CN114580971A

CN114580971A - Dynamic adjustment method, system and equipment based on performance digital twinning

Info

Publication number: CN114580971A
Application number: CN202210407586.8A
Authority: CN
Inventors: 张萌; 杜日煌; 李国喜; 王东; 赵存
Original assignee: National University of Defense Technology
Current assignee: National University of Defense Technology
Priority date: 2022-01-23
Filing date: 2022-04-19
Publication date: 2022-06-03

Abstract

The invention provides a dynamic adjustment method, a system and equipment based on performance digital twinning, wherein the method comprises the following steps: monitoring and collecting multi-source heterogeneous data in the assembly process; preprocessing real-time data and historical data of the multi-source heterogeneous data, extracting dynamic characteristic data corresponding to the real-time data and the historical data, performing characteristic fitting, comparing a real-time data fitting curve with a historical data fitting curve, and judging whether the data is abnormal or not; establishing a digital twin model in the assembly process to obtain simulation prediction data; analyzing the quality parameters of the products in process based on the simulation prediction data to determine the abnormal reason; and performing dynamic optimization regulation and control based on the abnormal reasons. The scheme improves the comprehensiveness of quality control in the assembly process, can predict the variation trend of the assembly quality characteristics, and reduces the blindness and subjectivity of the adjustment strategy in the assembly process.

Description

Dynamic adjusting method, system and equipment based on performance digital twins

Technical Field

The invention relates to the field of digital twins and intelligent manufacturing, in particular to a dynamic adjustment method, a dynamic adjustment system and dynamic adjustment equipment for the digital twins of a two-axis two-frame servo system structure.

Background

The complex product refers to a product with complex requirements, complex composition, complex technology, complex manufacturing process and complex production management, such as an airplane and the like. The development period of complex products is long, and the trial production cost is high. Quality is an inherent property of complex equipment and plays a very important role in its performance. Assembly is one of the most important links in the manufacturing process of complex products, the result of which has a direct impact on the quality, performance, lifetime and maintainability of the product. During the manufacturing process, the assembly time accounts for about 20-50% of the total production time, and the assembly cost accounts for 20-30% of the total manufacturing cost. In recent years, with the gradual maturity of the technologies of internet of things, big data and cloud computing, the assembly production of modern complex mechanical products is promoted to be developed towards digitalization, networking and intellectualization. The method promotes the deep fusion of physical information fusion, digital twin technology and manufacturing industry. The digital twin is a potential effective way for realizing intelligent interconnection and interactive fusion of a physical domain and an information domain in the equipment assembly process, and is recently highly concerned by experts at home and abroad.

The two-axis two-frame servo structure is used as a complex assembly structure and widely applied to optoelectronic system stabilization platforms such as a guide head, and the like, and the load is usually components such as an optical device or a gyroscope and the like, so that the precision is high and the vibration and other external interference are sensitive, and therefore, the two-axis two-frame servo structure has high requirements on the dynamic indexes of the two-axis two-frame servo structure. The assembly is the final link for realizing the functions of the two-shaft two-frame structure and is the key for achieving the expected dynamic performance of the whole machine. Taking a certain type of guide head as an example, the processing and manufacturing grades of parts of the guide head are already high, but the standard reaching rate of product performance still depends on assembly, the assembly mode of a servo mechanism is still manual assembly, and in some links, only experienced technicians assemble the guide head by hand feeling, so that the problems of low one-time assembly qualification rate, poor stability and reliability and high production cost and cycle exist in the assembly process.

The patent publications and literature documents at the present stage show that a scholars predict and control the quality of an intelligent assembly shop by means of digital twin, however, the existing quality prediction method mainly analyzes quality data by means of large data, the prediction mode is still positioned on the control of geometric elements such as assembly tolerance and the like, and quality characteristics generally have multidisciplinary attributes and are characterized in the form of physical quantities. Still other scholars use the manhattan algorithm to calculate the relative mass dispersion to obtain better assembly process parameters. However, when the assembly process becomes complicated, the application of the manhattan algorithm becomes very limited, and the processing capability is not sufficiently expressed for the non-linear problem existing in the assembly quality characteristic. In addition, in the existing assembly and adjustment method, the assembly process is a coordinate system based on the structure size, the shape and the position precision, and the static geometric parameters are used as assembly control input, so that good static precision can be obtained, but effective means for assembly simulation analysis, assembly performance prediction and assembly quality guarantee are lacked in the aspects of dynamic physical indexes related to vibration characteristics, such as frequency, damping, pretightening force and dynamic stiffness of a servo mechanism.

In summary, although the existing research results and methods can improve the assembly precision to a certain extent and realize the optimized management of complex product assembly workshops, the problems that the nonlinear problem processing capability is not enough, the quality characteristics only focus on the traditional geometric elements and the like exist, the quality characteristic parameter prediction results obtained by the system are difficult to adapt to complex discrete product assembly, and the given adjustment strategy cannot meet the requirement that the assembly quality of the whole product is stable and controllable.

Disclosure of Invention

In view of the above, the present invention provides a dynamic tuning method, system and device based on performance digital twin, so as to make up for the deficiencies in the prior art. Specifically, the invention provides the following technical scheme:

in one aspect, the invention provides a dynamic tuning method based on performance digital twinning, which comprises the following steps:

step 1, monitoring and collecting multi-source heterogeneous data in an assembly process, wherein the multi-source heterogeneous data comprises parameter data, dynamic information data and environmental data;

step 2, preprocessing the real-time data and the historical data of the multi-source heterogeneous data, extracting dynamic characteristic data corresponding to the real-time data and the historical data, performing characteristic fitting, comparing a real-time data fitting curve with a historical data fitting curve, and judging whether the data is abnormal or not;

step 3, establishing a digital twin model in the assembly process, wherein the digital twin model comprises a visual model, a dynamic simulation model and an assembly behavior model, and predicting the mode of the dynamic simulation model to obtain simulation prediction data;

step 4, analyzing the quality parameters of the products under production based on the simulation prediction data, and if the quality parameters are abnormal, determining abnormal reasons corresponding to the abnormal quality parameters;

and 5, carrying out dynamic optimization regulation and control based on the abnormal reasons.

Preferably, in step 1, the parameter data includes a pre-assembly parameter, a post-assembly parameter, and a robot arm parameter.

Preferably, in the step 2, the determining whether the data is abnormal includes: and fitting the change curve of the historical data in a time period, fitting based on the change curve of the collected real-time data in the time period with the same length, comparing the distribution peaks of the two fitting curves, and judging that the data are abnormal when the distribution peak of the real-time data fitting curve is deviated to the left or right relative to the distribution peak of the historical data fitting curve, or the peak value of the peak is too high or too low.

Preferably, in the step 3, the predicting of the modality further includes:

301, numbering parts and connecting pairs of equipment to be assembled to form a discrete complete machine model consisting of a plurality of equivalent rigid bodies;

step 302, discretizing a discrete whole machine model into a linear system with limited freedom, and calculating a rigidity matrix K and a damping matrix C based on the acquired dynamic characteristic data;

step 303, solving a vibration mode equation of the linear system based on the stiffness matrix K and the damping matrix C to obtain each order of natural frequency and a corresponding vibration mode of the linear system, wherein each order of natural frequency and corresponding vibration mode are simulation prediction data.

Preferably, the stiffness matrix K is:

wherein k is_a，k_r，k_r' respectively comprises axial rigidity, main radial rigidity and secondary radial rigidity of two angular contact ball bearings, k_α，k_β，k_γThe yaw stiffness of the inner frame system rotating around the x axis, the azimuth stiffness of the y axis and the pitch stiffness of the inner frame system rotating around the z axis are respectively.

Preferably, the calculation of each stiffness is:

in the above formula, k_θFor conversion of the torsional stiffness of angular contact ball bearings to the linear stiffness on the gear mesh line, /)₁、l₂Respectively the distance between the contact point of the two half shafts and the bearing support part and the mass center of the inner frame structure, K_xIs the linear contact stiffness of the torsion spring.

Preferably, the mode equation is:

and, | [ K ]]-ω²[M]|＝0

Wherein ω is the natural frequency of the structural vibration;

is the natural vibration mode of the system, [ M ]]Is the quality matrix of the system.

Preferably, in the step 3, the assembly behavior model refers to a time sequence model of an assembly process, and in this embodiment, the time sequence model of the whole assembly process may be first constructed;

after receiving all input information of a node in the assembly behavior model, predicting through the mode of the dynamic simulation model to serve as the output of the node;

and displaying the output of each node through a visualization model.

Preferably, in the step 4, the abnormality cause corresponding to the quality parameter abnormality is determined by:

step 401, determining abnormal reasons causing various abnormalities;

step 402, establishing a fuzzy relation matrix between various anomalies and anomaly reasons in a time period;

step 403, calculating the membership degree of the fuzzy abnormal mode based on the fuzzy relation matrix;

step 404, establishing a fuzzy relation equation between the abnormality Y and the abnormality reason X:

Y＝X°R

wherein,

for each component y in the sample_iFor abnormal Y_iThe degree of membership of (a) is,

the membership degree of X to various abnormal reasons and a min-max fuzzy synthesis operator;

and 405, solving the fuzzy relation equation based on the membership degree of the fuzzy abnormal mode in the step 403 to obtain the optimal solution of the fuzzy equation so as to determine the abnormal reason which has the maximum contribution to the abnormality.

In addition, the invention also provides a dynamic adjusting system based on the performance digital twin, which comprises: the system comprises a quality data acquisition module, a data processing and mapping module, a digital twin module and an intelligent management and control module;

the quality data acquisition module acquires multi-source heterogeneous data in the assembly process in real time in a physical entity, classifies and stores the acquired multi-source heterogeneous data, and transmits the preprocessed multi-source heterogeneous data to the digital twin module for digital twin analog simulation of assembly products;

the data processing and mapping module comprises a data preprocessing unit, a real-time quality parameter database unit and a communication unit; the data preprocessing unit is used for preprocessing the multi-source heterogeneous data to obtain preprocessed data; the real-time quality parameter database unit is used for storing normal quality parameter data and abnormal quality parameter data in the assembly process; the communication unit is used for mapping the preprocessed data to a digital twinning model and taking the mapped data as the input of the analysis and calculation of the digital twinning module;

the digital twinning module is used for establishing a digital twinning model, the digital twinning model comprises a visual model, a dynamic simulation model and an assembly behavior model, and the mode of the dynamic simulation model is predicted to obtain simulation prediction data;

and the intelligent control module is used for analyzing the quality parameters of the products under production based on the simulation prediction data, and determining the abnormal reason corresponding to the abnormal quality parameters if the quality parameters are abnormal.

Preferably, the system further comprises a physical entity module, wherein the physical entity module comprises a body unit, a measuring equipment unit, a detecting equipment unit and a signal transmission equipment unit in the assembly workshop;

the body unit refers to assembly equipment, materials and products in process used for product assembly; the measuring device unit comprises means for measuring the work in process; the detection equipment unit is a device for detecting the state of the assembly equipment; the signal transmission equipment unit refers to an equipment interface and a network for data transmission and exchange;

the related information of the physical entity of the assembly workshop is obtained through the physical entity module, and interactive feedback is realized through the dynamic link and the digital twin module.

Preferably, in the intelligent management and control module, the judgment method for the quality parameter abnormality is as follows:

and fitting the change curve of the historical data in a time period, fitting based on the change curve of the collected real-time data in the time period with the same length, comparing the distribution peaks of the two fitting curves, and judging that the data are abnormal when the distribution peak of the real-time data fitting curve is deviated to the left or right relative to the distribution peak of the historical data fitting curve, or the peak value of the peak is too high or too low.

Preferably, in the digital twin module, the mode prediction mode is as follows:

numbering parts and connecting pairs of equipment to be assembled to form a discrete complete machine model consisting of a plurality of equivalent rigid bodies;

discretizing a discrete whole machine model into a linear system with limited degree of freedom, and calculating a rigidity matrix K and a damping matrix C based on acquired dynamic characteristic data;

and solving a vibration mode equation of the linear system based on the rigidity matrix K and the damping matrix C to obtain each order of natural frequency and a corresponding vibration mode of the linear system, wherein each order of natural frequency and the corresponding vibration mode are simulation prediction data.

In yet another aspect, the present invention also provides a performance digital twin based dynamic tuning device comprising a processor, a storage device storing instructions readable by the processor; the processor is configured to invoke instructions in the memory device to perform the performance digital twin based dynamic tuning method as described above.

Compared with the prior art, the technical scheme of the invention at least has the following advantages:

firstly, the invention improves the comprehensiveness of quality control in the assembly process and realizes the description and characterization of all elements and the whole flow thereof in the assembly process by depicting and modeling the assembly process of the complex product from different levels and different dimensions. Secondly, the invention carries out virtual-real interaction with a physical assembly workshop through a digital twin model, can carry out digital twin model simulation on the basis of real-time quality data acquired in an assembly field, further predicts the variation trend of the assembly quality characteristics, and improves the initiative of quality control. Thirdly, the active regulation and control strategy in the assembly production process is generated by an intelligent decision optimization method and fed back to the physical entity module in real time, so that the blindness and subjectivity of the adjustment strategy in the assembly process are reduced.

Additional advantages, objects, and features of the invention will be set forth in part in the description which follows and in part will become apparent to those having ordinary skill in the art upon examination of the following. The objectives and other advantages of the invention may be realized and attained by the means of the instrumentalities and combinations particularly pointed out hereinafter.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.

FIG. 1 is a schematic view of a digital twin control system according to an embodiment of the present invention;

FIG. 2 is a flow chart of a digital twin management and control system according to an embodiment of the present invention;

FIG. 3 is a schematic diagram of a digital twinning construction process according to an embodiment of the present invention;

FIG. 4 is an exemplary diagram of an assembly scenario model according to an embodiment of the present invention;

FIG. 5 is an exemplary illustration of an assembly analysis model according to an embodiment of the present invention;

FIG. 6 is an exemplary diagram of a servo model according to an embodiment of the present invention;

FIG. 7 is a schematic diagram of a flexible connection relationship of a servo mechanism according to an embodiment of the present invention.

Detailed Description

Embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the described embodiments are only a few embodiments of the invention, and not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

It will be appreciated by those of skill in the art that the following specific examples or embodiments are a series of presently preferred arrangements of the invention to further explain the principles of the invention, and that such arrangements may be used in conjunction or association with one another, unless it is expressly stated that some or all of the specific examples or embodiments are not in association or association with other examples or embodiments. Meanwhile, the following specific examples or embodiments are only provided as an optimized arrangement mode and are not to be understood as limiting the protection scope of the present invention.

Referring to fig. 1 to 5, an embodiment of the present invention is described by taking the quality control of a certain type of equipment as an example. As shown in fig. 1, the digital twin dynamic tuning system of this embodiment mainly includes a physical entity module, a quality data acquisition module, a data processing and mapping module, a digital twin module, and an intelligent management and control module.

The quality data acquisition module acquires information data of an assembly process in a physical entity in real time, and after the data is processed, the data is transmitted to the digital twinning module to be assembled in a product digital twinning analog simulation; the assembly process information acquired by the quality data acquisition module comprises the starting and stopping states of assembly stations, the number of joints of the mechanical arm, the starting and stopping states of each joint, the torque of each joint and the rotating speed in the assembly process; the state of the mechanical arm speed reducer, the starting and stopping state of the motor and the electrical load; the position of the mechanical arm, the state of an encoder, the voltage of a motor and the like; the starting and stopping states of a rack temperature, a rack vibration signal, a conveying belt deviation amount and a deviation alarm of the conveying belt are set; the number, size and rotating speed of carrier rollers of the conveyor belt; starting and stopping states, rotating speed and voltage of the motor; the size of an assembly part, the staggered tooth number of an anti-backlash gear of the assembly part, a pre-tightening gap of a bearing, an initial contact angle of the bearing, the rigidity of the anti-backlash torsion spring, a pressure angle of the gear and the like; the acquired data is accurately mapped to the digital twin model in the digital twin module after being processed by the data processing and mapping module.

The quality data acquisition module is used for acquiring quality parameters in the assembly production process, at least comprises all elements of the production processes such as equipment, materials, processes, environment and the like, and classifying and storing the acquired data to provide data support for subsequent information analysis and processing; the data collected includes: 1. assembling quality parameters, wherein the assembling quality parameters comprise position parameters and size parameters of the cabin section butt joint hole, an initial contact angle of a rolling bearing, a bearing pre-tightening gap, spring stiffness of the anti-backlash gear and the like; 2. quality characteristic parameters of each component, such as a bearing pre-tightening clearance, a bearing contact angle, spring stiffness and the like of a servo frame of the seeker; 3. dynamic assembly information of the product to be assembled, including operation state data and the like in the assembly process; 4. and the assembly environment information comprises vibration information, noise information and the like in an assembly workshop. The collected data are from hardware collection obtained by adopting a sensor and software collection obtained by carrying out software secondary development by utilizing an interface protocol, and the collected data comprise existing static data in the assembly process and dynamic data in the assembly process obtained by depending on the sensing equipment.

And analyzing in an intelligent management and control module, and providing a dynamic adjustment strategy for the generated abnormal state. The digital twin module is in a core position and performs data interaction feedback with other modules to realize information exchange and closed-loop optimization.

The physical entity module comprises assembly equipment, materials, a body unit of a product in process, a measuring equipment unit, a detecting equipment unit and a signal transmission equipment unit in an assembly workshop; the body unit of the assembly equipment, the material and the product in process is the assembly equipment, the material and the product in process used for product assembly, and comprises a mechanical structure body of the assembly equipment, the assembly material and the product in process, an electric and hydraulic device for driving the assembly equipment to move and act, and a corresponding electromechanical hydraulic control system; the measuring device unit comprises means for measuring the dimensions, physical properties, etc. of the work in process; the detection equipment unit is a device for detecting the running state, the system moving path, the system working efficiency and other states of the assembly equipment; the signal transmission equipment unit refers to equipment interfaces and networks for data transmission and exchange, and the signal transmission supports an OPC-UA transmission protocol. The related information of the physical entity of the assembly workshop is obtained through the physical entity module, and interactive feedback is realized through the dynamic link and the digital twin module.

The data processing and mapping module comprises a data preprocessing unit, a real-time quality parameter database unit and a communication unit; the data preprocessing unit is used for data cleaning fusion and data normalization processing, and removing data noise and interference brought in the data acquisition process through the data cleaning fusion to obtain useful signals; then, standardizing and discretizing the data through data normalization processing to form an intelligent decision data system; the real-time quality parameter database unit is used for storing normal quality parameter data and abnormal quality parameter data in the assembly process, so that subsequent data comparison and data mining are facilitated; the communication unit is used for storing the data after the fusion and normalization processing, and mapping the processed data to a digital twin model through a communication protocol to be used as an input condition for analysis and calculation of the digital twin module.

The digital twinning module comprises a digital twinning model unit, the digital twinning model unit is used for constructing an entity model of an assembly workshop and a dynamic virtual model in the assembly process, the assembly process is described from different levels and different dimensions according to the assembly process of a complex product, model fusion processing is carried out on simulation models in different subject fields such as the field of kinematics subject, the field of dynamics subject and the like, different types and different scale scales, a more comprehensive and more complex digital twinning model is established, and the precision of the model is improved; the digital twin module is in a core position, provides samples for the intelligent control module through data interaction feedback, and provides guidance for system operation optimization.

The intelligent management and control module comprises a state judgment unit and an intelligent decision unit; the state judgment unit is used for judging whether the quality parameter calculated by the digital twin module is in a normal state or an abnormal state or a dangerous state needing emergency treatment, and providing a basis for intelligent decision; the intelligent decision unit comprehensively utilizes an intelligent analysis processing algorithm to carry out deep analysis on the expression form of the current abnormal condition and the formation reason of the abnormal condition.

As shown in fig. 2, the technical solution of the present invention may further provide a dynamic tuning method based on performance digital twinning, and in this embodiment, a servo mechanism in workshop assembly is taken as an example to describe a specific solution and an implementation manner of the present invention. In particular, the method may comprise the steps of:

1) and (3) starting a system: initializing software and hardware of each module in the system;

2) and (3) system monitoring operation: monitoring the starting and stopping states of an assembly station, the number of joints of a mechanical arm, the starting and stopping states of each joint, the torque and the rotating speed of each joint, the staggered tooth number of a mechanical anti-backlash gear, a pre-tightening gap of a bearing, an initial contact angle of the bearing, the rigidity of an anti-backlash torsion spring, a gear pressure angle and the like in the assembly process of an assembly workshop in real time;

3) multi-source heterogeneous quality data acquisition: collecting, storing and transmitting multi-source heterogeneous information in the complex equipment assembling process; further preferably, the multi-source heterogeneous information may include: status data from different devices, data from different parts such as size, surface roughness and bolt pre-tightening force; dynamic assembly information, such as start-stop states of assembly stations, measurement data of mechanical arms and conveyor belts, and the like; and collected data of external disturbances such as temperature, noise, vibration, etc. of the environment.

Parameters such as mechanical arm and equipment assembly in a workshop are taken as examples, and the steps can be detailed as follows:

collecting mechanical arm parameters, wherein the mechanical arm parameters comprise the number of joints of a mechanical arm, the starting and stopping states of each joint, the torque of each joint, the rotating speed and the like;

collecting parameters before equipment assembly, including an initial contact angle of a bearing, the rigidity of an anti-backlash torsion spring, a gear pressure angle and the like;

collecting parameters of the assembled equipment, including the number of staggered teeth of the mechanical anti-backlash gear, the pre-tightening clearance of the bearing and the like;

collecting dynamic information data including data such as an assembling process, whether assembling is carried out or not and the like;

environmental data is collected, including ambient temperature, noise, vibration, etc.

4) Data preprocessing and distinguishing: the method comprises the steps of cleaning, fusing, normalizing and other preprocessing the real-time data and the historical data, then extracting dynamic characteristic data, comparing a fitting curve of the real-time data with a fitting curve of the historical data through data fitting, judging whether the acquired data are abnormal or not and whether quality characteristic parameters of products in process are abnormal or not according to the comparison, and providing conditions for subsequent quality data prediction and intelligent decision.

During data cleaning, data beyond a reasonable data range are removed; in the embodiment, normalization processing maliciously adopts a min-max standardization method, for example, to normalize the quality data of different orders of magnitude in a reasonable interval, so that subsequent calculation and use are facilitated.

Continuing to take the assembly shop mechanical arm and the like in the embodiment as an example, cleaning the parameters before equipment assembly, the parameters after equipment assembly, the parameters before historical equipment assembly and the parameters after historical equipment assembly, wherein the cleaning comprises removing repeated data and data which exceed a specified range;

after cleaning, the data are normalized, and quality data with different orders of magnitude can be normalized to a reasonable interval by adopting a min-max standardization method for example;

fitting the normalized equipment parameter data and the normalized historical parameter data to obtain a data fitting curve and a historical data fitting curve;

and comparing the data fitting curve with the historical data fitting curve, and determining whether the acquired parameters before equipment assembly and the acquired parameters after equipment assembly are abnormal. More specifically, change curves of parameters before historical equipment assembly and parameters after historical equipment assembly are fitted, after the change curves exist, the change trends of the parameters before the equipment assembly and the parameters after the equipment assembly are fitted, if the distribution peaks of the parameters before the equipment assembly and the parameters after the equipment assembly are deviated to the left or the right relative to the historical change curves, and the peak values are too high or too low relative to the historical change curves, the equipment is abnormal, and when the equipment is abnormal, the part is disassembled and reassembled.

5) Digital twin analog simulation: and establishing a digital twinning model in the assembly process, performing digital twinning simulation by using the preprocessed data as an input driving model, predicting modal data of the established model, acquiring simulation data, and dynamically displaying to realize high-fidelity simulation in the assembly process.

And (3) predicting modal data, namely acquiring the number of staggered teeth of the anti-backlash gear of the assembly body, the axial pre-tightening gap of the bearing, the initial contact angle, the spring rigidity and the gear pressure angle as constraint conditions and excitation for establishing a simulation model, and performing simulation calculation to obtain the modal data of the assembly mechanism.

For the prediction of modal data, and as shown in fig. 6, the pilot servo is taken as an example, and the dynamic calculation performed in this embodiment is intended to calculate the natural frequencies of the respective stages of the pilot servo under different assembly parameters. In specific calculation, various collected parameters are reflected to a rigidity matrix in a dynamic equation. As shown in fig. 6, the seeker servo mechanism can be regarded as an independent whole, and is divided according to the functional structure, each independent part is numbered one by one, and small, scattered and miscellaneous components such as bolts and pins are omitted, as shown in table 1; the adjacent parts are often interacted by some kind of coupling pair, and the parts with contact or fit relation are also numbered one by one and classified as shown in table 2.

TABLE 1

Part number	Name of part	Part number	Part nameBalance
				1	Left half axle	8	Right half axle
2	Left flange	9	Motor stator
				3	Outer frame	10	Motor rotor
4	Left bearing	11	Pinion gear
				5	Inner frame	12	Big gear
6	Right bearing	13	Ground (earth)
				7	Right flange

TABLE 2

According to the hypothesis principle, each part listed in the system and the connection relation among the parts are combined, simplified and numbered to facilitate modeling as a starting point and to take main factors as a principle, the assembly body is subjected to grid division, as shown in table 3, a discrete whole machine model consisting of a plurality of equivalent rigid bodies is formed, the rigid bodies are mutually connected through flexible joint surfaces which are in contact with each other to form flexible nodes, the nodes can store and consume energy and can generate micro deformation, and the deformation coordination principle and physical geometric constraint of the nodes are met in the vibration process, so that the structures consisting of the rigid bodies and the nodes form independent units in the whole machine system, the units are used as a subsystem of the system, and the rigidity and damping matrix of the units can be independently described.

TABLE 3

The rigid bodies are flexibly connected with each other through the nodes to form basic units of the system, the units have independent dynamic characteristics and can be subdivided into bolt connection units, gear transmission units and bearing support units, and mapping relations between assembly process parameters and dynamic parameters of unit joint surfaces are respectively established according to different types of the units, as shown in a table 4. The bolt connection unit mainly obtains the rigidity and damping parameters of a joint surface according to the pretightening force or pretightening torque of the bolt; the gear transmission unit mainly obtains tooth surface contact rigidity and damping parameters according to the assembly center distance, the transmission torque and the staggered tooth number of the anti-backlash gear; the bearing supporting unit mainly obtains axial and radial rigidity and damping parameters according to the axial pre-tightening gap and the radial pre-tightening gap. According to the specific connection type of the unit, the detailed connection relation between equivalent rigid bodies can be constructed as shown in figure 7.

TABLE 4

The modal analysis is the basis for performing dynamic analysis such as random vibration, harmonic response, response spectrum and the like, and the natural frequency and the mode shape of each order of the structure can be obtained through the modal analysis. The structure of the seeker is an elastic body with infinite multiple degrees of freedom, in this embodiment, the seeker is discretized into a linear system with finite degrees of freedom, and the differential equation of motion of the system under the action of an external load is as follows:

in the formula: [ M ] A]、[C]、[K]Respectively sequentially providing a mass matrix, a damping matrix and a rigidity matrix of the system; { x (t) },

{ F (t) } are, in turn, the displacement column vector, velocity column vector, acceleration column vector, and the external excitations acting on each discrete particle of the system, respectively.

The inner frame, the two bearing outer rings and the anti-backlash gear are taken as a whole and are called an inner frame system, the two half shafts tightly push the bearing through adjusting the thickness of the gasket to play a role in supporting the inner frame system, and the driving gear transmits power to the anti-backlash gear through driving torque so as to drive the inner frame system to move.

Select the following generalized coordinates

q＝[x y z α β γ]^T (2)

The x, y, z, alpha, beta and gamma respectively represent the translation of the inner frame system in the x direction, the translation in the y direction, the translation in the z direction, the yaw rotation around the x axis, the azimuth rotation around the y axis and the pitching rotation around the z axis in sequence.

Rigidity matrix

Wherein k is_a，k_r，k′_rRespectively comprising axial rigidity, main radial rigidity and secondary radial rigidity, k_α，k_β，k_γThe yaw stiffness of the inner frame system rotating around the x axis, the azimuth stiffness of the inner frame system rotating around the y axis and the pitch stiffness of the inner frame system rotating around the z axis are respectively. The above respective rigidities are calculated as follows:

the damping matrix C is completely identical to the stiffness matrix K in structure, only K at a corresponding position needs to be replaced by C, and the solving mode of each parameter in the damping matrix C can be obtained by adopting the existing solving mode in the field, and is not described herein again.

In the above formula, k_θTorsional rigidity of angular contact bearings, /)₁、l₂Respectively the distance between the contact point of the two half shafts and the bearing support part and the mass center of the inner frame structure, K_xIs the linear contact stiffness of the torsion spring.

External force matrix:

F＝[0,0,0,0,0,T]^T (4)

no external force is input other than the driving torque T of the gear 8.

When the system does undamped free vibration, [ C ], { F (t) } are all zero, and the solving equation can be changed into:

usually the free vibration of the system is a simple harmonic vibration, so the solution of equation (5) can be assumed to be:

in the formula: omega is the natural frequency of the structural vibration;

is the natural mode shape of the system. By substituting equation (6) for equation (5), the mode equation of the system can be obtained:

equation (7) is a homogeneous system of linear equations with a sufficient requirement for a non-zero solution that the determinant of the coefficient matrix is zero, i.e.:

|[K]-ω²[M]|＝0 (8)

the determinant can obtain the natural frequency omega of each order of the system_i. Will omega_i ²Respectively substituted into the vibration mode equation (7) of the system, and the corresponding n-dimensional column vector can be solved

I.e. omega_iCorresponding mode shape, i.e. modal information.

The modal information of the mechanism is required to be solved, namely equation (1) is solved, and then the acquired data is used for calculating each item involved in the stiffness matrix K in equation (1), wherein the specific calculation formula is the calculation formula of each stiffness.

In a more preferred embodiment, described in connection with an actual component measurement value, the respective measurement values for the seeker include: the rolling contact system comprises a backlash elimination gear rigidity k0, an axial pretension amount a, a drive torque T1, an angular contact ball bearing contact angle bayer _ angle, a backlash elimination gear pressure angle aear _ angle, an axial displacement amount x, a radial displacement amount y, a rolling element diameter Db equal to 5, an inner ring raceway coefficient fi equal to 0.55, an outer ring raceway coefficient fo equal to 0.53, an inner ring diameter Di equal to 20, an outer ring diameter Do equal to 42, a contact body 1 elastic modulus E1 equal to 2.07E +5, a contact body 2 elastic modulus E2 equal to E1, a contact body 1 poisson ratio v1 equal to 0.3, a contact body 2 poisson ratio v2 equal to v1, a rolling element number Nb equal to 16, a dynamic friction coefficient u equal to 0.2, a rolling distribution angle ane _ ball pi/16 [0,1,2,3,4,5,6,7,8,9,10,11,12,13, 15, and the like.

Based on these measurable values, we can further obtain intermediate quantities at each calculation:

the gear contact force F is T1/10-T2/50;

the contact force F1 generated by the anti-backlash gear torsion spring is k0 pi 1 4;

the torsional spring linear stiffness kx is k 0;

the medium diameter dm is (Di + Do)/2;

two-contact equivalent elastic modulus E ═ 1/((1-v1^2)/E1+ (1-v2^ 2)/E2);

the central distance A between the inner raceway and the outer raceway is (fi + fo-1) × Db;

angular contact ball bearing contact angle 25 pi/180;

the anti-backlash gear pressure angle gear _ angle is 20 × pi/180;

the main curvature radius Rim of the rolling body contacting with the inner raceway is 0.5 Db fi/abs (fi-0.5);

the minor curvature radius Rin of the rolling body contacting the inner raceway is 0.5 Db (Di + Do-2 Db cos (bear _ angle))/4/cos (bear _ angle));

the equivalent radius Ri of the rolling body in contact with the inner roller way is sqrt (Rim) Rin;

the contact stiffness coefficient kic of the rolling body in contact with the inner raceway is 4 × E × sqrt (ri)/3;

the main curvature radius Rom of the rolling element contacting with the outer raceway is 0.5 Db fo/abs (fo-0.5);

the main curvature radius Ron of the contact of the rolling bodies and the outer raceway is 0.5 × Db (Di + Do +2 × Db × cos (bear _ angle))/4/cos (bear _ angle)/(0.5 × Db + (Di + Do +2 × Db × cos (bear _ angle))/4/cos (bear _ angle));

the equivalent radius Ro of the contact between the rolling bodies and the outer raceway is sqrt (Rom Ron);

the contact stiffness coefficient koc of the contact between the rolling elements and the outer raceway is 4 × E × sqrt (ro)/3;

the combined contact rigidity coefficient k of the rolling body contacted with the inner and the outer raceways is (1/((1/kic) ^ (2/3) + (1/koc) ^ (2/3))) (3/2)

The axial displacement x is sin (bear _ angle) + a/a;

the radial displacement y is cos (bear _ angle);

the deformation z on the rolling body contact line is (x ^2+ y ^2) ^ 0.5.

Based on the intermediate quantities calculated above, we can obtain the respective final quantities:

axial stiffness ka ═ 1.5 ^ k ^ A ^0.5 ^ (z-1) ^0.5 ^ 2^ 16 (x/z);

radial stiffness kr ═ sum (1.5 ^ k ^ A ^0.5 ^ z-1) ^0.5 ^ y/z ^ 2^ cos (angle _ ball) ^ 2);

bearing friction torque T2 ═ u × k × a ^ (3/2) × (z-1) ^ (3/2) × dm/2;

the linear contact rigidity k1 of the fixed teeth of the anti-backlash gear and the driving wheel is 8E/3 (F5 20 100 sin (gear _ angle)/pi ^3/E/2/120) 0.25;

the combined linear stiffness k _ gear1 when the ordinary gears are meshed is T2/2500 k1/(T2/2500+ k 1);

the contact force generated by the torsion spring of the anti-backlash gear enables the linear contact rigidity formed between the driving gear and the loading gear

k2＝8*E/3*(F1*5*20*100*sin(gear_angle)/pi^3/E/2/120)^0.25；

The combined linear contact stiffness generated by the torsion spring force k _ gear2 ═ k2 × kx/(k2+ kx);

the combined torsional stiffness ko of the bearing is (k _ gear1+ k _ gear2) × 2500. Thereby completing the corresponding values to be solved in each stiffness calculation.

It should be noted that the above-mentioned manner of calculating the stiffness of the seeker as an exemplary illustration for further illustrating the embodiments of the present invention should not be construed as limiting the scope of the present invention, wherein the calculation and parameters of each of the directly acquired data and the intermediate data and the final stiffness amount in the above-mentioned stiffness calculation may be adjusted based on different digital twin objects or equipment objects to be assembled.

In a more specific embodiment, as shown in fig. 3, the digital twin model building process of the present invention is implemented by building digital twin models for complex equipment assembly processes from different levels and dimensions, and the digital twin models for the assembly processes include process level, regional level, and overall level assembly models according to levels; according to the dimensionality, the assembly process digital twin model comprises an assembly scene model, an assembly process model and an assembly analysis model, a time sequence diagram of the whole assembly process is built through SysML, the assembly scene model of the assembly process is built through Plant Simulation, for example, as shown in FIG. 4, the assembly analysis model is built according to the embodiment, and the system frequency domain dynamic characteristic analysis is carried out on the servo frame of the seeker, as shown in FIG. 5.

In a more preferred embodiment, the digital twin model is composed of a visualization model, a dynamic simulation model, and an assembly behavior model.

The assembly behavior model is a time sequence model of the assembly process, in this embodiment, the time sequence model of the whole assembly process can be constructed first, the time sequence diagram includes the time sequence of the parts and information transmission of the assembly work of the whole workshop, for the input and output of each node in the time sequence, one part of the input is the real-time quality data collected in the assembly process corresponding to the current node, and one part of the input is the output data of the previous process. After receiving all input information of the node, a dynamic simulation model is used for calculating quality characteristics, the key quality characteristics are predicted to obtain the output of the node, the output of each node is displayed through a visual model, and the scene model is updated according to the advance of time sequence for the visual model.

In a more preferable mode, the assembly of the seeker is taken as an example, the digital twin technology is taken as a carrier, the assembly process of the seeker servo mechanism is reconstructed for the purposes of synchronizing three-dimensional dynamic scenes of a workshop, visually implementing assembly data and restoring the real-time state of the workshop, and a simulation model capable of visualizing the assembly process of the seeker servo mechanism is formed. The assembly process of the seeker servomechanism is a complex and time consuming process that consists of many steps and is linked in the order of assembly. The whole assembly line takes the servo mechanism outer frame as a main line, the motor installation is started, the whole seeker servo mechanism is completed, and finally the seeker servo mechanism is tested and then sent to the next workshop for assembly.

The entire seeker servomechanism is divided into sections, each of which performs an assembly operation, and each assembly station has an operator performing the assembly operation. Parts are produced from source objects, transported by stations and conveyors to assembly stations, and a portion of the assembly is transported by worker transport, AGV and robotic arm. The assembly stations are located at the end of the conveyor belt and, since the movement of the parts on the conveyor belt takes a certain amount of time, in order to prevent the overstocking of parts, the transport time of the various parts must be coordinated so that the operator has enough time to operate without overloading. At each station, a suitable number of operators are allocated. For each station, the parts (bolts, nuts, etc.) required for assembly are transported to a buffer zone, which has different capabilities depending on the range. Failure to provide the necessary components at a certain time can result in assembly line failure. The process operation data required for the model operation, including the part number, the name of the operation performed at each location, the time required, etc., are stored in an internal database.

In this example, Adams was used to simulate the dynamics. The main assembly quality parameters influencing the performance of the servo mechanism of the seeker are the initial contact angle of a rolling bearing, the pre-tightening clearance of the bearing and the spring stiffness of the anti-backlash gear, and the parameters determine the stiffness of the joint surface of the assembly body, so that the mode of the whole assembly body is determined. After corresponding assembly quality parameters are measured in the assembly process, the influence of different assembly quality parameters on the resonant frequency of the servo mechanism of the seeker is calculated through Adams, frequency domain analysis is carried out based on a virtualization module, a sinking connection pair is added to the part joint surface, a virtualization calculation experiment is carried out on the rigidity in six-degree-of-freedom directions of each sinking connection pair through definition, and the model is shown in figure 5 after being led into Adams.

6) Predicting the quality characteristic: and further analyzing the quality of the products under production according to the digital twin simulation result, and if the quality parameter analysis result fluctuates abnormally, diagnosing by adopting an intelligent decision algorithm and providing a proper adjustment strategy.

In the quality analysis, calculation is carried out by using an established simulation model, after simulation calculation is carried out, a simulation result is analyzed, and if the change trend of a curve of the resonance frequency obtained by the simulation calculation is different from that expected, the curve is abnormal, or if the frequency peak value is not in a reasonable interval, the curve is abnormal.

In a more preferred embodiment, the above intelligent decision-making diagnosis can be implemented by using a fuzzy equation: the method comprises the steps of expressing the relation between the quality parameter abnormity of the assembly and the reason causing the abnormity by adopting a fuzzy equation, describing the running state of the quality abnormity in the assembly process and the fuzziness and uncertainty information in the environment based on a membership function and a fuzzy relation matrix in a fuzzy set, constructing the fuzzy relation equation, solving the approximate solution of the fuzzy relation equation by adopting an improved BP algorithm through a fuzzy deconvolution problem, and finding out the corresponding reason causing the quality abnormity.

Taking the guide head in the embodiment as an example, the abnormal expression symptom of the assembly quality in the assembly process of the servo mechanism of the guide head mainly comprises three aspects,

a. the geometric accuracy and the positioning accuracy of the assembly body are over-standard and are represented by PR.

b. The resonant frequency distribution of the assembly body is in a positive bias type/negative bias type (the resonant frequency waveform is fixed, the positive bias type index degree coefficient is more than 0, and the negative bias type index degree coefficient is less than 0), and is expressed by PL.

c. The peak value of the resonance frequency is more than \ less than 21.3Hz and more than 2Hz, and is represented by RE.

The analysis summarized possible causes of quality abnormality causing the above-described symptoms of assembly quality abnormality by analysis of the influence factors on the assembly process are shown in table 5 below.

TABLE 5

For each type of anomaly, a fuzzy relation matrix can be obtained by mutual comparison and normalization processing between the anomalies:

(the general expression of the fuzzy relation matrix R is described below, and is not described here again), the membership degree result of the fuzzy abnormal pattern calculated in a certain period is set as:

PR_mf＝0.4432，PL_mf＝0.8697，RE_mf＝0.1856

PR, PL, and RE are the calculation results of the membership degrees corresponding to the causes of the three abnormal symptoms in the above example, and it is assumed that the set of all possible causes of quality abnormality of the assembly unit during the assembly process is X ═ X (X is₁,X₂,…X_m) Wherein m is an anomaly factor; the set of quality anomalies resulting from m anomaly causes is Y ═ Y (Y)₁,Y₂,…Y_n) Where n is the number of mass anomaly classes.

Let the observed mass anomaly feature number sample be (y)₁,y₂...y_n)，

For each component y in the sample_iFor symptom Y_iThe degree of membership, the quality anomaly can be represented by a fuzzy vector as:

if the quality abnormality is caused by an abnormality cause X,

for the membership degree of X to various abnormal causes, the abnormal cause can be represented by a fuzzy vector as:

due to the causal relationship between the abnormal cause and the abnormal symptom, the fuzzy relation equation between Y and X can be obtained:

Y＝X°R (11)

wherein. Is a min-max fuzzy synthesis operator,

substituting the data into the fuzzy relation equation (11) includes:

Y＝(UPR(n-PR))，(UPR(n-PL))，(UPR(n-RE))＝(0.4432，0.8697，0.1856)

then the fuzzy relation equation to be solved is:

in the established fuzzy equation of the quality abnormity of the assembly, the fuzzy membership a of the quality abnormity in the fuzzy relation equation of the formula (12)_i＝U_A(x_i) (i ═ 1,2, …, m), in this example we use the algorithm of BP neural network to solve the equation.

According to the fuzzy relation equation of the formula (2), the membership degree a of the quality abnormal reason_i∈[0，1]. In the present case, the domain quality abnormality expression symptom U and the abnormality causing reason V are connected, and U belongs to [0,1 ]]，V＝Y∈[0，1]When m is 14 and n is 3, the samples are distributed on U and V, i.e. the samples are

X＝(x₁，x₂，…，x₁₄)＝(0.0358824，0.117321，…，1.0000)

Y＝(y₁，y₂，y₃)＝(0.33333，0.66667，1.00000)

Initial values χ (0) and σ (0) for χ and σ are determined, where χ (0) is chosen over U, let χ (0) be 0.3, σ (0) be 1, the initial j be 1, and given ε be 0.0001. Wherein, χ and σ are free parameters in fuzzy relation calculation.

The learning rate is then calculated according to the following formula: eta

Where τ is a constant and k is the number of iterations. Let Y be (Y)₁，y₂，y₃) Respectively calculating Z as ab, and searching for the maximum Z_iCorresponding to i. If the corresponding max Z is found_iHas a value of maxZ_i＝(Z₁，Z₅，Z₁₄) Adjusting U_x(X_m) The parameters x and sigma of_A(X₁)，U_A(X₅)，U_A(X₁₄) That is, the section variation value when i is 1, 5, and 14, the section variation with the smallest value is the optimal solution of the fuzzy equation, that is, the degree of contribution of the abnormality cause corresponding to the number to the assembly quality abnormality is the largest.

7) Intelligent control of assembly quality: analyzing the quality characteristics of the components, the assemblies and the whole body according to the quality prediction result, if the quality characteristic prediction value is abnormal, generating an active regulation and control strategy of assembly by an intelligent decision optimization method and feeding the active regulation and control strategy back to an assembly workshop in real time, and carrying out dynamic optimization and control on the current or subsequent assembly process by automatic assembly equipment or assembly production personnel in the production workshop to realize real-time dynamic regulation and control on the assembly process.

In a more preferred embodiment, taking the guidance head quality characteristic prediction in this embodiment as an example, based on the cause of the abnormality obtained by the abnormality diagnosis, combining the results of the dynamic simulation, and according to the difference between the simulation result and the expected simulation result, for example, if the resonant frequency peak value is too low, the bolt tightening torque is increased, and the pre-tightening gap of the angular contact bearing is reduced; if the resonance frequency peak value is too high, the bolt tightening torque is reduced, and the pre-tightening clearance is increased. Thereby realizing the dynamic regulation and control of the assembly quality.

In yet another embodiment, the present solution can be implemented by means of a device, which can include corresponding modules for performing each or several steps in the above-mentioned respective embodiments. Thus, each step or several steps of the above described embodiments may be performed by a respective module, and the device may comprise one or more of these modules. The modules may be one or more hardware modules specifically configured to perform the respective steps, or implemented by a processor configured to perform the respective steps, or stored within a computer-readable medium for implementation by a processor, or by some combination.

The device may be implemented using a bus architecture. The bus architecture may include any number of interconnecting buses and bridges depending on the specific application of the hardware and the overall design constraints. The bus connects together various circuits including one or more processors, memories, and/or hardware modules. The bus may also connect various other circuits such as peripherals, voltage regulators, power management circuits, external antennas, and the like.

Any process or method descriptions in flow charts or otherwise described herein may be understood as representing modules, segments, or portions of code which include one or more executable instructions for implementing specific logical functions or steps in the process, and alternate implementations are included within the scope of the preferred embodiment of the present disclosure in which functions may be executed out of order from that shown or discussed, including substantially concurrently or in reverse order, depending on the functionality involved, as would be understood by those reasonably skilled in the art of implementing the present disclosure. The processor performs the various methods and processes described above. For example, method embodiments in the present scheme may be implemented as a software program tangibly embodied in a machine-readable medium, such as a memory. In some embodiments, some or all of the software program may be loaded and/or installed via memory and/or a communication interface. When the software program is loaded into memory and executed by a processor, one or more steps of the method described above may be performed. Alternatively, in other embodiments, the processor may be configured to perform one of the methods described above by any other suitable means (e.g., by means of firmware).

The logic and/or steps represented in the flowcharts or otherwise described herein may be embodied in any readable storage medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, processor-containing system, or other system that can fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions.

It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by a computer program, which can be stored in a computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. The storage medium may be a magnetic disk, an optical disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), or the like.

The above description is only for the specific embodiment of the present invention, but the scope of the present invention is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present invention are included in the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. A dynamic tuning method based on performance digital twinning, characterized in that the method comprises:

step 4, analyzing the quality parameters of the products under production based on the simulation prediction data, and if the quality parameters are abnormal, determining the abnormal reason corresponding to the abnormal quality parameters;

2. The method of claim 1, wherein in step 1, the parameter data comprises pre-assembly parameters, post-assembly parameters, and robot arm parameters.

3. The method of claim 1, wherein the step 2, determining whether the data is abnormal comprises: and fitting the change curve of the historical data in a time period, fitting based on the change curve of the collected real-time data in the time period with the same length, comparing the distribution peaks of the two fitting curves, and judging that the data are abnormal when the distribution peak of the real-time data fitting curve is deviated to the left or right relative to the distribution peak of the historical data fitting curve, or the peak value of the peak is too high or too low.

4. The method according to claim 1, wherein the step 3, the predicting of the modality, further comprises:

5. The method of claim 4, wherein the mode shape equation is:

and, | [ K ]]-ω²[M]|＝0

Wherein ω is the natural frequency of the structural vibration;

6. The method according to claim 1, wherein in the step 3, the assembly behavior model is a time sequence model of an assembly process, and in this embodiment, the time sequence model of the whole assembly process may be first constructed;

and displaying the output of each node through a visualization model.

7. The method according to claim 1, wherein in the step 4, the abnormality cause corresponding to the quality parameter abnormality is determined by:

step 401, determining abnormal reasons causing various abnormalities;

wherein,

the degree of membership of X to various causes of abnormality,

is min-max fuzzy synthesis operator;

8. A dynamic tuning system based on performance digital twinning, the system comprising: the system comprises a quality data acquisition module, a data processing and mapping module, a digital twin module and an intelligent management and control module;

9. The system according to claim 8, wherein the system further comprises a physical entity module comprising a body unit, a measuring device unit, a detecting device unit and a signal transmitting device unit in an assembly plant;

10. A dynamic tuning device based on performance digital twinning, characterized in that the device comprises a processor, a storage device, the storage device storing instructions readable by the processor; the processor is configured to invoke instructions in the memory device to perform the performance digital twinning based dynamic tuning method of any of claims 1-7.