Multiphysics Coupled Dynamic Evolution Modeling and Accuracy Control Method for Heavy-Duty Spindle
By using a multi-physics coupled dynamic evolution modeling and precision control method for heavy-duty spindles and iteratively solving the shaft displacement using a twin model, the problem of dynamic evolution of heavy-duty spindles during cutting was solved, real-time precision control was achieved, the consistency between simulation and reality was improved, and the engineering difficulty was reduced.
Patent Information
- Application Number
- CN202611636835.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-10-22
- Publication Date
- 2026-11-17
AI Technical Summary
Existing technologies cannot effectively reflect the dynamic evolution behavior of heavy-duty spindles caused by multi-physics coupling during cutting, and the accuracy control methods rely on estimated working conditions or simple table lookups, resulting in large deviations between simulation results and actual results and limited compensation effects.
A multi-physics coupled dynamic evolution modeling and precision control method for heavy-duty spindles is adopted. Through closed-loop iteration of two sub-models, namely the spindle body and bearing support, the spatial displacement of the shaft system is solved in real time, the rotation error is calculated and compensation commands are generated, thereby achieving active precision control.
It enables real-time precision control of heavy-duty spindles during cutting, overcomes the shortcomings of traditional models, improves the consistency between simulation and actual service conditions, avoids the accumulation of prediction errors, and reduces the difficulty of engineering implementation.
Smart Images

Figure CN123116747A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of precision control technology for heavy-duty spindles of CNC machine tools, specifically to a method for multi-physics field coupled dynamic evolution modeling and precision control of heavy-duty spindles. Background Technology
[0002] Heavy-duty spindles are widely used in large CNC gear hobbing machines, heavy-duty boring and milling machines, and other equipment to undertake cutting tasks with large cutting forces and high torques. During heavy-duty cutting, the spindle bearing is subjected to large radial and axial cutting loads, resulting in high bearing contact pressure and significant bearing frictional heat generation, accompanied by changes in the state of the lubricating oil film. The feedback coupling between the structural force field, temperature field, and flow field causes the bearing preload to drift and the support stiffness and damping to continuously change. Consequently, the dynamic characteristics of the entire spindle system evolve continuously with machining time, and the spindle rotation accuracy deteriorates continuously, seriously affecting the machining quality of heavy parts.
[0003] Currently, there are significant technical shortcomings in domestically produced heavy-duty spindle technology: 1. Existing technologies for modeling and analyzing heavy-duty spindles mostly adopt time-invariant dynamic models, setting the bearing support stiffness and damping as fixed constants. This can only characterize the spindle state under a certain static working condition and cannot reflect the dynamic evolution behavior caused by the coupling of multiple physical fields during the entire cutting process. Although some schemes perform unidirectional coupling simulation of force field and temperature field, they ignore the participation of flow field and bidirectional feedback between physical fields, resulting in a large deviation between the simulation results and the actual service state of the spindle.
[0004] 2. In the existing technology, some methods for controlling the precision of heavy-duty spindles rely on future estimated working conditions to predict errors. However, it is difficult to obtain estimated working conditions, and the prediction has cumulative errors. Other methods establish a table of correspondence between errors and working condition parameters through a large number of pre-calibration tests. During the machining process, the preset compensation amount is simply retrieved based on the collected working condition parameters, without considering the influence of the dynamic changes of bearing support characteristics with multiple physical fields, thus limiting the compensation effect.
[0005] Therefore, there is an urgent need for a method that can solve the multi-physics coupled dynamic state of the heavy-duty spindle system online in real time and achieve active control of accuracy based on real-time rotational error. Summary of the Invention
[0006] The purpose of this invention is to provide a method for multi-physics coupling dynamic evolution modeling and precision control of heavy-duty spindles. By using two sets of sub-models, namely the spindle body and bearing support, to achieve closed-loop iteration, the spatial displacement of the shaft system is solved in real time, the current rotation error is calculated, and compensation instructions are generated based on the error threshold judgment results, thereby realizing active control of the rotation precision of heavy-duty spindles.
[0007] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A method for multiphysics coupled dynamic evolution modeling and accuracy control of heavy-load spindles includes the following steps: S1. Real-time operating condition data acquisition: Acquire real-time operating condition data during the service of heavy-duty spindles. The real-time operating condition data includes cutting force, spindle speed, bearing temperature, spindle body temperature, and lubrication medium status. S2. Construct a multi-physics coupled dynamic evolution model for the heavy-duty spindle system: decompose the heavy-duty spindle system into a spindle body structure sub-model and a bearing support evolution sub-model; The spindle body structure sub-model receives the collected cutting force and spindle body temperature, maps the cutting force to generate a cutting load vector F(t), and calculates and maps the equivalent thermal load vector F from the spindle body temperature. T (t), a constant mass matrix M is pre-constructed; the bearing support evolution sub-model receives spindle speed, bearing temperature, lubrication medium state, and the relative deformation of bearing position fed back from step S3. The time-varying equivalent stiffness matrix K(Ψ(t)) and the time-varying equivalent damping matrix C(Ψ(t)) are calculated and output; where Ψ(t) is the bearing support state variable that changes with time; the two sub-models are coupled to obtain the complete main shaft system dynamic evolution model; S3. Solving the dynamic evolution model of the main spindle system and extracting displacements: Solve the complete dynamic evolution model of the main spindle system to obtain the discrete node displacement vector x(t) of the main spindle system. Extract the bearing position node displacements from x(t), and calculate the relative deformation of the bearing position based on the bearing position node displacements and the fixed position of the bearing outer ring. ,Will Feedback is sent to the bearing support evolution sub-model; S4. Rotation accuracy calculation and threshold identification: Based on the displacement of the spindle front node extracted from x(t). Calculate the current rotation error and compare it with a preset threshold. If the rotation error exceeds the preset threshold, output the rotation accuracy compensation amount; otherwise, set the compensation amount to zero. S5. Precision Active Control: Generate a precision active control command based on the rotational precision compensation amount to drive the spindle actuator to perform rotational precision active control.
[0008] Furthermore, in step S2, the cutting load vector F(t) and equivalent thermal load vector F output by the spindle body structure sub-model are... T The constant mass matrix M is coupled with the time-varying equivalent stiffness matrix K(Ψ(t)) and time-varying equivalent damping matrix C(Ψ(t)) output by the bearing support evolution sub-model to obtain the complete dynamic evolution model of the main shaft system. Its dynamic equations are:
[0009] In the formula, Axis acceleration vector The shaft system velocity vector, x(t) is the shaft system displacement vector, Ψ(t) is the bearing support state variable, M is the constant mass matrix, C(Ψ(t)) is the time-varying equivalent damping matrix, K(Ψ(t)) is the time-varying equivalent stiffness matrix, F(t) is the cutting load vector, and F... T (t) is the equivalent thermal load vector.
[0010] Furthermore, the spindle body structure sub-model incorporates spindle geometric parameters and material fixed parameters, and the constant mass matrix M is determined based on the spindle geometric parameters and material fixed parameters, and does not change with the real-time operating data.
[0011] Furthermore, in step S3, at the initial moment of iteration start, the initial relative deformation of the bearing corresponding to the static assembly is calculated offline based on the spindle geometric parameters and bearing preload, and used as the initial input of the bearing support evolution sub-model; starting from the first discrete time step, the relative deformation of the bearing position is extracted from the solution results of the complete spindle system dynamic evolution model.
[0012] Furthermore, the bearing support evolution sub-model solves for the rolling element-raceway solid contact stiffness based on Hertzian contact theory, and solves for the oil film stiffness and oil film damping based on the elastohydrodynamic lubrication equations. The solid contact stiffness and oil film stiffness are connected in series to obtain the total equivalent stiffness of the bearing.
[0013] Furthermore, the time-varying equivalent stiffness matrix K(Ψ(t)) and the time-varying equivalent damping matrix C(Ψ(t)) are the comprehensive equivalent matrix parameters of the support system composed of the bearing solid contact pair and the lubricating oil film, and are not the inherent stiffness and damping of the bearing metal parts material.
[0014] Furthermore, the bearing support evolution sub-model incorporates bearing geometric parameters, bearing material fixed parameters, and initial values of static assembly deformation. These initial values of static assembly deformation serve as one of the bases for calculating the time-varying equivalent stiffness matrix K(Ψ(t)) and the time-varying equivalent damping matrix C(Ψ(t)).
[0015] Furthermore, the bearing support state variable Ψ(t) characterizes the evolution of bearing preload drift and lubricating medium state under the coupling effect of structural force field, temperature field and flow field. The bearing support state variable Ψ(t) includes multiple factors such as bearing contact load, oil film thickness, oil film pressure, bearing preload, lubricating oil viscosity and thermal expansion caused by bearing temperature.
[0016] Furthermore, the bearing support state variable Ψ(t) is obtained by coupling the structural force field, temperature field, and flow field; wherein, the structural force field characterizes the bearing contact pressure and shaft elastic deformation under cutting load, the temperature field characterizes the temperature distribution and thermal expansion deformation caused by bearing frictional heat generation and motor loss heat generation, and the flow field characterizes the flow characteristics of bearing lubricating oil film and the heat transfer characteristics of cooling medium; the time-varying equivalent stiffness matrix K(Ψ(t)) and the time-varying equivalent damping matrix C(Ψ(t)) are input into the complete main shaft system dynamic evolution model for solution to obtain the shaft system vibration and dynamic rotation response.
[0017] Furthermore, the spindle actuator includes one or more of the following: a bearing preload adjustment mechanism, a cooling flow rate adjustment mechanism, and a CNC system coordinate compensation unit; the control methods include: adjusting the bearing preload, adjusting the cooling medium flow rate, and compensating for CNC coordinate axis errors.
[0018] The beneficial effects of the above technical solution are as follows: 1. This invention adopts a dual-model decoupling architecture, distinguishing between the constant inherent properties of the spindle body and the time-varying evolution characteristics of the bearing support. The heavy-duty spindle system is divided into a spindle body structure sub-model and a bearing support evolution sub-model. The spindle body sub-model outputs a constant mass matrix and cutting load and thermal load vectors, while the bearing support evolution sub-model outputs time-varying equivalent stiffness and damping matrices that change with the multi-physics field state. This constructs a complete dynamic evolution model that can characterize the force-thermal-fluid multi-physics field coupling of the heavy-duty spindle system, overcoming the limitation of traditional time-invariant dynamic models that can only characterize a single instantaneous working condition.
[0019] 2. This invention feeds back the relative deformation of the bearing position obtained from the dynamic solution to the bearing support evolution sub-model, forming a two-way coupled iteration. This feedback allows the bearing support characteristics to be updated in real time with the machining process. Compared with unidirectional coupled simulation, this invention considers the participation of the flow field and the two-way feedback between the physical fields, improving the consistency between the simulation state and the actual service state of the spindle.
[0020] 3. This invention calculates the rotational error based on the displacement of the spindle front end node obtained in real time, and outputs compensation commands according to the threshold judgment results to drive the spindle actuator to adjust the support state, thereby achieving active control of accuracy. Unlike simply looking up a table to retrieve a preset compensation amount, it can actively suppress the deterioration of rotational accuracy under heavy load and long-term cutting, and does not require prediction and deduction, avoiding the problems of difficult working condition estimation and prediction error accumulation. It is easy to implement in engineering and has less pressure for online real-time calculation. Attached Figure Description
[0021] Figure 1 This is a flowchart illustrating the overall process of the multiphysics coupling dynamic evolution modeling and precision control method for heavy-duty spindles according to the present invention. Figure 2 This is a flowchart of the method steps of the present invention. Detailed Implementation
[0022] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.
[0023] It should be noted that, unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.
[0024] Example 1:
[0025] This embodiment uses the heavy-duty spindle system of a certain type of heavy-duty CNC gear hobbing machine as an example to illustrate the complete implementation process of the method of the present invention. The spindle is supported by two sets of rolling bearings at the front and rear, and bears large cutting forces and large torques. During the machining process, the bearings generate significant heat, and the lubrication status changes with temperature, making the rotational accuracy prone to drift.
[0026] like Figure 1 and Figure 2 As shown, the multiphysics coupled dynamic evolution modeling and accuracy control method for heavy-duty spindles in this embodiment includes the following steps: S1, Real-time Operating Condition Data Acquisition During heavy-duty spindle operation, real-time operating data is collected through sensors. Specifically: a piezoelectric force gauge is used to collect three-dimensional cutting forces, including radial force, axial force, and cutting torque; a speed sensor mounted on the spindle is used to collect spindle speed; a temperature sensor is used to collect bearing temperature and spindle body temperature; and a lubricating oil condition sensor is used to collect the lubricating medium condition, including lubricating oil viscosity and supply pressure.
[0027] The aforementioned sensor signals are synchronously acquired by the data acquisition module and then sent to the subsequent spindle body structure sub-model and bearing support evolution sub-model. The synchronization period for data acquisition can be set to 2ms to meet the control requirements under heavy-load cutting.
[0028] S2. Construct a multi-physics coupled dynamic evolution model for the heavy-duty spindle system. A dual-model decoupling architecture is adopted, which splits the heavy-duty spindle system into a spindle body structure sub-model and a bearing support evolution sub-model. The two are modeled and output separately, and then coupled through dynamic equations, thereby separating the constant inherent properties of the spindle body from the time-varying evolution characteristics of the bearing support.
[0029] (1) Construction of the main shaft body structure sub-model The spindle body structure sub-model incorporates the spindle's geometric parameters and fixed material parameters, including the geometric dimensions, elastic modulus, and density of each segment of the spindle. Based on these parameters, a constant mass matrix M is constructed offline.
[0030] The offline construction method for the constant mass matrix M can employ the finite element discretization method: the spindle is discretized along its axis into several solid elements, and the element mass matrix is assembled according to the standard finite element format to obtain the overall mass matrix M of the spindle body. Since the spindle geometry and material are fixed, M does not change with real-time operating data and can be repeatedly called after being constructed offline once. Therefore, when solving the dynamic equations at each time step, it is not necessary to reassemble the mass matrix; only the time-varying stiffness matrix and damping matrix corresponding to the bearing support need to be updated, thereby reducing the computational load of online solution.
[0031] The spindle body structure sub-model receives the collected cutting force and spindle body temperature, and maps them into a load vector: the three-dimensional cutting force is mapped to the front end node of the spindle to generate the overall cutting load vector F(t); the thermal strain of each axis element is calculated based on the spindle body temperature field, and the equivalent thermal load vector F is generated. T (t).
[0032] (2) Construction of the bearing support evolution sub-model The bearing support evolution sub-model incorporates bearing geometric parameters (such as rolling element diameter, raceway curvature radius, bearing width, pitch circle diameter, etc.), fixed bearing material parameters (such as elastic modulus, Poisson's ratio, etc.), and initial values of static assembly deformation.
[0033] The initial value of static assembly deformation refers to the initial relative deformation of the inner and outer rings of the bearing under preload in the static assembly state of the spindle. At the initial moment of iteration start, the initial value of static assembly deformation is used as the initial input of the bearing support evolution sub-model to calculate the bearing support state variables Ψ(t0) such as bearing contact load and oil film thickness at the initial moment.
[0034] The bearing support evolution sub-model receives spindle speed, bearing temperature, lubrication medium condition, and the relative deformation of bearing position fed back from step S3. This is used as the bearing position boundary condition to update the bearing multiphysics state.
[0035] The bearing multiphysics state is represented by the bearing support state variable Ψ(t). Ψ(t) is a bearing support state quantity that varies with time, and its internal components include, but are not limited to: contact load between the rolling elements and the raceway, contact deformation, oil film thickness, oil film pressure, bearing preload, lubricating oil viscosity, and thermal expansion derived from the bearing temperature field. Ψ(t) characterizes the evolution of bearing preload drift and lubricating medium state under the coupling effects of the structural force field, temperature field, and flow field.
[0036] In this embodiment, the primary physical fields involved in the coupling include the structural force field, temperature field, and flow field; the shaft dynamic response is the dynamic behavior of the system obtained after the coupling of force-heat-flow multi-physics fields. The structural force field characterizes the bearing contact pressure and shaft elastic deformation under cutting load; the temperature field characterizes the temperature distribution and thermal expansion deformation caused by bearing frictional heat generation and motor loss heat generation; the flow field characterizes the flow characteristics of the bearing lubricating oil film and the heat transfer characteristics of the cooling medium. The above three types of physical fields are interconnected through the bearing support state variable Ψ(t): the structural force field and temperature field affect the bearing contact load and thermal expansion, and the flow field affects the oil film thickness and oil film pressure. The above physical quantities together constitute Ψ(t); Ψ(t) outputs time-varying equivalent stiffness and damping, which are input into the complete spindle shaft dynamic evolution model for solution, and finally the shaft vibration and dynamic rotation response are obtained.
[0037] The bearing support evolution sub-model calculates and outputs the time-varying equivalent stiffness matrix K(Ψ(t)) and the time-varying equivalent damping matrix C(Ψ(t)) based on the updated Ψ(t).
[0038] Calculation of the time-varying equivalent stiffness matrix K(Ψ(t)): First, the solid contact stiffness between the rolling element and the raceway is solved based on Hertzian contact theory. The contact stiffness between a single rolling element and the raceway can be expressed as: In the formula, The contact stiffness between a single rolling element and the raceway. The contact deformation amount is based on the initial value of static assembly deformation and the relative deformation of the bearing position fed back from step S3. Sure; The Hertzian contact coefficient is related to the rolling element material and geometry; thus, the solid contact stiffness sub-matrix of the bearing is obtained. .
[0039] Next, the oil film stiffness is solved based on the elastohydrodynamic lubrication equations. Oil film stiffness The oil film stiffness is related to the oil film thickness h, lubricating oil viscosity η, and entrainment velocity u. It can be obtained by solving the Reynolds equation to get the oil film pressure distribution, and then differentiating it with respect to the oil film thickness to obtain the scalar value of the oil film stiffness at each rolling element position. The oil film stiffness at each rolling element position is assembled according to the bearing rolling element arrangement to obtain the oil film stiffness submatrix. .
[0040] By concatenating the solid contact scalar stiffness with the oil film scalar stiffness, a single bearing total equivalent stiffness submatrix is obtained. : , .
[0041] Finally, the total equivalent stiffness sub-matrix corresponding to each bearing is mapped to the global coordinate system of the shaft system through coordinate transformation according to the installation position of each bearing in the main shaft system, and the matrix assembly is completed to obtain the time-varying equivalent stiffness matrix K(Ψ(t)).
[0042] Calculation of the time-varying equivalent damping matrix C(Ψ(t)): Oil film damping is solved based on the theory of elastohydrodynamic lubrication. Oil film damping is related to the oil film thickness h, lubricating oil viscosity η, and contact load, with the damping contribution mainly originating from the oil film compression effect. By solving the perturbation form of the Reynolds equation, the dynamic damping coefficient of oil film pressure on compression velocity is obtained, and the scalar oil film damping at each rolling element position is calculated. The aforementioned parameters, such as oil film thickness and lubricating oil viscosity, are all taken from the bearing support state variable Ψ(t).
[0043] Within the same bearing, only the oil film damping corresponding to the rolling elements under load participates in the assembly, while the damping contribution of the unloaded rolling elements is set to zero. The scalar oil film damping at each rolling element position is mapped to the radial, axial, and overturning direction coordinates of the bearing according to its spatial position, and an oil film damping sub-matrix is assembled. Under conditions of sufficient lubrication and large oil film thickness, the solid Hertzian contact damping of the rolling element-raceway is very small and is ignored. The total equivalent damping of the bearing is approximately approximated by the oil film damping sub-matrix.
[0044] The total equivalent damping sub-matrix corresponding to each bearing is mapped to the global coordinate system of the shaft system through coordinate transformation according to the installation position of each bearing in the main shaft system, and the matrix assembly is completed to obtain the time-varying equivalent damping matrix C(Ψ(t)).
[0045] It should be noted that the above-mentioned equivalent stiffness and equivalent damping of the bearing are comprehensive equivalent parameters of the support system composed of the bearing solid contact pair and the lubricating oil film, and are not the inherent stiffness and damping of the bearing metal parts material.
[0046] (3) Coupling of the complete principal shaft system dynamic evolution model The cutting load vector F(t) and equivalent thermal load vector F output by the spindle body structure sub-model are... T The constant mass matrix M is coupled with the time-varying equivalent stiffness matrix K(Ψ(t)) and time-varying equivalent damping matrix C(Ψ(t)) output by the bearing support evolution sub-model to obtain the complete dynamic evolution model of the main shaft system, whose dynamic equations are:
[0047] In the formula, The axis system acceleration vector. Let x(t) be the shaft system velocity vector, x(t) be the shaft system displacement vector, Ψ(t) be the bearing support state variable, M be the constant mass matrix, C(Ψ(t)) be the time-varying equivalent damping matrix, K(Ψ(t)) be the time-varying equivalent stiffness matrix, and F(t) be the cutting load vector. T (t) is the equivalent thermal load vector.
[0048] S3. Solving the dynamic evolution model of the main shaft system and extracting displacements. The above dynamic equations are solved in the time domain. In this embodiment, the Newmark-β implicit integration method is used with a time step of 2ms to obtain the discrete nodal displacement vector x(t) of the principal shaft system.
[0049] Extract the bearing position node displacement from x(t), and calculate the relative deformation of the bearing position in conjunction with the fixed position of the bearing outer ring. This feedback is then fed back to the bearing support evolution sub-model. This feedback enables the bearing support characteristics to be updated in real time as the manufacturing process progresses, forming a two-way coupled iteration of the structural force field, temperature field, flow field, and dynamic field.
[0050] Simultaneously, extract the displacement of the front node of the principal axis from x(t). The data is then fed into the rotation accuracy calculation module.
[0051] At the initial time t0 of the iteration, the initial relative deformation of the bearing corresponding to the static assembly is calculated offline based on the spindle geometry parameters and bearing preload, and used as the initial input to the bearing support evolution sub-model. From the first discrete time step, the relative deformation of the bearing position is extracted from the solution results of the complete spindle system dynamic evolution model.
[0052] S4. Rotation accuracy calculation and threshold identification Based on the displacement of the front node of the spindle Calculate the current rotation error. The method for calculating the rotation error is as follows: Take the x-direction displacement of the front node of the spindle in a plane perpendicular to the axis. With y-direction displacement Calculate radial rotational runout (radial rotational error): .
[0053] Current slew error With preset threshold contrast: like Then the output rotation accuracy compensation amount ; like If the compensation amount is zero, no control command will be output.
[0054] Preset threshold The accuracy is set according to the machining requirements of the heavy-duty spindle, for example, 2.5μm. Through an on-demand compensation mechanism, the compensation amount is output only when the rotational error exceeds the preset threshold; when the error is within the threshold range, the compensation amount is set to zero and no adjustment is made.
[0055] S5, Precision Active Control The rotational accuracy compensation amount is output based on the threshold discrimination result. Based on this compensation amount, an active accuracy control command is generated to drive the spindle actuator for adjustment. The spindle actuator includes one or more of the following: a bearing preload adjustment mechanism, a cooling flow rate adjustment mechanism, and a CNC system coordinate compensation unit. Control methods include: adjusting bearing preload, adjusting cooling medium flow rate, and CNC coordinate axis error compensation. After the actuator adjusts the spindle support state, the bearing support state variable Ψ(t) is updated accordingly, which in turn updates the time-varying equivalent stiffness matrix K(Ψ(t)) and the time-varying equivalent damping matrix C(Ψ(t)), suppressing rotational error evolution and drift, and achieving active control of the spindle shaft system rotational accuracy during heavy-duty machining.
[0056] The quantitative correspondence between the actuator's motion and rotational error can be determined as follows: For bearing preload adjustment mechanisms, the relationship between the preload adjustment amount and the rotation error compensation amount can be calibrated through bearing stiffness tests to obtain the preload-rotation error coefficient. During actual adjustment, the required preload adjustment amount can be obtained by multiplying the rotation error compensation amount by this coefficient.
[0057] For a cooling flow regulating mechanism, the relationship between the cooling medium flow rate regulation amount and the bearing temperature change amount can be determined by the thermal balance equation. Then, by using the relationship between thermal deformation and rotation error, the corresponding relationship between the cooling medium flow rate regulation amount and the rotation error compensation amount can be obtained.
[0058] For the coordinate compensation unit of the CNC system, the coordinate compensation amount corresponds directly to the rotation error compensation amount. The tool tip offset caused by rotation is compensated by real-time correction of the tool tip coordinate. In actual control, the rotation error compensation amount is directly input into the CNC system as the coordinate compensation amount.
[0059] The above three control methods can be used individually or in combination. After the actuator adjusts the spindle support state, the bearing support state variable Ψ(t) is updated accordingly, which in turn updates the time-varying equivalent stiffness matrix K(Ψ(t)) and the time-varying equivalent damping matrix C(Ψ(t)), forming a closed loop of "modeling-solving-control".
[0060] Example 2:
[0061] This embodiment illustrates the iterative timing and closed-loop update process of the bearing support evolution sub-model.
[0062] At the initial moment t0 of the iteration start, the initial relative deformation of the bearing corresponding to the static assembly is calculated offline based on the spindle assembly geometry parameters and bearing preload, and used as the initial input of the bearing support evolution sub-model.
[0063] Starting from the first discrete time step, the following process is executed at each time step: (1) The bearing support evolution sub-model is based on the relative deformation of the current bearing position. The spindle speed, bearing temperature, lubrication medium condition, and initial static assembly deformation values are used to update the bearing multiphysics state Ψ(t). (2) Calculate the time-varying equivalent stiffness matrix K(Ψ(t)) and the time-varying equivalent damping matrix C(Ψ(t)) based on the updated Ψ(t); (3) Combine K(Ψ(t)) and C(Ψ(t)) with the constant mass matrix M, load vector F(t) and F T (t) coupling, solve the dynamic equation, and obtain the node displacement vector x(t) at the current time step; (4) Extract the bearing position node displacement from x(t), and calculate the relative deformation of the bearing position in combination with the fixed position of the bearing outer ring. This is fed back to the bearing support evolution sub-model for Ψ(t) update in the next time step.
[0064] In terms of iteration timing, the bearing deformation in the current step is used to update K(Ψ(t)) and C(Ψ(t)) in the next time step, resulting in a single-step time lag. This embodiment focuses on the slow-varying drift caused by heat during long-term cutting, using a time step size of 2ms. The impact of millisecond-level lag is minimal and fully meets the control requirements of heavy-duty machine tool processing. If higher computational accuracy is required, sub-iterations can be added within each time step to repeatedly update K(Ψ(t)) and C(Ψ(t)) until convergence eliminates the lag.
[0065] Example 3:
[0066] This embodiment provides a specific application example to illustrate the implementation process of the method of the present invention.
[0067] Taking a certain type of heavy-duty spindle as an example, the spindle speed range is 10~500 rpm, the maximum radial component of the cutting force is approximately 20 kN, the maximum axial component is approximately 10 kN, and the maximum cutting torque is approximately 5 kN·m. The bearing is an angular contact ball bearing with a rolling element diameter of approximately 20 mm, a pitch circle diameter of approximately 150 mm, and a contact angle of approximately 40°. The time step is 2 ms. A preset rotational error threshold is used. Take 2.5μm.
[0068] At the initial moment of iteration, the initial relative deformation of the bearing is calculated offline based on the spindle geometry parameters and bearing preload, for example, 5μm, as the initial input of the bearing support evolution sub-model.
[0069] During the machining process, when the calculated rotational error e(t) exceeds 2.5μm, the output compensation amount is... = e(t) - 2.5μm, and generate control commands to drive the actuator. When e(t) does not exceed 2.5μm, the compensation amount is set to zero, and no control commands are output.
[0070] Through the above process, the present invention can solve the multi-physics coupled dynamic state of the heavy-duty spindle system online in real time, and achieve active control of accuracy based on real-time rotation error, thereby suppressing the drift of spindle rotation accuracy under heavy-duty long-term cutting.
[0071] Finally, it should be noted that any parts of this invention not described in detail are prior art. Those skilled in the art will understand that the above descriptions are merely preferred embodiments of the invention and are not intended to limit the invention. Although the invention has been described in detail with reference to the foregoing examples, those skilled in the art can still modify the technical solutions described in the foregoing examples or make equivalent substitutions for some of the technical features. All modifications and equivalent substitutions made within the spirit and principles of the invention should be included within the scope of protection of the invention.
Claims
1. A method for multi-physics coupled dynamic evolution modeling and accuracy control of heavy-load spindles, characterized in that, Includes the following steps: S1. Real-time operating condition data acquisition: Acquire real-time operating condition data during the service of heavy-duty spindles. The real-time operating condition data includes cutting force, spindle speed, bearing temperature, spindle body temperature, and lubrication medium status. S2. Construct a multi-physics coupled dynamic evolution model for the heavy-duty spindle system: decompose the heavy-duty spindle system into a spindle body structure sub-model and a bearing support evolution sub-model; The spindle body structure sub-model receives the collected cutting force and spindle body temperature, maps the cutting force to generate a cutting load vector F(t), and calculates and maps the equivalent thermal load vector F from the spindle body temperature. T (t), a constant mass matrix M is pre-constructed; The bearing support evolution sub-model receives spindle speed, bearing temperature, lubrication medium condition, and the relative deformation of bearing position fed back from step S3. The time-varying equivalent stiffness matrix K(Ψ(t)) and the time-varying equivalent damping matrix C(Ψ(t)) are calculated and output; where Ψ(t) is the bearing support state variable that changes with time; the two sub-models are coupled to obtain the complete main shaft system dynamic evolution model; S3. Solving the dynamic evolution model of the main spindle system and extracting displacements: Solve the complete dynamic evolution model of the main spindle system to obtain the discrete node displacement vector x(t) of the main spindle system. Extract the bearing position node displacements from x(t), and calculate the relative deformation of the bearing position based on the bearing position node displacements and the fixed position of the bearing outer ring. ,Will Feedback is sent to the bearing support evolution sub-model; S4. Rotation accuracy calculation and threshold identification: Based on the displacement of the spindle front node extracted from x(t). Calculate the current rotation error and compare it with a preset threshold. If the rotation error exceeds the preset threshold, output the rotation accuracy compensation amount; otherwise, set the compensation amount to zero. S5. Precision Active Control: Generate a precision active control command based on the rotational precision compensation amount to drive the spindle actuator to perform rotational precision active control.
2. The method for multi-physics coupled dynamic evolution modeling and accuracy control of heavy-duty spindles according to claim 1, characterized in that, In step S2, the cutting load vector F(t) and equivalent thermal load vector F output by the spindle body structure sub-model are... T The constant mass matrix M is coupled with the time-varying equivalent stiffness matrix K(Ψ(t)) and time-varying equivalent damping matrix C(Ψ(t)) output by the bearing support evolution sub-model to obtain the complete dynamic evolution model of the main shaft system. Its dynamic equations are: In the formula, Axis acceleration vector The shaft system velocity vector, x(t) is the shaft system displacement vector, Ψ(t) is the bearing support state variable, M is the constant mass matrix, C(Ψ(t)) is the time-varying equivalent damping matrix, K(Ψ(t)) is the time-varying equivalent stiffness matrix, F(t) is the cutting load vector, and F... T (t) is the equivalent thermal load vector.
3. The method for multi-physics coupled dynamic evolution modeling and accuracy control of heavy-duty spindles according to claim 1, characterized in that, The spindle body structure sub-model incorporates spindle geometric parameters and material fixed parameters. The constant mass matrix M is determined based on the spindle geometric parameters and material fixed parameters and does not change with the real-time operating data.
4. The method for multi-physics coupled dynamic evolution modeling and accuracy control of heavy-duty spindles according to claim 1, characterized in that, In step S3, at the initial moment of iteration start, the initial relative deformation of the bearing corresponding to the static assembly is calculated offline based on the spindle geometry parameters and bearing preload, and used as the initial input of the bearing support evolution sub-model; Starting from the first discrete time step, the relative deformation of the bearing position is extracted from the solution results of the complete spindle system dynamic evolution model.
5. The method for multi-physics coupled dynamic evolution modeling and accuracy control of heavy-duty spindles according to claim 1, characterized in that, The bearing support evolution sub-model is based on Hertzian contact theory to solve for the solid contact stiffness of the rolling element-raceway, and on the elastohydrodynamic lubrication equations to solve for the oil film stiffness and oil film damping. The solid contact stiffness and oil film stiffness are connected in series to obtain the total equivalent stiffness of the bearing.
6. The method for multi-physics coupled dynamic evolution modeling and accuracy control of heavy-duty spindles according to claim 1, characterized in that, The time-varying equivalent stiffness matrix K(Ψ(t)) and the time-varying equivalent damping matrix C(Ψ(t)) are the comprehensive equivalent matrix parameters of the support system composed of the bearing solid contact pair and the lubricating oil film, and are not the inherent stiffness and damping of the bearing metal parts.
7. The method for multi-physics coupled dynamic evolution modeling and accuracy control of heavy-duty spindles according to claim 1, characterized in that, The bearing support evolution sub-model incorporates bearing geometric parameters, fixed bearing material parameters, and initial values of static assembly deformation. These initial values of static assembly deformation serve as one of the bases for calculating the time-varying equivalent stiffness matrix K(Ψ(t)) and the time-varying equivalent damping matrix C(Ψ(t)).
8. The method for multi-physics coupled dynamic evolution modeling and accuracy control of heavy-duty spindles according to claim 1, characterized in that, The bearing support state variable Ψ(t) characterizes the evolution of bearing preload drift and lubricant state under the coupling effect of structural force field, temperature field and flow field. The bearing support state variable Ψ(t) includes multiple factors such as bearing contact load, oil film thickness, oil film pressure, bearing preload, lubricant viscosity and thermal expansion caused by bearing temperature.
9. The method for multi-physics coupled dynamic evolution modeling and accuracy control of heavy-duty spindles according to claim 1, characterized in that, The bearing support state variable Ψ(t) is obtained by coupling the structural force field, temperature field, and flow field. The structural force field characterizes the bearing contact pressure and shaft elastic deformation under cutting load; the temperature field characterizes the temperature distribution and thermal expansion deformation caused by bearing frictional heat generation and motor loss heat generation; and the flow field characterizes the flow characteristics of the bearing lubricating oil film and the heat transfer characteristics of the cooling medium. The time-varying equivalent stiffness matrix K(Ψ(t)) and the time-varying equivalent damping matrix C(Ψ(t)) are input into the complete main shaft system dynamic evolution model to solve for the shaft system vibration and dynamic rotation response.
10. The method for multi-physics coupled dynamic evolution modeling and accuracy control of heavy-duty spindles according to claim 1, characterized in that, The spindle actuator includes one or more of the following: a bearing preload adjustment mechanism, a cooling flow adjustment mechanism, and a CNC system coordinate compensation unit; The control methods include: adjusting the bearing preload, adjusting the cooling medium flow rate, and compensating for CNC coordinate axis errors.
