, set the maximum and minimum values of individual speed and search range. The search range is related to the parameters represented by the individual value of the particle. For this patent, the individual value of the particle represents the bearing stiffness, so the search range is generally taken as the empirical value of the bearing stiffness ±10 2 Order of magnitude, unit is N / m.
[0020] Step 5.3, initialize the population and calculate the fitness value. Randomly generate a population containing individuals of the population size in the search space. The speed and position of each individual are six-dimensional vectors, which are randomly assigned. Call the fitness function to calculate the fitness value of each individual. For this patent, the six-dimensional vectors of the particle individuals represent the stiffness values of the supporting bearing and the two intermediate bearings in the x-axis and y-axis directions respectively.
[0021] Step 5.4: Calculate individual and group extreme values. Compare the fitness values of each individual to obtain the group extreme value, and record the best position of the individual as the individual extreme value.
[0022] Step 5.5: Update the speed and position of the particles. Each particle generates the speed and position of the next generation of particles based on the individual extreme value and the group extreme value.
[0023] Step 5.6: Repeat steps 5.4 and 5.5 until the fitness value reaches the required accuracy. At this point, the particle reaches the optimal position, that is, the bearing stiffness closest to the actual counter-rotating twin rotors is obtained.
[0024] Step 5.7, update the counter-rotating twin-rotor model with the optimized value. The critical speed and vibration response of the counter-rotating twin-rotor model at this time are closer to the actual counter-rotating twin-rotor model, thereby improving the model accuracy.
[0025] Compared with the existing methods, the present invention has the following beneficial effects:
[0026] 1. It can be applied to the counter-rotating dual-rotor structure, taking into account the reverse gyroscopic torque generated by the counter-rotation of the dual rotors and the vibration coupling of the inner and outer rotors transmitted through the intermediate bearing.
[0027] 2. Multiple parameters are involved at the same time to perform multiple optimizations on the counter-rotating twin-rotor structure. At the same time, the critical speed and vibration response characteristics of the counter-rotating twin-rotor structure are taken into consideration to make the stiffness optimization results more accurate, so that the simulation model is closer to the actual counter-rotating twin-rotor structure in multiple angles.
[0028] 3. Compared with the fixed bearing stiffness calculated according to the empirical formula or obtained by looking up the table, this method can obtain the regular curve of the bearing stiffness changing with the speed by optimizing the bearing stiffness separately under different speed combinations, thereby realizing the variable bearing stiffness at different speeds, which is more in line with the real counter-rotating dual rotor structure. BRIEF DESCRIPTION OF THE DRAWINGS
[0029] Figure 1 This is a schematic diagram of the stiffness optimization of the counter-rotating twin rotors. The finite element model of the counter-rotating twin rotors is obtained by discretizing the actual counter-rotating twin rotors, and the critical speed and vibration response of the measured counter-rotating twin rotors are used as the optimization basis. The optimized counter-rotating twin rotor support stiffness including the variable intermediate bearing stiffness can be obtained through the particle swarm multi-optimization algorithm combined with the finite element program that can be quickly called, thereby realizing the correction of the counter-rotating twin rotor finite element model and improving the model accuracy.
[0030] Figure 2 It is a flow chart of particle swarm optimization. The steps of particle swarm optimization include: initializing particle speed and position, constructing fitness function, calculating fitness size, calculating individual optimum and group optimum, updating particle speed and position, and judging whether the end condition is met. The characteristics of this method are: modifying the finite element model by optimizing the bearing stiffness, improving the model accuracy through collaborative optimization of multiple optimization bases, and making the bearing stiffness variable at different speeds;
[0031] Figure 3 This is a flowchart that takes into account the nonlinearity of the intermediate bearing. The process of obtaining the nonlinear stiffness of the intermediate bearing includes: randomly setting the initial stiffness array of the intermediate bearing, calculating the fitness of the stiffness array through the fitness function, recording the individual optimum and the group optimum, updating the particle speed and position, judging whether the error is acceptable, outputting the optimal solution, and forming the nonlinear stiffness curve of the intermediate bearing based on the optimal stiffness array;
[0032] Figure 4 This is a flow chart for the inversion of dynamic balancing data. By performing finite element modeling on the actual counter-rotating twin rotors, a finite element model of the counter-rotating twin rotors is obtained. By analyzing the actual counter-rotating twin rotor structure, the vibration transmission law between the counterweight plate, the bearing seat measuring point, and the bearing is obtained. By combining this law with the historical dynamic balancing data and the counter-rotating twin rotor finite element model, the required dynamic stiffness parameters can be obtained. The dynamic stiffness parameters obtained by combining the measured initial amplitude and reversely calculating through the vibration transmission law can obtain the original imbalance size. DETAILED DESCRIPTION
[0033] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below using the optimization process of the counter-rotating dual-rotor test bench and in combination with relevant drawings. The specific embodiments described here are only used to understand the present invention and are not used to limit the present invention.
[0034] Example 1: Optimization of the intermediate bearing stiffness of a counter-rotating dual-rotor system
[0035] Step 1: Establish a dual-rotor simulation model. The finite element method is used to transform the actual counter-rotating dual-rotor test bench model into a discretized model with limited degrees of freedom. The dynamic simulation is performed through the written finite element program, and its steady-state unbalanced response is analyzed. The steady-state unbalanced response here is the linear superposition after the inner and outer shafts are respectively applied with unbalanced amounts.
[0036] Step 2: Perform modal analysis on the discretized contra-rotating twin-rotor model. Perform modal analysis on the contra-rotating twin-rotor finite element model to obtain the Campbell diagram corresponding to the speed ratio, thereby obtaining the first several critical speeds and corresponding vibration modes of the model.
[0037] Step 3: After completing the modal analysis, measure the critical speed of the actual counter-rotating twin rotors. Measure the Bode plot of the speed change with the specified speed ratio on the actual counter-rotating twin rotor test bench and extract the critical speed. The critical speed is the speed corresponding to the peak of the amplitude and the sharp change of the phase in the Bode plot.
[0038] Step 4: Measure the vibration response of the inner and outer shafts of the counter-rotating dual-rotor test bench under a given test weight. Measure the response of three measuring points of the inner and outer rotors at the same speed, or measure the response of one measuring point of the inner and outer rotors at three different speeds. Here, select the response of one measuring point of the inner and outer rotors at three different speeds.
[0039] Step 5: After completing the measurement of all required data of the actual counter-rotating dual rotors, based on the simulation model, the measured critical speed and vibration response are substituted into the particle swarm algorithm for optimization and multiple optimization is performed. Steps 5.1 to 5.7 below are the specific operations of multiple optimization.
[0040] Step 5.1: Define the particle swarm fitness function Among them, Function is the fitness, x 1 ~x n are multiple optimization parameters. For this method, x 1 ~x n They are respectively taken as: the absolute value of the ratio of the difference between the critical speed calculation result and the measured result to the measured result, and the absolute value of the ratio of the difference between the calculated result and the measured result of the response of each measuring point at the selected speed to the measured result. Here, n is taken as 7.
[0041] Step 5.2, start particle swarm optimization, first initialize parameters, set parameters such as number of iterations, learning factor and population size, set the maximum and minimum values of individual speed and search range. Here, the number of iterations is set to 50, the individual learning factor is set to 2, the group learning factor is set to 2, the population size is set to 20, and the individual speed boundary is set to -10 2 ~10 2 , the search range is set to 10 6 ~10 9 , unit: N / m.
[0042] Step 5.3, initialize the population and calculate the fitness value. Randomly generate a population containing individuals of the population size in the search space. The speed and position of each individual are six-dimensional vectors, which are randomly assigned. Call the fitness function to calculate the fitness value of each individual. Here, the six-dimensional vectors of the particle individual represent the stiffness values of the supporting bearing and the two intermediate bearings in the x-axis and y-axis directions respectively.
[0043] Step 5.4: Calculate individual and group extreme values. Compare the fitness values of each individual to obtain the group extreme value, and record the best position of the individual as the individual extreme value.
[0044] Step 5.5, update the speed and position of the particle. Each particle generates the speed and position of the next generation of particles based on the individual extreme value and the group extreme value. The formula for updating the speed and position of the particle is as follows:
[0045] v=w*v+c 1 *rand1*(i best -p)+c 2 *rand2*(g best -p)
[0046] Where v is the particle velocity, c 1 、c 2 are the individual learning factor and the group learning factor, respectively. rand1 and rand2 are random numbers between 0 and 1. best , g best are the best position of the individual and the best position of the group, respectively, and p is the current position of the population. w is the linearly decreasing inertia weight, and the calculation formula is:
[0047]
[0048] Among them, G k is the total number of iterations, i is the current number of iterations, and w max 、w min are the initial inertia weight and the final inertia weight, which are taken as 0.9 and 0.4 respectively here.
[0049] Step 5.6: Repeat steps 5.4 and 5.5 until the fitness value reaches the required accuracy. At this point, the particle reaches the optimal position, that is, the bearing stiffness closest to the actual counter-rotating twin rotors is obtained.
[0050] Step 5.7, update the rotor model with the optimized value. At this time, the critical speed and vibration response of the counter-rotating dual-rotor model are closer to the actual counter-rotating dual-rotor model, thereby improving the model accuracy.
Claims
1. A stiffness optimization method for a coaxial counter-rotating multi-support rotor system, characterized in that: The following steps are involved: Step 1, establish a finite element model of the counter-rotating double rotors; use the finite element method to discretize the actual counter-rotating double rotors into a discretized model with limited freedom; perform dynamic simulation on the counter-rotating double rotors through the written finite element program, and perform steady-state unbalanced response analysis of the single rotor on the inner and outer rotors respectively; Step 2, performing modal analysis on the discretized counter-rotating twin-rotor model; performing modal analysis on the counter-rotating twin-rotor finite element model to obtain a Campbell diagram corresponding to the speed ratio, thereby obtaining the first several critical speeds and corresponding vibration modes of the model; Step 3: After completing the modal analysis, measure the critical speed of the actual counter-rotating twin rotors; measure the Bode diagram of the speed change with the specified speed ratio on the actual counter-rotating twin rotors, and extract the critical speed; the critical speed is the speed corresponding to the peak value of the amplitude and the sharp change of the phase in the Bode diagram; Step 4, respectively measure the vibration response of the inner shaft and the outer shaft of the counter-rotating dual rotors under a given test weight; measure the response of three measuring points of the inner and outer rotors at the same speed or measure the response of one measuring point of the inner and outer rotors at three different speeds; Step 5: After completing the measurement of all required data of the actual counter-rotating dual rotors, based on the simulation model, the measured critical speed and vibration response are substituted into the particle swarm algorithm for optimization and multiple optimizations are performed.
2. The method for optimizing the stiffness of a coaxial counter-rotating multi-bearing rotor system according to claim 1, characterized in that: The specific implementation process of step 5 is as follows: Step 5.1, define the particle swarm fitness function Among them, Function is the fitness, x1~x n are multiple optimization parameters; for this method, x1~x n They are respectively taken as: the absolute value of the ratio of the difference between the critical speed calculation result and the measurement result to the measurement result, and the absolute value of the ratio of the difference between the calculation result and the measurement result of the response of each measuring point at the selected speed to the measurement result; Step 5.2, start particle swarm optimization, first initialize the parameters, set the number of iterations, learning factor and population size parameters, set the maximum and minimum values of individual speed and search range; the individual value of the particle represents the bearing stiffness, and the search range is taken as the empirical value of the bearing stiffness ± 10 2 Order of magnitude; Step 5.3, initialize the population and calculate the fitness value; randomly generate a population containing individuals of the population size in the search space, and the speed and position of each individual are six-dimensional vectors, which are randomly assigned; call the fitness function to calculate the fitness value of each individual; the six-dimensional vectors of the particle individual represent the stiffness values of the supporting bearing and the two intermediate bearings in the x-axis and y-axis directions respectively; Step 5.4, calculate individual and group extreme values; compare the fitness values of each individual, obtain the group extreme value, and record the best position of the individual as the individual extreme value; Step 5.5, update the speed and position of the particles; each particle generates the speed and position of the next generation of particles according to the individual extreme value and the group extreme value; Step 5.6, repeat steps 5.4 and 5.5 until the fitness value reaches the required accuracy; at this time, the particle reaches the optimal position, that is, the bearing stiffness closest to the actual counter-rotating dual rotor is obtained; Step 5.7, update the counter-rotating twin-rotor model with the optimized value. The critical speed and vibration response of the counter-rotating twin-rotor model at this time are close to those of the actual counter-rotating twin-rotor model.
