Please click "Follow" before reading the article, so that it is easy to discuss and share, and in order to return your support, I will update the quality content daily.
preface
With the continuous improvement of environmental awareness and relevant laws and regulations, whether it is the automotive industry or the construction machinery industry, more and more customers are beginning to pay more attention to extreme vibration and noise hazards.
At present, the competition in the construction machinery market is becoming increasingly fierce, and the performance cost of similar models in the conventional aspect is getting closer and closer, and major manufacturers choose to compete and technical direction to focus on improving the vibration comfort level of the whole vehicle in order to increase their product competitiveness.
The driver can usually directly feel the vibration performance of the cab, and during actual operation, the vibration level of the cab is too high to cause the driver to have various adverse reactions, such as lack of energy and slow response.
Therefore, cab design should not only consider safety and economy, but also consider driver comfort, so what should we do to reduce cab vibration and improve driver comfort?
Ride comfort
As the core system of cab vibration isolation, the cab bracket and its influence on the vibration comfort performance of the whole vehicle are of great significance.
In order to improve the driver's riding comfort, in addition to considering the main vibration sources such as working devices, engines, and tracks, it is necessary to reduce vibration from the source, so isolating the vibration source of cab vibration has become our main improvement direction.
We knew from the beginning that by optimizing the design of the cab bracket, we could reduce the amplitude of vibration transmitted from the frame to the cab.
Therefore, the performance of the cab bracket will directly affect the vibration level of the cab, which plays a vital role in improving the vibration comfort of the cab.
The cab parameter information acquisition and matching analysis of the installation system and the optimization process require in-depth theoretical research in combination with experimental design and finite element analysis.
However, in this process, the traditional scaffold system design and optimization process still faces many obstacles.
On the one hand, there are technical difficulties, because this means that in the design of the cab bracket system for construction machinery and the actual project we want to optimize, there will be many conflicting requirements and conditions, and it is very difficult to meet at the same time.
On the other hand, because of the many parameters of the cab installation system, small parameter changes will lead to major changes in the performance of the installation system, which also brings great technical difficulties to improving the vibration isolation performance of the bracket.
In the optimization of the cab bracket system, it is also necessary to carry out decoupling optimization, vibration isolation optimization and robustness optimization based on the vehicle coordinate system.
However, these need to be adjusted in repeated iterations with the help of cab simulation models and mathematical models, and the purpose of rapid optimization cannot be achieved at all.
Especially because we have to meet the accuracy requirements of multiple mounts while also completing the task of optimal matching of stiffness in different directions, which leads to long installation system development cycles and low efficiency.
Mount system design
There are many ways to predict the stiffness of cab brackets in construction machinery, including traditional finite element calculations based on mechanisms and data mining methods, and data-driven methods.
However, these data-driven methods also have their drawbacks, such as their disadvantages when faced with rubber.
As a nonlinear material, rubber's characteristics lead to a great difference from metal in mechanical use, because rubber brackets do not have a suitable stiffness calculation formula.
Regardless of this situation, with the development of technology has also changed somewhat, with the maturity of rubber material hyperelasticity theory and the development of computer technology, we can also apply some finite element software to calculate and study rubber supports.
In the 40s of the 19th century, two scholars formed a damping mechanism composed of rubber and hydraulic damping devices, and successfully applied for a patent.
After that, they made an axisymmetric finite element model of the spring, calculated the force and deformation curve of the rubber main spring, and then used ANSYS to calculate the volume flexibility problem of the axisymmetric rubber mainspring, and gave the setting method of boundary conditions and the final results of the simulation.
In order to analyze the influence of different vibration modes on ride comfort, some scholars chose to establish a vehicle model for simulation, and significantly improve ride comfort by changing the position of the bracket relative to the body.
On the basis of this research, the influence of stiffness and damping of the cab bracket system on front longitudinal and lateral acceleration, vertical seat displacement, vertical seat speed and vertical seat acceleration was analyzed through a large number of experiments, and then the parameters of the cab bracket system were designed according to the analysis results, which greatly reduced the displacement and acceleration of the cab.
However, if the finite element method is to be used, extensive theoretical studies of the stiffness of the suspension rubber and cab system need to be analyzed in conjunction with experimental design, because this method makes the modeling process time-consuming and inefficient, which is not conducive to rapid optimization.
Compared with the traditional finite element method, the data mining method for predicting the stiffness of the mounting can eliminate the complex investigation and complex modeling process, and significantly improve the prediction and optimization efficiency.
At the same time, in order to identify the positive and negative mechanical models of magnetorheological hydraulic supports, the research team of Chongqing University compared the model recognition accuracy of BP neural network and GA-BP neural network algorithms by using the dynamic characteristics of magnetorheological supports as training data.
The final results show that the positive and negative models of the magnetorheological hydraulic support using the GA-BP neural network converge faster and have higher recognition accuracy, and this method helps to control the magnetorheological bracket, and the rheological mount is used for the control system.
After that, the team of Tongji University used loudness, roughness, sharpness, jitter and speech intelligibility as input parameters and subjective evaluation scores as output parameters, and established a dynamic noise quality prediction model for electric vehicles using brackets.
The vector machine algorithm and particle swarm algorithm are used to optimize the model, and the final results show that the prediction model has high accuracy and the average error is only 2%.
Later, some people chose to use the advanced hybrid neural network (AHNN) friction component model, use the data obtained from the powertrain dynamometer test bench to train the model, and finally predict the dynamic parameters of the gearbox when changing speeds.
Improved decision tree models have also been used to simulate driver behavior in racing simulations to balance the accuracy of machine and logic work with the accuracy of human work.
Another part of the scholars used genetic algorithms and random forest classifiers for gearbox failure detection, and the classification accuracy rate exceeded 97%.
Later, in order to view the increase in the demand for electric vehicle charging, a domestic scholar proposed a method for predicting the charging load of electric vehicles based on random forest algorithm and load data of individual charging stations.
This method determines the application form of the current data in the algorithm, and verifies the accuracy and reliability of the prediction algorithm, which indicates that it is feasible to use the data mining method to predict the installation stiffness, but the requirements for evaluating the predicted installation stiffness are higher.
The evaluation methods for cab brackets can be divided into the following two categories: one is the mounting system level, where the stiffness of the mount is constrained by the decoupling rate as a boundary condition, and the decoupling rate in the main vibration direction can be maximized by setting the stiffness of the appropriate installation.
The second is the whole vehicle level, which mainly depends on the vibration isolation performance analysis of the system, in which the vibration transmission enters the cab system through the frame, and the stiffness of the bracket is again constrained by the vibration isolation performance as a boundary condition.
Feasibility of stiffness prediction
These studies prove the feasibility of data mining technology for installation stiffness prediction, and use data mining methods to establish the mapping relationship between powertrain parameters and mount stiffness for nonlinear continuous numerical variables such as mount system stiffness.
Moreover, engineering evaluation indicators such as decoupling rate and vibration isolation rate of cab mounting system are introduced as boundary conditions of multiple regression prediction model, which avoids the traditional complex modeling process, and realizes the direct prediction of mounting stiffness and rapid optimization of the mounting system.
After these studies, some scholars have carried out some experiments on the prediction of cab bracket stiffness and achieved effective results, but there are still some common problems in the prediction and evaluation of construction machinery cab installation stiffness.
For example, in the previous cab bracket stiffness prediction research, most of the traditional finite element methods were used, which led to increased modeling complexity and reduced prediction optimization efficiency.
Coupled with the use of data-driven data mining methods by some academics, in which no multi-output model is introduced, and due to the correlation between the isotropic stiffness of each mount, the single-output prediction result ultimately leads to a low confidence in the predicted stiffness of the mount.
Construction machinery cab bracket system
Combining these valuable experiences, we chose to bypass the traditional complex modeling process from the outset when investigating the cab mounting system for construction machinery.
It is proposed to introduce multi-objective regression forest (MRF) in the design and optimization design of the installation system, and combine the installation parameters with the MRFs method to realize the multi-point installation stiffness prediction, and then take the engineering evaluation indicators such as system decoupling rate and vibration isolation rate as the boundary conditions.
After comparing the other two multi-output regression methods, we find a satisfactory multi-objective prediction model method and establish a set of high-confidence installation system stiffness prediction model.
In the case of avoiding the complex modeling process, we establish the mapping relationship between cab parameters and installation stiffness, which solves the problems of technical difficulty, low data mining rate and long development cycle in the development of bracket system, and applies it to the optimization of cab bracket system in engineering.
Multi-objective regression forest model
From the current research status of the design of the cab mount of construction machinery, it can be seen that the optimal design of the mounting part mainly focuses on the optimization of the stiffness of the bracket, in the dynamic modeling, the mount is usually simplified to three linear springs perpendicular to each other, and the mount provides stiffness separately in each direction.
However, due to the correlation of isotropic stiffness between mounts, when mounting stiffness is used as the prediction target of model tuning, it will inevitably affect other models, so the correlation between targets needs to be considered.
The advantages of multi-objective regression are mainly that the multi-objective regression model is usually smaller than the model with multiple single targets, and the multi-objective model can better identify the dependence between different target variables, and the model prediction effect is better.
MRF is a data tree that can predict multiple consecutive targets at the same time, and internally it is a binary tree structure that divides the data into two subsets, each node has an optimization base.
The optimization basis for multiple objectives is to replace the sum of squared errors of a single variable with the sum of squared errors of multiple variables, repeat the division process until the stopping condition is met, and finally generate a decision tree with the median or average value of the leaf nodes as the prediction result.
In fact, the MRFs algorithm can better handle the interaction between the input features of the installation stiffness prediction model, and the results generated are highly interpretable.
The robustness of this algorithm is also very good when adding noisy data or feature changes during MRFs model construction, which can automatically detect interactions between variables and handle missing values in element variables with minimal information loss.
If the construction machinery cab installation system is taken as the research object, the left front mount, the right front suspension, the left rear suspension, and the right rear mount are arranged, the corresponding installation system is arranged flat, and the stiffness coordinate system of each elastic bracket is parallel to the center of mass coordinate system of the cab.
conclusion
According to the final results, after the optimization of the MRFs model, the decoupling rate in each degree of freedom direction is higher than that of the multi-SVR and MLPR models, and only the MRFs model significantly improves the vibration isolation rate compared with the original state after optimization, while the other models even decrease in some directions.
After the advantages of MRF in installing system stiffness prediction, we solve the shortcomings of traditional finite element design methods and traditional data mining methods such as high difficulty and low robustness, and provide a guiding direction for suspension system stiffness prediction.
If you like my article, please click "Follow", welcome to discuss with me in the comment area, thank you.
END