A press-fitting device for bearing machining and a press-fitting method thereof

Abstract

The application is suitable for the technical field of mechanical manufacturing, and provides a press-fitting device for bearing machining and a press-fitting method thereof, wherein the press-fitting method for bearing machining comprises the following steps: S1. constructing a press-fitting resistance expected risk model according to a theoretical interference amount of a bearing and a shell, surface roughness of a fitting surface and cleanliness of the fitting surface, and outputting a press-fitting resistance expected deviation; S2. constructing a press-fitting process health evaluation model according to a press-fitting degree, a press-fitting speed, bearing pressure uniformity and bearing and shell centering degree, and outputting a press-fitting process health index. The press-fitting method for bearing machining systematically solves problems such as inaccurate resistance prediction and missing process state evaluation in the traditional press-fitting process by constructing a closed-loop management system integrating risk expectation, process evaluation, risk early warning and dynamic control.

Description

Technical Field

[0001] This invention belongs to the field of mechanical manufacturing technology, and in particular relates to a press-fitting device and press-fitting method for bearing processing. Background Technology

[0002] As a core component of mechanical equipment, bearings directly determine the overall machine's operational accuracy, service life, and stability through their performance and reliability. In bearing assembly, the press-fitting process is crucial for embedding the bearing outer ring into the housing with an interference fit. This process significantly impacts quality attributes such as the interference fit, coaxiality, and residual stress distribution. However, traditional press-fitting processes typically employ preset constant pressure or constant speed control modes, failing to fully consider the complex interactions of multiple dynamic factors during the press-fitting process. This results in significantly insufficient process robustness and difficulty in ensuring consistent quality.

[0003] Specifically, predicting and controlling press-fit resistance faces significant challenges. Press-fit resistance is influenced by multiple factors, including theoretical interference fit and surface roughness of mating surfaces. Current technologies rely solely on operator experience to set pressure thresholds, lacking a quantitative mechanism for predicting resistance changes. This empirical approach cannot identify risks of abnormal resistance deviations, easily leading to sudden pressure increases, causing damage to bearing raceways or deformation of the housing structure, and in severe cases, resulting in assembly failure.

[0004] Meanwhile, real-time status assessment during the pressing process has significant shortcomings; traditional methods have failed to establish a multi-dimensional parameter fusion assessment model. The lack of real-time quantitative analysis of pressure uniformity and alignment results in abnormal conditions such as skewing and jamming not being identified and addressed in a timely manner, further exacerbating assembly quality fluctuations. Existing technologies lack a comprehensive risk assessment framework, making it impossible to provide quantitative early warnings before deformation occurs, thus hindering proactive risk mitigation. Therefore, existing technologies urgently need improvement to address these issues. Summary of the Invention

[0005] The purpose of this invention is to provide a pressing device and pressing method for bearing processing, in order to solve the above-mentioned problems.

[0006] This invention is implemented as follows: a press-fitting method for bearing processing includes the following steps: S1. Construct a press-fit resistance expected risk model based on the theoretical interference fit between the bearing and the housing, the surface roughness of the mating surface, and the cleanliness of the mating surface, and output the expected deviation of the press-fit resistance; S2. Construct a health assessment model for the pressing process based on pressing force, pressing speed, bearing pressure uniformity, and bearing-housing alignment, and output the pressing process health index. S3. Based on the temperature difference between the bearing and the housing, the expected deviation of the pressing resistance, the health index of the pressing process, and the bearing vibration intensity during the pressing process, a pressing deformation risk assessment model is constructed, and the pressing deformation risk index is output. S4. Construct a pressing speed control model based on the pressing deformation risk index, the pressing process health index, and the current pressing speed, and output the target pressing speed.

[0007] In a further technical solution, in step S1, the theoretical interference fit between the bearing and the housing, the surface roughness of the mating surface, and the cleanliness of the mating surface are sequentially substituted into the maximum-minimum value formula for normalization, and the theoretical interference fit index, surface roughness index, and mating surface cleanliness index are generated sequentially; in the press-fit resistance expected risk model: The theoretical interference index and surface roughness index are multiplied by their respective preset influence coefficients and summed. The result is then subtracted from the difference between the mating surface cleanliness index and its corresponding preset influence coefficient, and a preset bias constant is added. The final result is then input into the Sigmoid function for nonlinear mapping, and the expected deviation of the press-fit resistance is output. The sum of each preset influence coefficient is 1.

[0008] In a further technical solution, step S2 involves substituting the bearing pressure uniformity, bearing-housing alignment, pressing speed, and reference pressing force into a maximum-minimum normalization formula for normalization, and sequentially generating the pressure uniformity index, alignment index, pressing speed index, and reference pressing force index; in the pressing process health assessment model: First, calculate the reference pressing force corresponding to the current speed based on the linear relationship between the baseline pressing force index and the pressing speed index; Secondly, a weighted power product of the pressure uniformity index and the neutrality index is used as the basic assessment value for process health. Finally, by calculating the absolute deviation between the actual pressing force and the reference pressing force, and introducing a preset penalty coefficient to scale the absolute deviation, the scaling result is limited to the range of 0 to 1. Then, the penalty adjustment factor is obtained by subtracting the limit value from 1. The basic evaluation value is multiplied by the penalty adjustment factor to output the pressing process health index.

[0009] In a further technical solution, step S3 involves substituting the temperature difference between the bearing and the housing, as well as the bearing vibration intensity during the press-fitting process, into a maximum-minimum normalization formula for normalization. This normalization process generates the temperature difference index and the bearing vibration intensity index, which are then incorporated into the press-fitting deformation risk assessment model. Calculate "1 plus (the product of temperature difference index and its risk coefficient)", "1 plus (the product of expected deviation of press-fit resistance and its risk coefficient)" and "1 plus (the product of bearing vibration intensity index and its risk coefficient)" respectively, and multiply the three together to obtain the positive risk factor. Simultaneously, the inhibition factor is obtained by subtracting the inhibition factor from the health index inhibition coefficient multiplied by the power operation result of the health index of the pressing process from 1; then the inhibition factor is subtracted from the positive risk factor and divided by a normalized benchmark value to finally obtain the pressing deformation risk index; where each preset risk coefficient, inhibition coefficient and weight coefficient are positive values.

[0010] A further technical solution involves step S4, where the current pressing speed is substituted into the maximum-minimum normalization formula for normalization, and a pre-pressing speed index is generated; in the pressing speed control model: The target pressing speed index is based on the current pressing speed index. The total attenuation is calculated by multiplying the pressing deformation risk index by the preset first attenuation coefficient, and the unhealthy degree of the pressing process health index (calculated by subtracting the health index from 1) by the preset second attenuation coefficient. The result of subtracting the total attenuation from 1 is multiplied by the current pressing speed index to ensure that the adjustment result is not lower than the set minimum allowable speed index of the process. The attenuation coefficients mentioned above are all positive values.

[0011] Further technical solutions also include step S5, which specifically involves constructing an adaptive optimization model for lubricant dosage between batches based on the lubricant spraying amount of this batch of bearings, the pressure deformation risk index, the bearing vibration intensity during the pressure process, and the alignment between the bearing and the housing, and outputting the recommended spraying amount for the next batch, thereby controlling the lubricant spraying amount of the next batch of bearings.

[0012] A further technical solution involves substituting the amount of bearing lubricating oil sprayed into the maximum-minimum normalization formula for normalization, and generating an index for the amount of lubricating oil sprayed in this batch; in the batch-to-batch adaptive optimization model for lubricant dosage: First, subtract the pressure deformation risk index from 1 and the bearing vibration intensity index from 1, respectively, and multiply them by their corresponding preset influence weights. Then, add them to the product of the neutrality index and its influence weight to calculate the comprehensive performance score. Secondly, the difference between the comprehensive performance score and the preset target performance score is multiplied by the preset learning rate and the estimated value of the influence direction of the lubricant dosage to obtain the adjustment amount; Finally, the adjustment amount is added to the lubricant spraying amount index of this batch, and the result is limited to the allowable range to output the recommended spraying amount index for the next batch; wherein the sum of the preset influence weights is 1, the learning rate is between 0 and 1, and the estimated value of the influence direction of the lubricant dosage is a positive or negative constant.

[0013] A further technical solution includes a frame, in which an electric clamping assembly for fixing the housing and a press-fitting machine for pressing down the bearing are disposed, the press-fitting machine being disposed on the upper side of the electric clamping assembly, and further comprising: Memory, used to store executable instructions; The processor, when executing executable instructions stored in the memory, implements the above-described press-fitting method for bearing machining.

[0014] Compared with the prior art, the beneficial effects of the present invention are as follows: This solution comprehensively considers multiple influencing factors, such as the theoretical interference fit between the bearing and the housing, the surface roughness of the mating surfaces, and the cleanliness of the mating surfaces, and outputs the expected deviation of the press-fit resistance in a quantitative form. This enables the press-fit resistance expected risk assessment model to shift from traditional qualitative judgments based on experience to quantitative predictions based on data and models. This allows for early warning of potential resistance anomalies before or during the press-fitting process, effectively preventing bearing or housing damage caused by sudden increases in resistance, and significantly improving the robustness of the press-fitting process and the consistency of product quality.

[0015] The bearing press-fitting method proposed in this application systematically solves a series of technical challenges in traditional press-fitting processes, such as inaccurate resistance prediction, lack of process status assessment, insufficient early warning of deformation risks, and lag in parameter adjustment, by constructing a closed-loop management system that integrates risk prediction, process assessment, risk early warning, and dynamic control. This significantly improves the intelligence level and reliability of bearing press-fitting.

[0016] This application effectively addresses the problem of insufficient early warning of press-fit deformation risks in existing technologies. The solution standardizes the temperature difference between the bearing and the housing, as well as the bearing vibration intensity during the press-fit process, and integrates these with the expected deviation of press-fit resistance and the press-fit process health index to construct a comprehensive press-fit deformation risk assessment model. This enables the system to dynamically and quantitatively assess deformation risks during the press-fit process, providing accurate early warnings before deformation occurs. This comprehensive risk assessment mechanism considers not only thermodynamic and kinetic factors but also process health status, significantly improving the accuracy and timeliness of risk assessment. It provides a reliable basis for subsequent process parameter adjustments, effectively avoiding quality problems caused by bearing or housing deformation, and improving the consistency and reliability of product assembly quality.

[0017] This application effectively addresses the lack of standardized processing and adaptive calculation for lubricant dosage adjustment in traditional bearing press-fitting processes. Specifically, by normalizing the lubricant spraying amount of this batch of bearings, a batch lubricant spraying amount index is generated, eliminating the influence of data with different dimensions on model calculations and providing a unified and comparable data foundation for subsequent intelligent optimization. The batch-to-batch adaptive lubricant dosage optimization model comprehensively considers the press-fitting deformation risk index, bearing vibration intensity index, and neutrality index, and introduces adjustable weighting coefficients to achieve a comprehensive and quantitative evaluation of the current batch's press-fitting performance. This multi-dimensional, weighted fusion evaluation method avoids the one-sidedness of a single indicator, enabling the system to more accurately grasp the true state of press-fitting quality. Based on the difference between the comprehensive performance score and the target performance score, combined with the learning rate and the estimated value of the lubricant dosage influence, the model can intelligently calculate the recommended lubricant spraying amount for the next batch. Attached Figure Description

[0018] Figure 1 A schematic diagram illustrating the steps of a press-fitting method for bearing machining; Figure 2 This is a schematic diagram of a press-fitting device for bearing processing.

[0019] In the attached diagram: 1. Frame; 2. Electric clamp assembly; 3. Press machine. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0021] The specific implementation of the present invention will be described in detail below with reference to specific embodiments.

[0022] like Figure 1 As shown, a bearing pressing method for machining is provided in one embodiment of the present invention, which includes: S1. Based on the theoretical interference fit between the bearing and the housing, the surface roughness of the mating surfaces, and the cleanliness of the mating surfaces, construct a risk model for the expected press-fit resistance, and output the expected deviation of the press-fit resistance: S2. Construct a health assessment model for the pressing process based on pressing force, pressing speed, bearing pressure uniformity, and bearing-housing alignment, and output the pressing process health index. S3. Based on the temperature difference between the bearing and the housing, the expected deviation of the pressing resistance, the health index of the pressing process, and the bearing vibration intensity during the pressing process, a pressing deformation risk assessment model is constructed, and the pressing deformation risk index is output. S4. Construct a pressing speed control model based on the pressing deformation risk index, the pressing process health index, and the current pressing speed, and output the target pressing speed.

[0023] In this embodiment, the press-fit method for bearing processing refers to the process of pressing or assembling the bearing into the housing through an interference fit during bearing manufacturing or assembly. This method aims to ensure a stable and reliable connection between the bearing and the housing to meet the functional and performance requirements of the product.

[0024] The press-fit resistance expected risk assessment model is a functional unit that predicts and assesses the magnitude of potential resistance and its deviation from the normal range during the bearing press-fitting process based on preset input parameters. This model quantifies potential resistance anomalies by establishing a mathematical model or algorithm. The expected deviation of press-fit resistance is a quantitative index output by the press-fit resistance expected risk assessment model, used to represent the degree of difference between the actual press-fit resistance and the theoretical or expected resistance. This index reflects the risk level of press-fit resistance deviating from expectations.

[0025] The pressing process evaluation model is a functional unit that monitors and analyzes various dynamic parameters during the pressing process in real time, and comprehensively evaluates the overall status of the current pressing operation. This model aims to identify any abnormal or unhealthy conditions during the pressing process. The pressing process health index is a quantitative indicator output by the pressing process evaluation model, used to represent whether the current pressing process is operating within an ideal or healthy range. A higher health index generally indicates a smooth and controlled pressing process, while a lower index may indicate potential problems.

[0026] The press-fit deformation risk assessment model is a functional unit that comprehensively considers multiple influencing factors to assess the likelihood of plastic deformation or damage to bearings during press-fitting. This model aims to provide early warning of potential deformation risks to avoid product quality defects. The press-fit deformation risk index is a quantitative indicator output by the press-fit deformation risk assessment model, used to represent the degree of risk of bearing deformation during press-fitting. A higher index indicates a greater likelihood of deformation, requiring intervention measures.

[0027] The pressing speed control model is a functional unit that dynamically adjusts the operating speed of the pressing equipment based on real-time risk assessments and process status. This model aims to improve pressing quality and efficiency by optimizing the pressing speed to adapt to constantly changing process conditions. The target pressing speed refers to the ideal pressing speed output by the pressing speed control model, used to guide the pressing equipment's operation. This speed is dynamically calculated based on current process conditions and risk assessment results to achieve the best pressing effect.

[0028] This embodiment provides a press-fitting method for bearing processing, which achieves intelligent management of the press-fitting process through the collaborative work of multiple models.

[0029] In a preferred embodiment of the present invention, the theoretical interference fit between the bearing and the housing, the surface roughness of the mating surface, and the cleanliness of the mating surface are sequentially substituted into the maximum-minimum value formula for normalization, and the theoretical interference fit index, surface roughness index, and mating surface cleanliness index are generated sequentially; in the press-fit resistance expected risk model: ; in , Input value; The interference effect coefficient is... This is the roughness influence coefficient. This is the cleanliness impact coefficient. ; This is the theoretical over-excess index. It is the surface roughness index. To match the surface cleanliness index, This represents the expected deviation of the press-fit resistance.

[0030] In this embodiment, raw data on the theoretical interference fit between the bearing and the housing, the surface roughness of the mating surfaces, and the cleanliness of the mating surfaces are first collected. Specifically, the theoretical interference fit can be directly obtained from the design drawings or process specifications, showing the difference in the theoretical fit dimensions between the bearing and the housing. The surface roughness of the mating surfaces can be achieved by acquiring surface morphology images in real time on the production line using an online visual inspection system, and then extracting roughness features using image processing algorithms. The cleanliness of the mating surfaces can be assessed by scanning the mating surfaces with an optical cleanliness detection system (such as one based on CCD or laser scattering principles) to detect contaminants such as oil and dust. Alternatively, the surface cleanliness can be indirectly evaluated using infrared spectroscopy analysis or a contact angle meter. The data can be transmitted to the control system in real time.

[0031] The theoretical interference fit between the bearing and housing, the surface roughness of the mating surfaces, and the cleanliness of the mating surfaces are successively substituted into the maximum-minimum formula for normalization. This aims to convert data with different dimensions or units to the same dimension, or to map data to a specific interval (e.g., [0, 1]). The maximum-minimum normalization formula can use X... norm =(X -X min ) / (X max -X min ), where X is the original data, X min and X max These are the minimum and maximum values ​​in the dataset, respectively.

[0032] The press-fit resistance expected risk model is a mathematical function used to comprehensively consider multiple input parameters (theoretical interference ratio index, surface roughness index, and mating surface cleanliness index) to predict the degree of resistance deviation that may occur during the press-fitting process. This model provides forward-looking risk information for subsequent press-fitting process assessment and control by quantifying the expected deviation of press-fitting resistance. This application uses a logistic regression form based on the Sigmoid function to map the linear combination of inputs to the (0,1) interval, representing the probability or degree of deviation of the risk occurrence; alternatively, other machine learning models, such as support vector machines, decision trees, or neural networks, can be used to establish a mapping relationship between input parameters and press-fitting resistance deviation by training historical data.

[0033] in, It is a sigmoid activation function, often used to compress arbitrary real-valued inputs into the (0,1) interval, making its output interpretable as a probability or some kind of indicator. In this model, it converts the result of a linear combination into a value between 0 and 1, i.e., the expected deviation of the press-fit resistance. This makes the deviation interpretable; for example, a value closer to 1 indicates a higher risk of deviation. The function can be calculated directly using its mathematical expression, or by calling the Sigmoid function from the standard math library in the software implementation.

[0034] The interference effect coefficient is... This is the roughness influence coefficient. This is the cleanliness impact coefficient, and These coefficients are weights in the model used to measure the influence of each input index on the expected deviation of the pressing resistance. They reflect the importance of different factors influencing resistance during the actual pressing process. By adjusting these coefficients, the model can better fit the actual situation. The constraint that the weights sum to 1 ensures a reasonable allocation of weights. These coefficients can be obtained through machine learning training using historical pressing data, for example, by optimizing model parameters using gradient descent; or they can be initially set based on expert experience and process knowledge, and then fine-tuned using a small amount of data in practical applications.

[0035] This is the theoretical over-excess index. It is the surface roughness index. To match the surface cleanliness index, This represents the expected deviation of the press-fit resistance. These are the variables in the model, representing the normalized input characteristics and the model's output, respectively. They are the core components of the model's calculations and outputs, ensuring that the model can receive standardized inputs and produce quantified risk assessment results, which are then directly substituted into the model's formulas for calculation.

[0036] This application's solution quantifies and models the key factors affecting press-fitting resistance, achieving a precise assessment of the expected risk of press-fitting resistance. Specifically, firstly, the three original physical quantities—the theoretical interference fit between the bearing and the housing, the surface roughness of the mating surface, and the cleanliness of the mating surface—are normalized using a maximum-minimum formula. This process maps the original data with different dimensions and numerical ranges to a uniform range of 0 to 1, thereby generating the theoretical interference fit index, surface roughness index, and mating surface cleanliness index. This standardization eliminates the differences in dimensions among the factors, ensuring their fairness and comparability in subsequent model calculations. Based on this, these normalized indices are input into the expected risk model of press-fitting resistance. Cleanliness index The negative sign at the beginning reflects physical reality: the higher the cleanliness of the mating surfaces, the lower the risk of misalignment during press-fitting resistance. (Sigmoid function) Mapping the result of the linear combination to a range of 0 to 1 allows for a deviation of the expected output press-fit resistance. It can intuitively represent the degree of quantification of risk; the larger the value, the higher the risk of deviation from expected resistance.

[0037] In a preferred embodiment of the present invention, the bearing pressure uniformity, bearing-housing alignment, pressing speed, and reference pressing force are substituted into a maximum-minimum normalization formula for normalization, and pressure uniformity index, alignment index, pressing speed index, and reference pressing force index are generated sequentially; in the pressing process health assessment model: ; ; in The uniformity influence coefficient is... For the moderate impact coefficient, For speed influence coefficient, Dimensionless If the intensity deviates from the penalty coefficient, ,and , , as well as All are greater than 0; The pressure uniformity index. For the moderate index, For the pressing speed index, For real-time pressing force index, As the benchmark press-fit force index, For the health index of the pressing process, Indicates a specific pressing speed The ideal pressing force threshold.

[0038] In this embodiment, the original physical quantities such as bearing pressure uniformity, bearing and housing alignment, pressing speed, and reference pressing force are converted into dimensionless exponents using a maximum-minimum normalization formula.

[0039] This normalization process can acquire raw data in real time using sensors. For example, a pressure sensor array can be used to measure the uniformity of bearing pressure, a vision system or laser displacement sensor can be used to measure alignment, and an encoder or speed sensor can be used to measure pressing speed. This raw data is then input into a processor, which executes a maximum-minimum normalization algorithm to linearly scale the data to a preset range. Alternatively, theoretical maximum and minimum values ​​for each parameter can be preset. During the pressing process, the parameter values ​​acquired in real time are compared with the preset extreme values ​​and normalized accordingly.

[0040] The press-fitting process health assessment model is a comprehensive mathematical expression used to quantitatively evaluate the health status of the bearing press-fitting process. It generates a single health index by integrating the pressure uniformity index, alignment index, and the deviation between the actual press-fitting force and the reference press-fitting force. This index directly reflects whether the press-fitting process is smooth, whether the alignment is good, and whether the force is reasonable, thus enabling timely detection of potential anomalies. This model can be implemented in the controller of the press-fitting equipment or a standalone industrial computer. The processor receives the normalized indices in real time and performs mathematical calculations within the model according to preset coefficients. Calculation results... It can be displayed in real time on the operation interface, or used as input for subsequent control models.

[0041] The reference pressing force model is used to calculate the pressing force at a specific pressing speed. The ideal pressing force threshold that should theoretically or empirically be achieved. It reflects the coupling relationship between pressing force and pressing speed; that is, during the pressing process, the required reasonable pressing force changes accordingly with the change in pressing speed. This is achieved by comparing the actual pressing force... By comparing the two models, it's possible to assess whether the pressing force deviates from expectations, thus determining the health of the pressing process. This model is typically pre-programmed into the control system of the pressing equipment. Before or during pressing, the current pressing speed index is used to... and the preset benchmark pressing force index and speed influence coefficient Real-time calculation This calculation result was subsequently used in the health assessment model for the press-fitting process.

[0042] Uniformity Influence Coefficient , moderate impact coefficient and speed influence coefficient Used to adjust the weights of different factors in the model on the health status of the pressing process. and These represent the relative importance of pressure uniformity and centrality in health assessment, respectively, and their sum is 1, ensuring a reasonable allocation of weights. This indicates the degree of influence of the pressing speed on the reference pressing force. The force deviation penalty coefficient. Controlling the real-time pressing force index Deviation from the ideal pressing force threshold At that time, the health index of the pressing process The severity of the punishment These coefficients determine the system's sensitivity and tolerance to fluctuations in pressurization force. , , as well as The setting of these coefficients directly affects the model's sensitivity and accuracy, enabling it to better reflect actual process requirements. These coefficients can be determined through expert experience, process experiments, or optimization algorithms. For example, by conducting a series of controlled pressing experiments, the impact of different uniformity, centering, or speed variations on pressing quality can be observed, and then the coefficients can be adjusted to ensure that the health index output by the model matches the actual quality assessment results.

[0043] This application normalizes key parameters such as bearing pressure uniformity, bearing-housing alignment, press-fitting speed, and reference press-fitting force to generate corresponding indices, thereby eliminating dimensional differences between different physical quantities and ensuring data consistency and comparability within the model. In the press-fitting process health assessment model, the pressure uniformity index is first used... And the moderate index Weighted product term To evaluate the geometric alignment and stress distribution during the press-fitting process, the uniformity influence coefficient is used. And the coefficient of moderate influence The design allows the model to flexibly adjust the relative importance of these two factors according to actual process requirements. Simultaneously, to accurately reflect the dynamic coupling relationship between pressing force and pressing speed, this application constructs a reference pressing force model. The model is based on the current press-fit speed index. and the preset benchmark pressing force index Combined with the speed influence coefficient The ideal pressing force at that speed is calculated in real time. Then, the actual pressing force is... With this ideal pressing force threshold Comparison, through the deviation term This quantifies the degree to which the pressing force deviates from the expected value. This deviation is then processed... This process ensures that even under extreme deviations, the impact on the health index is reasonably limited, avoiding the model's oversensitivity to outliers. Finally, the press-fitting process health assessment model integrates the assessment results of geometric alignment and force distribution with the assessment results of press-fitting force deviation, generating the press-fitting process health index through multiplication. This multiplicative approach ensures that a significant deterioration in any key factor will lead to a marked decrease in the health index, thus providing a comprehensive and real-time reflection of the overall health status of the pressing process. By integrating and modeling multi-dimensional process parameters in real-time, this solution addresses the difficulty of traditional methods in promptly detecting anomalies such as skewness and jamming during the pressing process, significantly improving the accuracy and reliability of pressing process status assessment.

[0044] In a preferred embodiment of the present invention, the temperature difference between the bearing and the housing, as well as the bearing vibration intensity during the press-fitting process, are substituted into a maximum-minimum normalization formula for normalization, and a temperature difference index and a bearing vibration intensity index are generated sequentially in the press-fitting deformation risk assessment model: ; in For temperature difference risk factor, The resistance deviation risk coefficient, For vibration risk factor, The health index inhibition coefficient, This is the weighting coefficient for the health index. , , as well as All are greater than 0, and Greater than 1; The temperature difference index. The expected deviation of the press-fit resistance. For the health index of the pressing process, The bearing vibration intensity index. It is a very small positive number. This is the pressure deformation risk index.

[0045] In this embodiment, the temperature difference data between the bearing and the housing is first collected in real time using sensors, such as infrared thermometers or thermocouples, and the vibration intensity data of the bearing during the press-fitting process is acquired, for example, using an accelerometer. The temperature difference between the bearing and the housing, as well as the bearing vibration intensity during the press-fitting process, are then substituted into a maximum-minimum normalization formula for normalization of the temperature difference index. and bearing vibration intensity index As a normalized result, it represents the relative strength of each factor in the range of 0 to 1, providing a standardized input for subsequent risk assessment models.

[0046] The press-fit deformation risk assessment model is a comprehensive mathematical model whose core lies in quantifying the deformation risk during the press-fit process by integrating multiple key parameters. The model outputs... The pressure fitting deformation risk index is a comprehensive quantitative indicator of risk. In the model... The temperature difference index reflects the temperature difference between the bearing and the housing. This difference may lead to thermal expansion and contraction, which in turn affects the fit accuracy and the risk of deformation. The expected deviation of the pressing resistance is derived from the output of the expected pressing resistance risk assessment model. It represents the degree of deviation between the actual pressing resistance and the expected resistance and is an important indicator for measuring potential anomalies in the pressing process. The press-fitting process health index is derived from the output of the press-fitting process evaluation model. It comprehensively reflects the process health status under factors such as press-fitting force, press-fitting speed, bearing pressure uniformity, and bearing-housing alignment. The bearing vibration intensity index reflects the dynamic impact and vibration level that the bearing experiences during the press-fitting process. Excessive vibration intensity may lead to localized stress concentration and deformation.

[0047] In the model For temperature difference risk factor, The resistance deviation risk coefficient, These are vibration risk coefficients, which are used to adjust the contribution weight of each factor to the total risk. The larger the value, the more significant the impact of the corresponding factor on the deformation risk. The influence of temperature difference on deformation can be analyzed based on thermal expansion coefficient experiments or finite element simulations, and the results can be obtained by fitting historical data. Based on the press-fit force-deformation relationship test, the contribution weight of resistance deviation to deformation risk can be determined through regression analysis. The impact of vibration intensity on deformation can be evaluated through vibration-deformation correlation experiments or spectrum analysis, and settings can be optimized in conjunction with process data.

[0048] The health index inhibition coefficient, It can be identified by combining expert experience with historical anomaly data; This is the weighting coefficient for the health index. The coefficients are typically determined through model sensitivity analysis or trial-and-error optimization, and are generally set to values ​​greater than 1 to enhance the inhibitory effect on healthy states. These two coefficients work together to influence the health index of the pressing process. This is to suppress or reduce the risk of deformation that occurs when the press-fitting process is in good health. It is a very small positive number, and its function is to prevent the denominator of the model from being zero, thus ensuring the stability of the calculation.

[0049] This application's solution normalizes the temperature difference between the bearing and the housing, as well as the bearing vibration intensity during the press-fitting process, using a maximum-minimum value normalization process to generate a temperature difference index and a bearing vibration intensity index. This standardizes parameters with different dimensions, facilitating model integration and avoiding errors caused by differences in magnitude. Based on this, the press-fitting deformation risk assessment model organically combines the temperature difference index, the expected deviation of press-fitting resistance, the health index of the press-fitting process, and the bearing vibration intensity index. The model uses multiplication terms... To amplify the risks caused by temperature difference, drag deviation, and vibration intensity, where the coefficient , , It can flexibly adjust the degree of impact of various risk factors. At the same time, through subtraction... To introduce a health index for the pressing process The inhibitory effect, when the health of the pressing process is good (i.e. A higher value indicates a lower overall risk assessment value, while a lower value indicates a higher risk value. The denominator is used to normalize the risk index and introduce a minimum normal value. To ensure numerical stability of the calculations, this design integrates multiple dynamic factors, enabling accurate quantitative assessment of press-fit deformation risk. The deviation from the expected press-fit resistance output by the expected press-fit resistance risk assessment model is used to determine this. And the health index of the pressing process output by the pressing process evaluation model. By combining these factors, this solution can more comprehensively consider various influencing factors from initial interference fit and surface condition to real-time pressing process conditions, thereby providing a more accurate and comprehensive deformation risk assessment and effectively making up for the shortcomings of traditional methods in risk warning.

[0050] In a preferred embodiment of the present invention, the current pressing speed is substituted into the maximum-minimum normalization formula for normalization processing, and a front pressing speed index is generated; in the pressing speed control model: ; ; in This is the risk index decay coefficient. The coefficient representing the attenuation of unhealthiness. , All are greater than 0. This represents the current pressing speed index. The pressure deformation risk index, For the health index of the pressing process, The minimum speed index allowed by the set process, The target pressing speed index, It is an arbitrary constant.

[0051] In this embodiment, the pressing speed control model can be deployed in an industrial controller, such as a programmable logic controller or an industrial PC. This controller can receive data in real time from sensors (such as force sensors, displacement sensors, vibration sensors, etc.) on pressing force, pressing speed, bearing pressure uniformity, bearing and housing alignment, bearing and housing temperature difference, and bearing vibration intensity.

[0052] The current pressing speed is substituted into the maximum-minimum normalization formula for normalization, which aims to convert the actual measured current pressing speed into a dimensionless standardized value so that the model can perform unified calculations and comparisons.

[0053] The pressing speed control model is the core algorithm used to dynamically adjust the pressing speed. Its goal is to calculate the target pressing speed for the next moment or stage based on the current risk and health status. This model combines the current pressing speed index... Press-fit deformation risk index Health index of the pressing process The pressing speed is adaptively adjusted. Among them, The minimum speed index allowed by the set process ensures that the pressing speed does not fall below the minimum limit required by the process. and This is an adjustable parameter used to control the weighting of the impact of risk and unhealthiness on speed adjustment. (Function) A truncation operation is defined to ensure that its input value It is never less than zero.

[0054] In the press-fit speed control model, it is used to ensure the speed adjustment factor. The value is always non-negative, thus avoiding the calculation of a negative target pressing speed, which conforms to the actual physical meaning and process requirements. This is the risk index attenuation coefficient, used to quantify the risk index of press-fit deformation. The degree of influence on the adjustment of the pressing speed. When At higher speeds, to reduce the risk of deformation, the pressing speed needs to be reduced. The larger the value, the greater the speed reduction. This coefficient can be set and adjusted based on actual process experience, experimental data, or optimization algorithms to balance pressing efficiency and quality risks. This is a decay coefficient for the degree of unhealthiness, used to quantify the health index of the pressing process. The degree of impact on the adjustment of the pressing speed. Due to To represent the degree of health, the model uses... This indicates the degree of unhealthiness. When the unhealthiness level is high, the pressing speed needs to be reduced to improve the process health. The larger the value, the greater the speed reduction. This coefficient can also be set and adjusted based on actual process experience, experimental data, or optimization algorithms.

[0055] The current pressing speed index is a normalized numerical representation of the current pressing speed, which serves as the basis for model calculations. The press-fit deformation risk index, output by the press-fit deformation risk assessment model, reflects the potential risk of bearing deformation during the current press-fit process. A higher value indicates a greater deformation risk. In the press-fit speed control model, it acts as a negative adjustment factor, prompting the system to reduce the press-fit speed when the risk is high. The pressing process health index, output by the pressing process evaluation model, reflects the overall health status of the current pressing process. A higher value indicates a healthier process. In the pressing speed control model, it is determined through… It acts as a negative adjustment factor, meaning that when the process is unhealthy, it prompts the system to reduce the pressing speed. The minimum permissible speed index for the set process is the minimum press-fit speed allowed by the normalized process. It serves as the lower limit of the target press-fit speed, ensuring that the press-fit speed will not fall below an acceptable minimum even under extreme risk or unhealthy conditions, in order to maintain basic press-fit efficiency and avoid other process problems. The target pressing speed index is the recommended pressing speed for the next moment or stage, calculated by the pressing speed control model and normalized. This index will serve as the control command for the actual pressing equipment, guiding the dynamic adjustment of the pressing process.

[0056] The pressing speed control model in this application achieves adaptive and dynamic adjustment of the pressing speed by standardizing the current pressing speed and combining it with the pressing deformation risk index and the pressing process health index. Specifically, the system first obtains the real-time current pressing speed and substitutes it into the maximum-minimum normalization formula to generate the current pressing speed index. This standardization process ensures that speed data with different dimensions can be uniformly input into the model for calculation, improving the model's generalization ability and accuracy. Subsequently, the pressing speed control model utilizes the current pressing speed index. The press-fit deformation risk index output by the press-fit deformation risk assessment model and the health index of the pressing process output by the pressing process evaluation model. The target pressing speed index is calculated using a preset mathematical formula. The core of this model lies in its adaptive adjustment mechanism: when the pressure deformation risk index... When the value increases, it indicates an increased risk of bearing deformation. The model will then use a risk index decay coefficient. The effect of this is to reduce the target pressing speed; at the same time, when the health index of the pressing process is... Reduce (i.e., unhealthy level) When the pressure rises, it indicates an abnormality or deviation in the pressing process. The model will then use an unhealthy attenuation coefficient. This mechanism further reduces the target pressing speed. This dual feedback adjustment mechanism allows the pressing speed to respond in real-time to risks and health conditions during the process, achieving refined control. Furthermore, a truncation function is introduced into the model. This ensures that the speed adjustment factor is always non-negative, thus fundamentally avoiding the calculation of a negative target pressing speed and guaranteeing the physical feasibility of the control commands. Simultaneously, the calculated target speed is compared with the set minimum allowable speed index for the process. By comparing the results, it is ensured that the final output target pressing speed index will not be lower than the minimum speed allowed by the process, thereby maintaining a certain level of production efficiency while ensuring pressing quality. Through the above mechanism, the pressing speed control model of this application can comprehensively consider information from multiple dimensions such as the expected deviation of pressing resistance, the health index of the pressing process, and the pressing deformation risk index, forming a closed-loop, adaptive pressing speed control strategy. This strategy not only solves the problems of lag and rigidity in speed adjustment in traditional methods, but also effectively reduces the risk of bearing deformation and improves the stability of the pressing process and the consistency of product quality through refined speed management.

[0057] As a preferred embodiment of the present invention, it also includes S5, specifically, constructing an inter-batch adaptive optimization model for lubricant dosage based on the amount of bearing lubricant sprayed in this batch, the pressure deformation risk index, the bearing vibration intensity during the pressure process, and the alignment between the bearing and the housing, and outputting the recommended spraying amount for the next batch, thereby controlling the amount of bearing lubricant sprayed in the next batch.

[0058] In this embodiment, the batch-to-batch adaptive optimization model for lubricant dosage aims to intelligently adjust the amount of lubricant sprayed during bearing press-fitting to address potential process fluctuations between different batches, thereby improving batch consistency in press-fitting quality. It can be an independent software model running on an industrial control computer or embedded system, responsible for data acquisition, model calculation, and instruction output; or it can be a functional unit integrated into the press-fitting equipment control system, communicating with sensors and actuators to dynamically optimize the amount of lubricant sprayed. The lubricant spraying amount for this batch of bearings refers to the actual amount of lubricant applied to the batch of bearings currently undergoing press-fitting before press-fitting. This amount can be monitored and recorded in real time by a flow meter or metering pump on the lubricant spraying equipment, serving as one of the benchmark inputs for evaluating the current batch's press-fitting performance; or it can be estimated based on parameters such as spraying time, spraying pressure, and nozzle diameter using a preset calculation model. The press-fitting deformation risk index, output by the press-fitting deformation risk assessment model, quantifies the likelihood of plastic deformation of the bearing during press-fitting. In this model, the bearing vibration intensity during the pressing process is a key indicator for evaluating the quality of the current batch. A higher value indicates a greater risk of bearing deformation during the pressing process, suggesting that the lubricant dosage for the next batch needs to be adjusted to reduce this risk. The bearing vibration intensity during the pressing process reflects the dynamic impact and friction state experienced by the bearing during pressing and can be monitored in real time using vibration sensors. In this model, it serves as an indicator for evaluating the stability and lubrication effect of the pressing process. Higher vibration intensity may indicate insufficient or uneven lubrication, requiring optimization of the lubricant dosage for the next batch. Bearing-housing alignment refers to the degree of alignment between the bearing's central axis and the central axis of the housing hole during the pressing process, which can be measured using methods such as visual inspection systems. In this model, it serves as an indicator for evaluating pressing accuracy and lubricant distribution uniformity. Lower alignment may be related to uneven lubricant distribution or insufficient lubrication leading to frictional imbalance, thus affecting the adjustment of the lubricant dosage for the next batch. The inter-batch adaptive lubricant dosage optimization model is the core algorithm for achieving intelligent adjustment of the lubricant dosage. Its inputs include the amount of lubricant sprayed on the bearing in this batch, the pressing deformation risk index, the bearing vibration intensity during the pressing process, and the bearing-housing alignment. This model can be built based on machine learning algorithms (e.g., reinforcement learning, neural networks, support vector machines, etc.) to establish a mapping relationship between input parameters and the optimal lubricant dosage by learning from historical batch pressing data and corresponding quality results. Alternatively, it can be designed based on expert experience rules and fuzzy logic, comprehensively judging input parameters according to a preset rule set to derive the recommended lubricant dosage. The output recommended spraying amount for the next batch is calculated by the inter-batch lubricant dosage adaptive optimization model and is a suggested lubricant spraying amount for the next batch of bearing pressing operations. This recommended amount can be a specific value (e.g., milliliters / time, grams / piece) or an adjustment ratio relative to the current batch spraying amount.The purpose is to provide an optimized target value for subsequent lubricant spraying equipment to guide the application of lubricant in the next batch. Controlling the lubricant spraying amount for the next batch of bearings refers to setting parameters or issuing commands to the lubricant spraying equipment based on the recommended spraying amount output by the batch-to-batch adaptive optimization model of lubricant dosage, to ensure that the next batch of bearings can be lubricated according to the optimized dosage before press-fitting. This can be achieved by receiving the recommended spraying amount through an industrial controller and adjusting the operating time, flow rate, or nozzle opening degree of the spraying pump; or by displaying the recommended value to the operator through a human-machine interface, allowing the operator to manually adjust the spraying equipment.

[0059] The batch-to-batch adaptive optimization model for lubricant dosage in this application aims to improve the batch consistency and reliability of bearing press-fitting by intelligently adjusting the lubricant spraying amount for the next batch through a comprehensive evaluation of key performance indicators of the current batch press-fitting process. Its working principle is as follows: First, after completing the press-fitting of the current batch of bearings, the system collects and acquires the actual data of the lubricant spraying amount for this batch of bearings, and combines this with the press-fitting deformation risk index output by the press-fitting deformation risk assessment model, as well as key performance indicators such as bearing vibration intensity and bearing-housing alignment obtained from sensors. These indicators comprehensively reflect the quality status and potential problems of the current batch press-fitting process. Subsequently, this multi-source data is input into the batch-to-batch adaptive optimization model for lubricant dosage. This model does not simply respond to a single indicator, but rather uses complex algorithmic logic to comprehensively analyze and weigh these interrelated indicators. For example, if the press-fitting deformation risk index is high and the bearing vibration intensity is also high, the model may determine that the current lubricant dosage is insufficient or the lubrication effect is poor, thus tending to increase the lubricant dosage for the next batch. Conversely, if all indicators perform well, the model may maintain or fine-tune the lubricant dosage. The core of this model lies in establishing a feedback mechanism between the current batch lubricant dosage and the press-fit quality performance, and inferring the lubricant dosage that can optimize the press-fit quality of the next batch. Ultimately, the inter-batch lubricant dosage adaptive optimization model outputs a calculated and optimized recommended spraying amount for the next batch. This recommended amount is then sent to the control system of the lubricant spraying equipment, precisely controlling the amount of lubricant sprayed on the next batch of bearings before press-fitting via instructions or parameter settings. Through this closed-loop feedback and adaptive adjustment mechanism, the proposed solution effectively overcomes the limitations of fixed lubricant dosage or adjustment based solely on experience in traditional press-fitting processes, enabling the application of lubricant to dynamically adapt to actual process conditions and quality requirements. This model, together with the press-fitting resistance expected risk assessment model, press-fitting process assessment model, press-fitting deformation risk assessment model, and press-fitting speed control model, constitutes a more complete intelligent press-fitting system. Key data such as the press-fitting deformation risk index, bearing vibration intensity during the press-fitting process, and bearing-housing alignment are all derived from the real-time monitoring and evaluation results of the aforementioned models. This means that lubricant optimization is based on a comprehensive and in-depth understanding of the press-fitting process, rather than being done in isolation. For example, when the press-fitting deformation risk index is high, it not only triggers adjustments to the press-fitting speed but also serves as an important basis for optimizing the lubricant dosage, improving press-fitting conditions from the source. This information sharing and decision-making among multiple models makes the control of the entire press-fitting process more refined and intelligent, significantly improving the stability and consistency of press-fitting quality and effectively solving the problem of batch-to-batch quality fluctuations.

[0060] In a preferred embodiment of the present invention, the amount of bearing lubricating oil sprayed in this batch is substituted into the maximum-minimum normalization formula for normalization processing, and an index of the amount of lubricating oil sprayed in this batch is generated; in the batch-to-batch adaptive optimization model for lubricating dosage: ; ; ; in The weight is affected by the deformation during press fitting. Weighting for the impact of bearing vibration. For the moderate impact weight, and , For learning rate, , The estimated value is the effect of lubricant dosage. , Score the target performance. For comprehensive performance scoring, The pressure deformation risk index, The bearing vibration intensity index. For the moderate index, This refers to the coating amount index of this batch of lubricating oil. The recommended coating amount index for the next batch is given, where a, b, and c are all arbitrary constants.

[0061] In this embodiment, the amount of bearing lubricating oil sprayed in this batch is substituted into the maximum-minimum normalization formula for normalization processing, and an index of the amount of lubricating oil sprayed in this batch is generated. The purpose is to convert the actual amount of lubricating oil sprayed (which may be expressed in milliliters, grams, or spraying time, etc.) into a dimensionless index. In the maximum-minimum normalization formula, the maximum value and the minimum value are the minimum and maximum spraying amounts set historically or by the process, respectively.

[0062] In the batch-to-batch adaptive optimization model for lubricant dosage, the overall performance score is... This formula defines the overall performance score. This is used to quantify the overall quality performance of the current batch of bearing press-fitting. It comprehensively considers the press-fitting deformation risk index. Bearing vibration intensity index And the moderate index By and use and This approach transforms all indicators into positive metrics where "better performance equates to higher numerical values," facilitating unified weighting. The model can be implemented by a separate computational unit that receives data from the press-fitting deformation risk assessment model, the press-fitting process assessment model (providing a neutrality index), and vibration sensors (providing a bearing vibration intensity index), and performs calculations based on preset weighting coefficients. Alternatively, the model can be integrated into a central control system and implemented through a software algorithm model that periodically acquires the latest data from various sensors and assessment models, updating the overall performance score in real time.

[0063] in, The weight is affected by the deformation during press fitting. Weighting for the impact of bearing vibration. For the moderate impact weight, and Weighting coefficients , , Used to adjust for press-fit deformation risk, bearing vibration, and alignment in the overall performance score. The relative importance of each indicator is considered. Their sum equals 1, ensuring the reasonableness of the score and preventing some indicators from being overemphasized or ignored. These weights can be determined based on actual process requirements, experiential knowledge, or historical data analysis. These weight coefficients can be stored as system parameters in a configuration database, allowing process engineers to manually adjust and optimize them according to actual production conditions. Alternatively, these weight coefficients can be adaptively learned using machine learning algorithms; for example, by analyzing the correlation between various indicators and final product quality in a large amount of historical pressing data, the weights can be automatically adjusted to maximize the optimization effect.

[0064] Recommended coating volume index for the next batch This formula is used to calculate the recommended lubricant spray coverage index for the next batch. It is based on the coating coverage index of the current batch. And score based on the overall performance of the current batch. Performance score Adjust for the differences between them. Learning rate Control the step size of the adjustment, and This indicates the direction of the effect of lubrication dosage on performance. The clip function ensures... The parameters are always kept within an effective range (0 to 1). This calculation can be performed by a standalone controller or software model that performs calculations based on collected performance data after each batch of press-fitting is completed, and sends the results to the lubricant spraying equipment to adjust the spraying amount for the next batch. Alternatively, the calculation can be integrated into the main control PLC or industrial PC as part of the overall press-fitting control system, enabling real-time or near real-time inter-batch parameter adjustments.

[0065] in, The learning rate is the learning rate. The learning rate is a parameter between 0 and 1 that determines the step size for each adjustment of the lubricant spray amount. A smaller learning rate leads to smoother, more stable adjustments but slower convergence; a larger learning rate may lead to faster convergence but may also cause system oscillations or instability. The learning rate can be set to a fixed value, such as 0.1 or 0.05, to strike a balance between stability and convergence speed. The learning rate can also be adaptive; for example, a larger learning rate can be used when the system performance deviates significantly from the target, while gradually decreasing the learning rate as it approaches the target to achieve fast convergence and fine-tuning.

[0066] The estimated value is the effect of lubricant dosage. . This is an estimate indicating the direction of the effect of lubricant dosage on press-fit performance. If increasing the lubricant dosage typically improves performance, then... +1; if increasing the lubricant dosage typically degrades performance, then It is -1. This parameter allows the model to be correctly adjusted according to the actual mechanism of action of the lubricant. Pre-set parameters can be based on the experience of domain experts or prior experimental data. For example, for most interference fit press fittings, appropriate lubrication can usually reduce resistance and wear, thereby improving performance. It can be set to +1. In more complex scenarios, It can also be determined dynamically through online learning or A / B testing. For example, the system can try slightly increasing or decreasing the lubricant dosage, observe the performance change trend, and thus automatically infer... The value of .

[0067] Score the target performance. For comprehensive performance scoring, The pressure deformation risk index, The bearing vibration intensity index. For the moderate index, This refers to the coating amount index of this batch of lubricating oil. The recommended coating amount index is used for the next batch. These are key variables used in the model, representing different states and outcomes during the pressing process. These indices and scores can be calculated through sensor data acquisition, preprocessing, and the aforementioned pressing resistance expected risk assessment model, pressing process assessment model, and pressing deformation risk assessment model. It can be set by process experts based on product quality requirements and historical best practices, or it can be dynamically adjusted during the production process through optimization algorithms to adapt to constantly changing production conditions.

[0068] . The function is a truncation function used to truncate a value. Limit to the specified minimum value and maximum value Between. If Less than Then return ;if Greater than Then return Otherwise return In this model, it ensures that the calculated recommended coating amount index for the next batch is accurate. Always maintain the coating within the effective physical or technological limits (e.g., 0 to 1) to prevent unreasonable or excessive coating amounts from exceeding the equipment's capabilities. This function can be implemented using built-in functions in a programming language or a user-defined function. At the hardware level, the corresponding comparison and selection operations can be performed by a digital signal processor or an arithmetic logic unit in a microcontroller.

[0069] This batch-to-batch adaptive optimization model for lubricant dosage achieves intelligent adjustment of the lubricant spraying amount during bearing press-fitting by establishing a feedback closed loop. Its core operating logic involves first standardizing the lubricant spraying amount for the current batch of bearings, converting it into a dimensionless index of the lubricant spraying amount for that batch. This provides a unified input basis for subsequent mathematical model calculations. Based on this, the model comprehensively evaluates the pressing performance of the current batch. It utilizes the pressing deformation risk index. Bearing vibration intensity index And the moderate index These three key indicators are used to calculate the overall performance score through a weighted summation. .in, and Converted to its complement and This ensures that all metrics are aligned with the principle of "better performance, higher scores," thereby guaranteeing a consistent scoring system. It can accurately reflect the overall pressing quality of the current batch. Weighting coefficient. , , This configuration allows the system to flexibly adjust the contribution of different risk factors to the overall score based on the actual sensitivity of the process to those factors. Subsequently, the system calculates the comprehensive performance score for the current batch. Compared with the preset target performance score Compare. If Below This indicates that the current batch's press-fit performance is poor and requires adjustment of the lubricant dosage to improve it; conversely, if Reaching or exceeding If so, it may be necessary to fine-tune or maintain the current lubrication dosage. The adjustment range depends on the learning rate. Control measures ensured the smoothness and stability of the adjustment process. The estimated effect of lubricant dosage... This guides the direction of adjustments; for example, if increasing the lubricant typically improves performance, then... For a +1 value, the system will adjust the coating amount accordingly based on the performance difference. Ultimately, through... The function will calculate the recommended spraying amount index for the next batch. By limiting the amount of lubricant sprayed to an effective range of 0 to 1, unreasonable or excessive amounts exceeding physical limits are avoided. This process forms a continuous optimization cycle: the results of each batch of press-fitting are fed back into the model to guide the adjustment of the lubricant dosage for the next batch, thereby achieving adaptive optimization of the lubricant spraying amount. This scheme is closely integrated with pre-processing techniques such as press-fitting deformation risk assessment models, press-fitting process evaluation models, and bearing vibration intensity measurements. The press-fitting deformation risk index, bearing vibration intensity index, and centering index are all derived from real-time monitoring and evaluation of the press-fitting process. By using these real-time, quantitative performance indicators as inputs to the lubricant dosage optimization model, this scheme can make decisions based on actual press-fitting quality performance, rather than relying solely on preset parameters. This deep integration makes lubricant adjustment no longer a blind, experience-based operation, but a data-driven intelligent optimization, significantly improving the consistency and reliability of press-fitting quality between batches and effectively solving the problem of lagging and rigid lubricant dosage adjustment in traditional processes.

[0070] like Figure 2 As shown, a bearing processing press-fitting device, applied to the bearing processing press-fitting method in the above embodiments, includes a frame 1. The frame 1 is equipped with an electric clamping assembly 2 for fixing a housing and a press-fitting machine 3 for pressing down the bearing. The press-fitting machine 3 is disposed on the upper side of the electric clamping assembly 2. The device also includes: Memory, used to store executable instructions; The processor, when executing executable instructions stored in the memory, implements the bearing pressing method in the above embodiments.

[0071] In this embodiment, frame 1 serves as the basic support structure of the pressing device. Its main function is to provide a stable working platform and sufficient rigidity to withstand the forces and vibrations generated during the pressing process, ensuring the stability and accuracy of the entire device. The electric clamp assembly 2 is used to precisely fix the housing during the pressing process, preventing displacement or vibration of the housing under stress, thereby ensuring the alignment accuracy of the bearing and the housing and the pressing quality. The electric clamp assembly 2 can be implemented in various ways; for example, an electro-hydraulic clamp can be used, with an electric pump providing hydraulic power to drive the clamping mechanism. The pressing machine 3 is the core component for realizing the bearing pressing operation. Its function is to provide controllable pressing force and displacement, precisely pressing the bearing into the housing. The pressing machine 3 can employ various driving methods; for example, a servo electric cylinder can be used, with a servo motor driving a ball screw or planetary roller screw to achieve high-precision force / displacement control.

[0072] The memory stores the executable instructions for the press-fitting method used in bearing machining. These instructions include the logic, parameters, and control strategies of the intelligent press-fitting algorithm. The memory can take various forms; for example, non-volatile memory such as flash memory or solid-state drives can be used to ensure that instructions are not lost after power failure; random access memory can also be used as working memory to temporarily store program execution data and intermediate results. The processor is the "brain" of the press-fitting device, responsible for parsing and executing the executable instructions stored in the memory, thereby achieving real-time monitoring, data processing, and control of the press-fitting process. The processor can be of various types; for example, a high-performance industrial-grade central processing unit can be used to handle complex algorithms and large amounts of data; a microcontroller can be used, suitable for control tasks with high real-time requirements and relatively low computational complexity; or a field-programmable gate array (FPGA) can be used to achieve high-speed parallel computing and customized logic control.

[0073] The bearing pressing device of this application organically combines mechanical structure with intelligent control system to form a closed-loop automated pressing workstation. Specifically, frame 1, as the physical foundation of the entire device, provides stable support and precise installation reference, ensuring the mechanical accuracy and stability of the pressing process. The electric clamp assembly 2 is responsible for firmly fixing the housing to be pressed in a preset position before pressing, ensuring that the housing will not experience any displacement or vibration during the pressing process, thus providing a prerequisite for precise alignment and pressing of the bearing. The pressing machine 3, as the actuator, has precise force / displacement control capability, which is key to realizing the intelligent pressing method. It can press the bearing into the housing at a preset or dynamically adjusted speed and force according to control commands. Furthermore, the memory pre-stores all executable instructions for the bearing pressing method, covering core algorithms such as pressing resistance expectation risk assessment, pressing process health assessment, pressing deformation risk assessment, and pressing speed control. The processor reads and executes these instructions in real time, interacting with the press machine 3, the electric clamp assembly 2, and other sensors to transmit control signals, thereby achieving intelligent management of the entire press process. Based on the real-time collected press data and the algorithm model in memory, the processor dynamically calculates the expected deviation of press resistance, the health index of the press process, and the risk index of press deformation, and adjusts the press speed of the press machine 3 accordingly, thus achieving adaptive control of the press process. This integrated hardware and software design enables the device to respond to various dynamic changes during the press process in real time, effectively avoiding the problem of unimplemented intelligent control in traditional press processes due to a lack of dedicated hardware support.

[0074] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A press-fitting method for bearing machining, characterized in that, Includes the following steps: S1. Construct a press-fit resistance expected risk model based on the theoretical interference fit between the bearing and the housing, the surface roughness of the mating surface, and the cleanliness of the mating surface, and output the expected deviation of the press-fit resistance; S2. Construct a health assessment model for the pressing process based on pressing force, pressing speed, bearing pressure uniformity, and bearing-housing alignment, and output the pressing process health index. S3. Based on the temperature difference between the bearing and the housing, the expected deviation of the pressing resistance, the health index of the pressing process, and the bearing vibration intensity during the pressing process, a pressing deformation risk assessment model is constructed, and the pressing deformation risk index is output. S4. Construct a pressing speed control model based on the pressing deformation risk index, the pressing process health index, and the current pressing speed, and output the target pressing speed.

2. The press-fitting method for bearing processing according to claim 1, characterized in that, In step S1, the theoretical interference fit between the bearing and the housing, the surface roughness of the mating surface, and the cleanliness of the mating surface are successively substituted into the maximum-minimum value formula for normalization, and the theoretical interference fit index, surface roughness index, and mating surface cleanliness index are generated sequentially; in the press-fit resistance expected risk model: The theoretical interference index and surface roughness index are multiplied by their respective preset influence coefficients and summed. The result is then subtracted from the difference between the mating surface cleanliness index and its corresponding preset influence coefficient, and a preset bias constant is added. The final result is then input into the Sigmoid function for nonlinear mapping, and the expected deviation of the press-fit resistance is output. The sum of all preset influence coefficients is 1.

3. The press-fitting method for bearing processing according to claim 1, characterized in that, In step S2, the bearing pressure uniformity, bearing and housing alignment, pressing speed and reference pressing force are substituted into the maximum-minimum normalization formula for normalization, and the pressure uniformity index, alignment index, pressing speed index and reference pressing force index are generated in sequence. In the health assessment model for the pressing process: First, calculate the reference pressing force corresponding to the current speed based on the linear relationship between the baseline pressing force index and the pressing speed index; Secondly, a weighted power product of the pressure uniformity index and the neutrality index is used as the basic assessment value for process health. Finally, by calculating the absolute deviation between the actual pressing force and the reference pressing force, and introducing a preset penalty coefficient to scale the absolute deviation, the scaling result is limited to the range of 0 to 1. Then, the penalty adjustment factor is obtained by subtracting the limit value from 1. The basic evaluation value is multiplied by the penalty adjustment factor to output the pressing process health index.

4. The press-fitting method for bearing processing according to claim 1, characterized in that, In step S3, the temperature difference between the bearing and the housing, as well as the bearing vibration intensity during the press-fitting process, are substituted into the maximum-minimum normalization formula for normalization. The temperature difference index and the bearing vibration intensity index are then generated sequentially in the press-fitting deformation risk assessment model. Calculate "1 plus (the product of temperature difference index and its risk coefficient)", "1 plus (the product of expected deviation of press-fit resistance and its risk coefficient)" and "1 plus (the product of bearing vibration intensity index and its risk coefficient)" respectively, and multiply the three together to obtain the positive risk factor. Simultaneously, the inhibition factor is obtained by subtracting the inhibition factor from the health index inhibition coefficient multiplied by the power operation result of the health index of the pressing process from 1; then the inhibition factor is subtracted from the positive risk factor and divided by a normalized benchmark value to finally obtain the pressing deformation risk index; where each preset risk coefficient, inhibition coefficient and weight coefficient are positive values.

5. The press-fitting method for bearing processing according to claim 1, characterized in that, In step S4, the current pressing speed is substituted into the maximum-minimum normalization formula for normalization and the front pressing speed index is generated. In the press-fit speed control model: The target pressing speed index is based on the current pressing speed index. The total attenuation is calculated by multiplying the pressing deformation risk index by the preset first attenuation coefficient, and the unhealthy degree of the pressing process health index (calculated by subtracting the health index from 1) by the preset second attenuation coefficient. The result of subtracting the total attenuation from 1 is multiplied by the current pressing speed index to ensure that the adjustment result is not lower than the set minimum allowable speed index of the process. The attenuation coefficients mentioned above are all positive values.

6. The press-fitting method for bearing processing according to claim 1, characterized in that, It also includes step S5, which specifically involves constructing an adaptive optimization model for lubricant dosage between batches based on the amount of bearing lubricant sprayed in this batch, the pressure deformation risk index, the bearing vibration intensity during the pressure process, and the alignment between the bearing and the housing, and outputting the recommended spraying amount for the next batch, thereby controlling the amount of bearing lubricant sprayed in the next batch.

7. The press-fitting method for bearing processing according to claim 6, characterized in that, The amount of bearing lubricating oil sprayed in this batch is substituted into the maximum-minimum normalization formula for normalization, and the lubricating oil spraying amount index for this batch is generated; in the batch-to-batch adaptive optimization model for lubricating dosage: First, subtract the pressure deformation risk index from 1 and the bearing vibration intensity index from 1, respectively, and multiply them by their corresponding preset influence weights. Then, add them to the product of the neutrality index and its influence weight to calculate the comprehensive performance score. Secondly, the difference between the comprehensive performance score and the preset target performance score is multiplied by the preset learning rate and the estimated value of the influence direction of the lubricant dosage to obtain the adjustment amount; Finally, the adjustment amount is added to the lubricant spraying amount index of this batch, and the result is limited to the allowable range to output the recommended spraying amount index for the next batch; wherein the sum of the preset influence weights is 1, the learning rate is between 0 and 1, and the estimated value of the influence direction of the lubricant dosage is a positive or negative constant.

8. A bearing processing press-fitting device, applied to the bearing processing press-fitting method of any one of claims 1-7, comprising a frame (1), wherein the frame (1) is provided with an electric clamp assembly (2) for fixing a housing and a press-fitting machine (3) for pressing down the bearing, the press-fitting machine (3) being disposed on the upper side of the electric clamp assembly (2), characterized in that, Also includes: Memory, used to store executable instructions; The processor, when executing executable instructions stored in the memory, implements the press-fitting method for bearing machining according to any one of claims 1-7.

Images