An intelligent hammering system and method with a dangerous gradient field and effective dose coupled closed loop

Abstract

The application discloses a dangerous gradient field and effective dose coupling closed loop intelligent hammering system and method, relates to the technical field of intelligent manufacturing, and the system comprises a controller, a motion unit, a variable compliance execution head, a multi-modal sensing unit and a tracing unit; wherein the variable compliance execution head is installed at the tail end of the motion unit and is used for adjusting the plastic strengthening of the metal component surface by the hammering tool; the controller is in communication connection with the motion unit, the variable compliance execution head and the multi-modal sensing unit, is used for dividing the surface of the metal component to be processed into a plurality of spatial units; and each unit is controlled and processed respectively. The application integrates the spatial risk gradient, the real plastic action energy, the thermal force constraint, the resonance damping and the compliance impedance regulation of the execution head into a unified closed loop, realizes quantitative, nondestructive and self-adaptive intelligent hammering strengthening of the residual stress of the metal component, and realizes full-link traceable and reproducible process data.

Description

Technical Field

[0001] This application relates to the field of train delay prediction, and in particular to an intelligent hammer impact system and method that couples a dangerous gradient field with an effective dose in a closed loop. Background Technology

[0002] After welding, additive manufacturing, and surface cladding of metal components, high-amplitude residual tensile stresses are easily generated in the surface layer, interface, weld toe, and thin-walled transition area. Under alternating loads and corrosive environments, fatigue cracks are prone to initiation, significantly reducing the service life of the components. Plastic impact processes such as hammering, ultrasonic impact, and mechanical rolling can offset residual tensile stresses through plastic extension of the surface layer and are the mainstream strengthening methods in the industry.

[0003] Existing intelligent hammer control solutions have four major flaws: 1. Risk assessment is static and singular, without spatial gradient constraints: Existing technologies simply superimpose risk indicators such as residual stress, thickness, and defects, without considering the abrupt gradient of risk on the curved surface; in high gradient areas such as weld toe and thin-walled edge, only increasing the single impact energy can easily cause local pressure damage, crack propagation, and thin-wall buckling, making it impossible to achieve differentiated control of "accumulating sufficient plastic deformation through multiple light impacts".

[0004] 2. Distorted processing dosage, unable to distinguish invalid resonant energy: Existing solutions use the number of impacts and the nominal energy of the instruction as the basis for processing measurement; during actual hammering, a large amount of energy excites resonance in the workpiece and fixture, without generating plastic stress release. The resonant energy is incorrectly included in the processing dosage, leading to excessive hammering, surface ripples, and equipment fatigue wear. At the same time, temperature, contact stiffness, and contact quality directly affect the plastic conversion efficiency, and fixed dosage standards cannot adapt to the variable state of fully curved materials.

[0005] 3. Static curing of temperature and stress thresholds: Traditional processes rely on manual experience to set fixed temperature and stress target ranges without considering local stiffness, defect risk, and impact energy for dynamic adjustment; insufficient plasticity conversion at low temperatures and easy surface damage and hardness reduction at high temperatures; insufficient stress control precision in high-risk areas, which can easily lead to stress overshoot or insufficient strengthening.

[0006] 4. The compliance of the execution head is uncontrollable and multiple parameters are independently adjustable: the stiffness of conventional hammer heads is fixed and cannot be adapted to various scenarios such as thin walls, thick walls and narrow weld toes; the frequency, step distance, damping and stroke parameters are independently adjustable, the impact coverage density is unbalanced after frequency adjustment, there is no impedance closed-loop matching mechanism, the surface normal fit is poor and the contact energy fluctuation is large.

[0007] In summary, existing technologies cannot integrate spatial risk gradient, real plastic effective energy, thermal safety constraints, resonance vibration reduction, and actuator compliance impedance into a unified closed loop, making it difficult to achieve quantitative, non-destructive, and adaptive intelligent hammering strengthening of complex heterogeneous metal components. A new integrated coupling control scheme is urgently needed. Summary of the Invention

[0008] In view of this, this application provides an intelligent hammering system and method with a closed-loop coupling of a hazardous gradient field and an effective dose. Addressing the shortcomings of existing technologies such as lack of risk gradient constraints, dose measurement distortion, static thresholds, and independent parameter control, this invention provides an intelligent hammering system and method with a closed-loop coupling of a hazardous gradient field and an effective dose. The system uses a hazardous field numerical value to characterize local comprehensive risk and a hazardous gradient to characterize the severity of risk changes, inversely constraining single impact energy and positively increasing the cumulative processed dose. It constructs a realistic stress release effective dose model that eliminates resonance energy and corrects for multiple material states. Based on a material response model, it solves for dynamic temperature and stress dual-constraint windows. Modal recognition and contact impedance feedback are used to adjust frequency, step size, damping, and stroke. Combined with a parallel magnetorheological compliant actuator, this forms a full-link closed-loop adaptive control, resolving the reinforcement contradiction in high-risk abrupt change regions and achieving precise and controllable control of residual stress.

[0009] A smart hammer impact system and method with a closed-loop coupling of a hazardous gradient field and an effective dose includes: Controller, motion unit, variable compliant actuator, multimodal sensing unit, and tracing unit; The motion unit is used to drive the hammering tool to move in relation to the workpiece being processed. A variable compliance actuator head is mounted at the end of the motion unit to adjust the hammering tool to perform hammering plastic strengthening on the surface of metal components; The multimodal sensing unit is used to acquire the state and impact response of the area to be processed; The traceability unit is used to store processed data; The controller is communicatively connected to the motion unit, the variable compliant actuator head, and the multimodal sensing unit to divide the surface of the metal component to be processed into multiple spatial units. The cumulative effective stress release dose D of each spatial unit was obtained by combining the state of the area to be treated and the impact response. eff(x) Compared with the estimated residual stress after the next impact; based on the cumulative effective stress release dose D eff(x) The estimated residual stress after the next impact controls the stopping of the motion unit; Obtain the processing priority, target cumulative effective dose D*(x), local step size p(x), and single impact energy limit E for each spatial unit. max(x) It generates control commands for adjusting the impact frequency, stroke, and local step distance of the motion unit, as well as control commands for the variable compliant actuator damping.

[0010] In one possible technical solution, the cumulative stress release effective dose D of each spatial unit is obtained by combining the state of the area to be processed and the impact response. eff(x)The estimated residual stress after the next impact includes: By obtaining the residual stress estimates, geometric risk quantities, defect risk quantities, temperature risk quantities, and service load risk quantities of each spatial unit based on the state of the region to be treated, a nonlinearly coupled danger field H(x) of each risk quantity on the surface of the metal component to be treated is constructed, and the surface gradient G of the danger field along the surface to be treated is calculated. H(x) ; According to H(x) and G H(x) Determine the processing priority, target cumulative effective dose D*(x), local step size p(x), and upper limit of single impact energy E for each spatial unit. max(x) ; And based on H(x) and G H(x) Dynamically determine the target stress window [σ] low(x) ,σ high(x) ]; During each impact, impact force, hammer velocity or displacement, contact duration, impact acceleration, temperature, and residual stress proxy signals are collected. The actual input energy per unit area is obtained by integrating the impact force and hammer velocity or displacement based on the contact duration. Local stiffness is obtained from the impact force-displacement response, and the resonant energy ratio is obtained from the impact acceleration spectrum. The effective stress release dose D is then accumulated by combining the material state correction factor and contact quality factor. eff(x) ; Predict the estimated residual stress after the next impact based on temperature and residual stress proxy signals.

[0011] In one possible technical solution, the effective dose D based on cumulative stress release... eff(x) The motion unit is stopped based on the estimated residual stress after the next impact, including: When the predicted next impact will cause the residual stress estimate to exceed σ low(x) Or the residual stress estimate enters the target stress window [σ low(x) ,σ high(x) And D eff(x) When D*(x) is reached, or when the marginal stress release rate is lower than the threshold for a preset number of consecutive cycles, the motion unit is controlled to stop.

[0012] In one possible technical solution, the dangerous field H(x) of the nonlinearly coupled risk quantities on the surface of the metal component to be treated is as follows: (2); in, The normalized m-th risk quantity includes at least the residual stress risk quantity. Geometric risk Defect risk level Temperature risk level and service load risk ; The risk sensitivity coefficient is limited to the preset upper and lower limits.

[0013] In one possible technical solution, the surface gradient GH(x) = || SH(x)||, where S is the gradient operator along the tangent plane of the surface to be processed; The controller generates a processing queue from high to low priority based on the combination of H(x) and GH(x), and ensures that the local step size satisfies... (6).

[0014] In one possible technical solution, the variable compliant actuator head includes: The system includes a housing, a ball joint or a two-degree-of-freedom micro-pendulum connecting seat connected to the housing, an electromagnetic linear drive unit located inside the housing, a hammer rod driven by the electromagnetic linear drive unit, a magnetorheological damping branch located between the hammer rod and the housing and connected in parallel with the impact force transmission path, a guide sleeve for guiding the hammer rod, a force sensor and a detachable hammer head connected in series at the end of the hammer rod, and an accelerometer and a displacement sensor located on the hammer rod or the moving mass.

[0015] In one possible technical solution, the detachable hammer head includes: At least two of the following: hemispherical hammerhead, flat-head hammerhead, and narrow-arc hammerhead; The controller selects the hammer type and radius based on curvature, weld toe angle, local thickness and defect risk. For thin-walled areas, it uses smaller single impact energy, larger damping and smaller stroke. For weld toe areas, it uses a narrow path along the weld toe tangent and a restricted normal deflection angle.

[0016] In one possible technical solution, the multimodal sensing unit includes: Residual stress sensors, temperature sensors, profile or displacement sensors, impact force sensors, hammer displacement or velocity sensors, and accelerometers; In one possible technical solution, the storage unit stores the danger field value, danger gradient, path number, temperature allowable window, target stress window, target cumulative effective dose, actual effective dose trajectory, impact parameter timing, local stiffness, contact impedance, resonance energy ratio, hammerhead identification, residual stress estimate and timestamp for each spatial unit, forming a process fingerprint that can be used for reproduction processing and updating risk sensitivity coefficient.

[0017] A smart hammer-impact method with a closed-loop coupling of a hazardous gradient field and an effective dose, the method comprising: Obtain the status and impact response of the area to be processed; Based on and combined with the state of the region to be treated and the impact response, the cumulative effective stress release dose D of each spatial unit is obtained. eff(x) Compared with the estimated residual stress after the next impact; based on the cumulative effective stress release dose D eff(x) The estimated residual stress after the next impact controls the stopping of the motion unit; Obtain the processing priority, target cumulative effective dose D*(x), local step size p(x), and single impact energy limit E for each spatial unit. max(x) ; Generate control commands for adjusting the impact frequency, stroke, and local step distance of the motion unit, as well as control commands for the variable compliant actuator damping; The motion unit and variable compliant actuator are adjusted by control commands; Save the processed data for reproducing the process and updating the process fingerprint of the risk sensitivity coefficient.

[0018] This application provides an intelligent hammer impact system and method with a closed-loop coupling of a hazardous gradient field and an effective dose, which has the following technical advantages compared with the prior art: Simultaneously, process parameters are allocated using two indicators: the hazardous field (total risk level) and the hazardous gradient (risk change rate). In areas of high mutation, the cumulative processing volume is increased, the single impact energy is limited, and the processing step distance is reduced. Sufficient plastic strengthening is achieved by relying on multiple light impacts, thus eliminating weld toe, thin-walled heavy impact cracking, and buckling defects.

[0019] Abandoning the crude measurement method of nominal energy and number of impacts, the system deducts the ineffective vibration energy of resonance in real time, and corrects the plastic conversion efficiency by combining temperature, stiffness, and contact quality. The effective dosage can be uniformly benchmarked across materials and curved surfaces, and the residual stress can be quantitatively controlled.

[0020] Based on the plasticity and damage calibration model, the temperature safety range is automatically solved, and the stress control range is adaptively tightened according to the risk gradient. The low temperature and high temperature protection are automatic, and the stress control accuracy in high-risk areas is significantly improved, avoiding insufficient strengthening or stress overshoot.

[0021] Impact frequency, feed step, stroke, and magnetorheological damping are optimized in a coordinated manner to avoid the resonance frequency band while ensuring uniform impact coverage; the parallel magnetorheological compliant actuator adjusts stiffness through a closed-loop contact impedance, and one set of equipment is suitable for all types of metal components, including flat, curved, thin-walled, and narrow weld toe components.

[0022] The storage unit fully records all processing parameters, sensing timing, and stress evolution data of each grid cell, forming a standardized process fingerprint. The processing effect can be completely reproduced for components in the same batch. Batch production data can be used to iteratively update the risk sensitivity coefficient and material correction model to continuously optimize control accuracy. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is a schematic diagram of the overall structure and closed-loop data flow of the intelligent hammer impact system of the present invention; Figure 2 This is a flowchart of the intelligent hammering control method of the present invention; Figure 3 A schematic diagram of the spatial allocation of path, step size, target dose, and single-click energy driven by a hazardous gradient field; Figure 4 This is a schematic diagram of a variable compliant actuator head structure; Figure 5 A schematic diagram of dynamic temperature window, dynamic stress window, and predicted shutdown status; Figure 6 This is a schematic diagram of self-aligning frequency anti-resonance and step-damping linkage; Figure 7 This diagram illustrates the effective dose accumulation, stress recovery, and shutdown criteria. Detailed Implementation

[0025] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0026] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0027] like Figure 1 As shown, an intelligent hammer impact system with a closed loop coupling a dangerous gradient field and an effective dose includes a controller, a motion unit, a variable compliant actuator, a multimodal sensing unit, and a tracing unit. The motion unit is used to drive the hammering tool to move in relation to the workpiece being processed. A variable compliance actuator head is mounted at the end of the motion unit to adjust the hammering tool to perform hammering plastic strengthening on the surface of metal components; The multimodal sensing unit is used to acquire the state and impact response of the area to be processed; The traceability unit is used to store processed data; The controller is communicatively connected to the motion unit, the variable compliant actuator head, and the multimodal sensing unit to divide the surface of the metal component to be processed into multiple spatial units. The cumulative effective stress release dose D of each spatial unit was obtained by combining the state of the area to be treated and the impact response. eff(x) Compared with the estimated residual stress after the next impact; based on the cumulative effective stress release dose D eff(x) The estimated residual stress after the next impact controls the stopping of the motion unit; Obtain the processing priority, target cumulative effective dose D*(x), local step size p(x), and single impact energy limit E for each spatial unit. max(x) It generates control commands for adjusting the impact frequency, stroke, and local step distance of the motion unit, as well as control commands for the variable compliant actuator damping.

[0028] like Figure 3 This diagram illustrates the spatial allocation of path, step size, target dose, and single-click energy driven by a hazardous gradient field, where D*(x) varies with H(x) and G. H(x) Increase and increase, E max(x) At least with G H(x) The risk decreases as either the defect risk or the thin-wall risk increases; One possible design is to obtain the cumulative stress release effective dose D of each spatial unit by combining the state of the region to be processed and the impact response. eff(x) The estimated residual stress after the next impact includes: By obtaining the residual stress estimates, geometric risk quantities, defect risk quantities, temperature risk quantities, and service load risk quantities of each spatial unit based on the state of the region to be treated, a nonlinearly coupled danger field H(x) of each risk quantity on the surface of the metal component to be treated is constructed, and the surface gradient G of the danger field along the surface to be treated is calculated. H(x) ; According to H(x) and G H(x) Determine the processing priority, target cumulative effective dose D*(x), local step size p(x), and upper limit of single impact energy E for each spatial unit. max(x) ; And based on H(x) and G H(x) Dynamically determine the target stress window [σ] low(x) ,σ high(x) ]; During each impact, impact force, hammer velocity or displacement, contact duration, impact acceleration, temperature, and residual stress proxy signals are collected. The actual input energy per unit area is obtained by integrating the impact force and hammer velocity or displacement based on the contact duration. Local stiffness is obtained from the impact force-displacement response, and the resonant energy ratio is obtained from the impact acceleration spectrum. The effective stress release dose D is then accumulated by combining the material state correction factor and contact quality factor. eff(x) ; Predict the estimated residual stress after the next impact based on temperature and residual stress proxy signals.

[0029] One possible design is that the effective dose D based on cumulative stress release... eff(x) The motion unit is stopped based on the estimated residual stress after the next impact, including: When the predicted next impact will cause the residual stress estimate to exceed σ low(x) Or the residual stress estimate enters the target stress window [σ low(x) ,σ high(x) And D eff(x) When D*(x) is reached, or when the marginal stress release rate is lower than the threshold for a preset number of consecutive cycles, the motion unit is controlled to stop.

[0030] One possible design is to consider the m-th original state variable. The risk quantity is obtained by truncating and normalizing. ∈[0,1]. The residual stress risk can be determined based on the relative position between the calibrated residual stress estimate and the allowable stress and target stress; the geometric risk can be determined by the principal curvature, thickness gradient, weld toe angle and edge safety distance; the defect risk can be determined by the defect probability given by ultrasonic, eddy current, visual or historical defect models; the temperature risk can be determined by the surface temperature gradient and cooling rate; the service load risk can be obtained by finite element or measured load mapping.

[0031] (1) The hazard field is represented by a product coupling approach, rather than a simple linear weighted summation of the individual risk quantities: (2) In the formula, This is the risk sensitivity coefficient. Productive coupling amplifies the effect when multiple risk factors increase simultaneously, thus distinguishing between single high-risk situations and risks arising from multiple factors. To avoid arbitrary weight settings, the system updates online in calibration batches and subsequent production batches using the error between the predicted and measured stress retreat amounts. .

[0032] (3) (4) In the formula, γ is the learning rate, ε is a positive number to prevent the denominator from being zero, and Π represents the projection to the preset upper and lower limits. When the actual stress regression is less than the predicted value, the error... A positive value indicates an increase in the sensitivity coefficient of the dominant risk factor in the current spatial unit; when the actual stress regression is greater than the predicted value, the corresponding sensitivity coefficient decreases. This answers the question of how risk weights are updated based on the processing results.

[0033] The surface hazard gradient is the gradient magnitude of the hazard field within the tangent plane of the curved surface: (5) The hazard gradient is used to identify risk mutation boundaries and, together with the hazard value, determines the path, step size, cumulative dose, and maximum click energy.

[0034] (6) (7) (8) The above relationship enables high-risk areas to be adequately treated with smaller step sizes and larger cumulative doses, while reducing the risk of defect propagation, thin-wall buckling, and weld toe damage by lower single-shot energy limits, thus forming a differentiated control that "increases cumulative investment and limits single-shot impact".

[0035] In the formula, p0, D0, and E0 are the reference step distance, the reference cumulative effective dose, and the reference single impact energy, respectively; , , , , and The non-negative sensitivity coefficient obtained through calibration; , , , , , These are the engineering constraint limits for the corresponding variables.

[0036] In the formula, the controller will allocate each spatial unit according to its overall priority. The parameters are sorted from high to low; within the same continuous region, the path direction is synthesized by the base coverage direction and the negative gradient direction of the danger field, enabling the execution head to smoothly transition from high-risk areas to low-risk areas. The coefficients mentioned above... This is the hazard gradient weighting coefficient, used to adjust the contribution of the hazard gradient term to the overall priority. Its value is determined by calibration based on the normalization results of the hazard field value and hazard gradient, as well as the control sensitivity. The larger this coefficient is, the higher the processing priority of gradient abrupt change regions.

[0037] The actual input energy per unit area of ​​the i-th impact is calculated from the measured signals of the impact force and the hammer rod velocity or displacement: (9) In the formula, The effective contact area is determined based on the hammer radius, indentation size, or contact model. and Let be the contact force and hammer rod velocity during the i-th impact, respectively; s be the hammer rod contact displacement. Local stiffness and contact resistance are defined as follows: (10) In the formula, and These represent the force increment and displacement increment for the i-th impact loading segment, respectively. Peak impact force; The initial hammer velocity is given. The contact quality factor is calculated from the ratio of contact duration to rebound velocity. (11) Power spectrum calculation of the impact acceleration signal, assuming the energy within the forbidden frequency band is... The total bandwidth energy is Then the resonance energy ratio = / In the formula, For effective contact duration, For reference contact duration, For rebound speed, and These represent the lower and upper limits of the contact quality factor. The effective dose increment for the i-th impact is: (12) (13) Temperature, stiffness, and hardness correction factors can be obtained from one-dimensional or multi-dimensional lookup table functions derived from calibration tests, or bounded analytical functions can be used. Since equation (12) deducts the energy falling into the modal ban band from the effective dose, the system will not mistake the overall vibration of the component for local plastic action.

[0038] The stress release efficiency per unit effective dose is updated online using exponential smoothing. Where, For smoothing coefficients, To prevent positive numbers with a denominator of zero, and These represent the lower and upper limits of stress relief efficiency, Δ. For the effective dose increment predicted based on the parameters of the next impact: (14) (15) in, Figure 5 This diagram illustrates the dynamic temperature window, dynamic stress window, and predicted shutdown conditions. The temperature window is not a fixed empirical range. The system pre-establishes a predictive plastic strain model through a small number of calibration impacts, finite element calculations, or material databases. and damage probability model The lower limit of the temperature window is defined as the lowest temperature at which the minimum effective plastic strain can be produced under the minimum allowable click energy. (16) The upper limit of the temperature window is defined as the highest temperature at which the probability of damage does not exceed a threshold and the hardness loss does not exceed a threshold, under the current upper limit of click energy. (17) If T < T1 and the temperature difference does not exceed the first threshold, the system only allows low-energy, low-frequency impact detection; if the temperature difference exceeds the first threshold, the system pauses and waits for heating or initiates preheating. If T > T2, the impact is stopped and cooling is waited for until T ≤ T2. ΔT hys After recovery, this is to avoid frequent starts and stops near the control boundary. Wherein, ΔT hys This represents the temperature hysteresis value, used to form a hysteresis interval between the stop threshold and the recovery threshold, avoiding frequent start and stop of hammering caused by temperature measurement fluctuations or thermal response lag; its value is determined by calibration based on temperature measurement error, temperature control response lag, and the allowable temperature margin of the material.

[0039] The lower limit T of the temperature allowable window 1(x) The lowest temperature at which the predicted plastic strain reaches the minimum effective plastic strain, given local stiffness and material hardness. The upper limit T of the temperature allowable window 2(x) To determine the highest temperature at which, given a maximum single impact energy, a dangerous gradient, and a defect risk level, the predicted surface damage probability does not exceed a damage threshold and the predicted hardness loss does not exceed a hardness loss threshold, the value of T is... 1(x) and T 2(x) The solution is obtained by calibrating the response model or by interpolation from the calibration data table.

[0040] When the measured temperature is lower than T 1(x) But with T 1(x) When the temperature difference does not exceed the first temperature difference threshold, the controller operates with a detection impact frequency and energy lower than the normal operating frequency and normal single-shot energy; when the temperature difference exceeds the first temperature difference threshold, the hammering is paused and the system waits for heating or starts preheating; when the measured temperature is higher than T... 2(x) Stop hammering and wait for it to cool down until the temperature drops to T. 2(x) Subtract the temperature hysteresis value.

[0041] The target stress window dynamically tightens according to the hazardous field.

[0042] In the formula, δ0 is the reference half-width. and These are the tightening coefficients corresponding to the hazardous field and the hazardous gradient, respectively: (18) (19) A narrower target stress window is used in high-risk or high-gradient areas. Before each impact, the controller uses equation (15) to predict the stress state after the next impact. When the predicted value will exceed... When necessary, reduce the click energy or stop the process to avoid stress overshoot and over-processing.

[0043] The controller performs short-time Fourier transform or power spectral density estimation on the impact response acceleration to identify modal peaks. Damping ratio And establish a disabled frequency band for each mode: (20) Within the permissible frequency set of the actuator, the controller selects the operating frequency f* by minimizing a combination of objectives: frequency offset, resonance risk, and dose tracking error. When the feed rate v... feed Given that p≈v feed / f* is used to synchronize and correct the step size, avoiding the formation of repetitive impact stripes after frequency modulation.

[0044] In the formula, I0 is the reference magnetorheological damping current. and These are the gain coefficients for resonance risk and contact impedance deviation, respectively. For reference contact impedance, I The maximum permissible magnetorheological damping current. The magnetorheological damping current is adjusted according to the following formula: (twenty one) Therefore, this invention does not change the frequency alone, but rather adjusts the frequency, step size, stroke, and damping as coupled variables in a coordinated manner. Figure 6 shows a schematic diagram of self-tuning anti-resonance and step-damping linkage; One possible design is, such as Figure 4 This is a schematic diagram of a variable compliant actuator head structure; The system includes a housing, a ball joint or a two-degree-of-freedom micro-pendulum connecting seat connected to the housing, an electromagnetic linear drive unit located inside the housing, a hammer rod driven by the electromagnetic linear drive unit, a magnetorheological damping branch located between the hammer rod and the housing and connected in parallel with the impact force transmission path, a guide sleeve for guiding the hammer rod, a force sensor and a detachable hammer head connected in series at the end of the hammer rod, and an accelerometer and a displacement sensor located on the hammer rod or the moving mass.

[0045] As one possible design, the detachable hammerhead includes: At least two of the following: hemispherical hammerhead, flat-head hammerhead, and narrow-arc hammerhead; The controller selects the hammer type and radius based on curvature, weld toe angle, local thickness and defect risk. For thin-walled areas, it uses smaller single impact energy, larger damping and smaller stroke. For weld toe areas, it uses a narrow path along the weld toe tangent and a restricted normal deflection angle.

[0046] As one possible design, the multimodal sensing unit includes: Residual stress sensors, temperature sensors, profile or displacement sensors, impact force sensors, hammer displacement or velocity sensors, and accelerometers; Before formal processing, the controller applies low-energy probing impacts to representative spatial units to calibrate the relationship between residual stress proxy signals and standard stress values, material state correction factors, local damage thresholds, and stress release efficiency per unit effective dose. The multimodal sensing unit synchronously acquires residual stress proxy signals, temperature, profile or displacement, impact force, hammer displacement or velocity, and acceleration, mapping the load-displacement-vibration response of the probing impact to the workpiece state, providing input for the aforementioned calibration. The resulting mapping relationship between the sensing signals and residual stress, damage risk, and effective dose parameters forms the sensing basis for subsequent closed-loop control.

[0047] One possible design is that the storage unit stores the hazard field value, hazard gradient, path number, temperature allowable window, target stress window, target cumulative effective dose, actual effective dose trajectory, impact parameter timing, local stiffness, contact impedance, resonant energy ratio, hammerhead identification, residual stress estimate and timestamp for each spatial cell, forming a process fingerprint that can be used for reproducibility processing and updating the risk sensitivity coefficient.

[0048] A smart hammer-impact method with a closed-loop coupling of a hazardous gradient field and an effective dose, the method as follows: Figure 2 include: Obtain the status and impact response of the area to be processed; Based on and combined with the state of the region to be treated and the impact response, the cumulative effective stress release dose D of each spatial unit is obtained. eff(x)Compared with the estimated residual stress after the next impact; based on the cumulative effective stress release dose D eff(x) The estimated residual stress after the next impact controls the stopping of the motion unit; Obtain the processing priority, target cumulative effective dose D*(x), local step size p(x), and single impact energy limit E for each spatial unit. max(x) ; Generate control commands for adjusting the impact frequency, stroke, and local step distance of the motion unit, as well as control commands for the variable compliant actuator damping; The motion unit and variable compliant actuator are adjusted by control commands; Save the processed data for reproducing the process and updating the process fingerprint of the risk sensitivity coefficient.

[0049] In this embodiment, the CAD surface of the component to be processed is registered with the 3D scanned surface, and discretized into spatial units with a grid side length of 1 mm to 5 mm. For each spatial unit, the temperature, curvature, local thickness, weld toe angle, defect probability, and service load mapping values ​​are obtained. Several representative units are selected for X-ray diffraction, blind hole method, or other standard residual stress detection, and calibration relationships are established with Barkhausen noise, surface wave ultrasound, or eddy current proxy signals.

[0050] On a representative unit, 3 to 10 impacts were performed with a detection energy lower than the lower limit of formal processing. F(t), δ(t), v(t), and a(t) were collected, and the local stiffness, contact impedance, resonant energy ratio, and actual input energy were calculated. Under the premise of no visible indentation, cracks, or excessive hardness, the temperature or impact energy was gradually increased, and a lookup table function was established.

[0051] Table 1. Calibration objects and their functions

[0052] For a curved repair component with a cladding layer-base material transition zone, a spatial element is selected with normalized residual stress risk, geometric risk, defect risk, temperature risk, and service load risk of 0.70, 0.35, 0.40, 0.20, and 0.50, respectively; and initial risk sensitivity coefficients of 0.30, 0.20, 0.20, 0.15, and 0.15, respectively. Substituting into equation (2), H(x) is approximately 0.455. G is calculated using the surface finite difference method based on the H values ​​of adjacent elements. H .

[0053] When the unit is located at the edge of the cladding layer and G H When the value is large, the controller increases D* while decreasing E. max And reduce the step size according to formula (6). This strategy differs from simply increasing the single-click energy according to the danger value; it can achieve high cumulative dose input through multiple low-energy impacts.

[0054] Table 2 Calculation output and control effect of the hazard gradient field

[0055] For steel components, calculations are performed point-by-point within the allowable temperature range given in the material database. and Assume that for a certain spatial element, under its current local stiffness and hardness, the minimum single-click energy cannot produce the specified minimum plastic strain when the temperature is below 160 ℃, and the minimum plastic strain is above 265 ℃ at E max If the predicted damage probability or hardness loss exceeds a threshold, the dynamic temperature window for that space element is set to [160 ℃, 265 ℃]. This value is only a calculation example; the actual lower limit is determined by the corresponding material and hammer calibration data.

[0056] If the measured temperature is 150℃ and the first temperature difference threshold is 15℃, the controller enters detection mode, limiting the frequency and single-shot energy to 30% to 60% of the normal value, allowing only a small number of impacts and reassessing the temperature and plastic response. If the temperature is below 145℃, the controller pauses and waits for the temperature to rise or starts preheating. If the temperature exceeds 265℃, the controller stops impacting until the temperature is below 265℃ minus the set hysteresis value before resuming.

[0057] In a certain impact, the controller obtains the actual input energy e per unit area by force-velocity integration. i ; calculate k based on the slope of the loading segment local q is calculated based on the ratio of contact duration to rebound speed. i Calculate R based on the acceleration spectrum i Substituting the above parameters, along with the temperature and hardness correction factors, into equation (12), we obtain ΔD. i And accumulate to D eff .

[0058] In the initial stage of treatment, the stress release efficiency η per unit effective dose is relatively high; as local plasticity approaches saturation, η gradually decreases. When the current... It has entered the target stress window and D eff Stop when D ≥ D*. Even if D eff If D* has not yet been reached, and equation (15) predicts that the next impact will exceed σ, then... low First, reduce the energy; if the minimum allowable energy still exceeds the limit, stop. If the marginal stress release rate of 3 to 10 consecutive impacts is lower than the threshold, also stop processing that unit.

[0059] Table 3 Closed-loop shutdown and protection logic

[0060] When processing thin-walled curved surfaces, 3D scanning or CAD models provide local thickness and curvature, which the controller uses to improve r.t And reduce E max The actuator uses a large-radius hemispherical hammer, which increases the magnetorheological damping current and reduces the stroke. The ball joint connection allows the hammer axis to follow the normal of the curved surface. In the weld toe area, the system uses a narrow-width circular arc hammer with its path arranged tangentially to the weld toe, and limits the angle of the hammer relative to the local normal, so that the impact is concentrated in the weld toe transition zone and does not directly hit the root of the acute angle.

[0061] When the local contact impedance is higher than the target impedance, the controller increases magnetorheological damping and reduces the stroke; when the local contact impedance is low and there is no risk of thin wall, the damping can be reduced and the stroke appropriately increased. By using impedance closed-loop rather than fixed stiffness settings, the same actuator head can be adapted to different curvatures, thicknesses, and material combinations.

[0062] Before and during the actual hammering, periodic impact detection is performed to identify the modal peaks of the workpiece-fixture system. If the current operating frequency falls into a certain prohibited frequency band, the controller selects a nearby non-prohibited frequency and recalculates the step distance based on the feed rate; if changing the frequency causes a decrease in the number of impacts per unit length, the step distance is reduced accordingly or the number of local reciprocating strokes is increased to maintain D* tracking. At the same time, the magnetorheological damping current is increased according to equation (21).

[0063] As the processing position changes, local stiffness and fixture boundary conditions change, and modal peaks may drift. The controller updates the spectrum with a sliding time window, and updates the disabled frequency band when the modal peak drift exceeds the set frequency difference to prevent the fixed vibration damping frequency from re-entering the resonance state in subsequent regions.

[0064] Table 4. Range of Main Parameters

[0065] The control algorithm can be implemented in an industrial computer, robot controller, programmable logic controller, or edge computing unit. The residual stress proxy sensor can select from Barkhausen noise, surface wave ultrasound, eddy current impedance, springback, or other online calibrable signals based on material magnetism and field conditions. The motion unit can be a six-axis robot, a gantry CNC platform, or a dedicated curved surface following mechanism.

[0066] To verify the technical effectiveness of this invention, a control group with the same components, initial state, and target stress conditions can be set up: the first group uses fixed parameters for equidistant hammering; the second group allocates nominal doses only according to the danger value; the third group uses the danger gradient field, effective dose, dynamic dual window, and modal-impedance synergistic closed loop of this invention. After treatment, the spatial proportion of residual stress entering the target window, total input energy per unit area, maximum indentation depth, surface waviness, resonant energy ratio, treatment time, and batch dispersion are compared.

[0067] This embodiment provides verification methods and judgment indicators, and does not use data that has not been confirmed by actual experiments as implementation results. After completing the prototype or test, the applicant may save the measured data as supplementary experimental records, or use it in subsequent responses to examination opinions to demonstrate unexpected technical effects.

[0068] Table 5. Comparison and Validation Indicators

[0069] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0070] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A smart hammer impact system with a closed-loop coupling of a hazardous gradient field and an effective dose, characterized in that the system... include: Controller, motion unit, variable compliant actuator, multimodal sensing unit, and tracing unit; The motion unit is used to drive the hammering tool to move in relation to the workpiece being processed. A variable compliance actuator head is mounted at the end of the motion unit to adjust the hammering tool to perform hammering plastic strengthening on the surface of metal components; The multimodal sensing unit is used to acquire the state and impact response of the area to be processed; The traceability unit is used to store processed data; The controller is communicatively connected to the motion unit, the variable compliant actuator head, and the multimodal sensing unit to divide the surface of the metal component to be processed into multiple spatial units. The cumulative effective stress release dose D of each spatial unit was obtained by combining the state of the area to be treated and the impact response. eff(x) Compared with the estimated residual stress after the next impact; based on the cumulative effective stress release dose D eff(x) The estimated residual stress after the next impact controls the stopping of the motion unit; Obtain the processing priority, target cumulative effective dose D*(x), local step size p(x), and single impact energy limit E for each spatial unit. max(x) It generates control commands for adjusting the impact frequency, stroke, and local step distance of the motion unit, as well as control commands for the variable compliant actuator damping.

2. The intelligent hammer impact system with a closed-loop coupling of a hazardous gradient field and an effective dose according to claim 1, characterized in that, The cumulative effective stress release dose D of each spatial unit is obtained by combining the state of the area to be treated and the impact response. eff(x) The estimated residual stress after the next impact includes: By obtaining the residual stress estimates, geometric risk quantities, defect risk quantities, temperature risk quantities, and service load risk quantities of each spatial unit based on the state of the region to be treated, a nonlinearly coupled danger field H(x) of each risk quantity on the surface of the metal component to be treated is constructed, and the surface gradient G of the danger field along the surface to be treated is calculated. H(x) ; According to H(x) and G H(x) Determine the processing priority, target cumulative effective dose D*(x), local step size p(x), and upper limit of single impact energy E for each spatial unit. max(x) ; And based on H(x) and G H(x) Dynamically determine the target stress window [σ] low(x) ,σ high(x) ]; During each impact, impact force, hammer velocity or displacement, contact duration, impact acceleration, temperature, and residual stress proxy signals are collected. The actual input energy per unit area is obtained by integrating the impact force and hammer velocity or displacement based on the contact duration. Local stiffness is obtained from the impact force-displacement response, and the resonant energy ratio is obtained from the impact acceleration spectrum. The effective stress release dose D is then accumulated by combining the material state correction factor and contact quality factor. eff(x) ; Predict the estimated residual stress after the next impact based on temperature and residual stress proxy signals.

3. The intelligent hammer impact system with a closed-loop coupling of a hazardous gradient field and an effective dose according to claim 1, characterized in that, The effective dose D based on cumulative stress release eff(x) The motion unit is stopped based on the estimated residual stress after the next impact, including: When the predicted next impact will cause the residual stress estimate to exceed σ low(x) Or the residual stress estimate enters the target stress window [σ low(x) ,σ high(x) And D eff(x) When D*(x) is reached, or when the marginal stress release rate is lower than the threshold for a preset number of consecutive cycles, the motion unit is controlled to stop.

4. The intelligent hammer impact system with a closed-loop coupling of a hazardous gradient field and an effective dose according to claim 2, characterized in that, The nonlinearly coupled danger field H(x) of the various risk quantities on the surface of the metal component to be treated is as follows: （2）； in, The normalized m-th risk quantity includes at least the residual stress risk quantity. Geometric risk Defect risk level Temperature risk level and service load risk ; The risk sensitivity coefficient is limited to the preset upper and lower limits.

5. The intelligent hammer impact system with a closed-loop coupling of a hazardous gradient field and an effective dose according to claim 2, characterized in that, The surface gradient GH(x) = || SH(x)||, where S is the gradient operator along the tangent plane of the surface to be processed; The controller generates a processing queue from high to low priority based on the combination of H(x) and GH(x), and ensures that the local step size satisfies... (6).

6. The intelligent hammer impact system with a closed-loop coupling of a hazardous gradient field and an effective dose according to claim 1, characterized in that, The variable compliant execution head includes: The system includes a housing, a ball joint or a two-degree-of-freedom micro-pendulum connecting seat connected to the housing, an electromagnetic linear drive unit located inside the housing, a hammer rod driven by the electromagnetic linear drive unit, a magnetorheological damping branch located between the hammer rod and the housing and connected in parallel with the impact force transmission path, a guide sleeve for guiding the hammer rod, a force sensor and a detachable hammer head connected in series at the end of the hammer rod, and an accelerometer and a displacement sensor located on the hammer rod or the moving mass.

7. The intelligent hammer impact system with a closed-loop coupling of a hazardous gradient field and an effective dose according to claim 6, characterized in that, The detachable hammerhead includes: At least two of the following: hemispherical hammerhead, flat-head hammerhead, and narrow-arc hammerhead.

8. The intelligent hammer impact system with a closed-loop coupling of a hazardous gradient field and an effective dose according to claim 1, characterized in that, The multimodal sensing unit includes: Residual stress surrogate sensors, temperature sensors, profile or displacement sensors, impact force sensors, hammer displacement or velocity sensors, and accelerometers.

9. The intelligent hammer impact system with a closed-loop coupling of a hazardous gradient field and an effective dose according to claim 1, characterized in that, The storage unit stores the hazard field value, hazard gradient, path number, temperature allowable window, target stress window, target cumulative effective dose, actual effective dose trajectory, impact parameter time series, local stiffness, contact impedance, resonance energy ratio, hammerhead identification, residual stress estimate and timestamp for each spatial cell, forming a process fingerprint that can be used for reproduction processing and updating the risk sensitivity coefficient.

10. A smart hammering method with a closed-loop coupling of a hazardous gradient field and an effective dose according to claim 1, used to control a smart hammering system with a closed-loop coupling of a hazardous gradient field and an effective dose according to any one of claims 1-9, characterized in that, The method includes: Obtain the status and impact response of the area to be processed; Based on and combined with the state of the region to be treated and the impact response, the cumulative effective stress release dose D of each spatial unit is obtained. eff(x) Compared with the estimated residual stress after the next impact; based on the cumulative effective stress release dose D eff(x) The estimated residual stress after the next impact controls the stopping of the motion unit; Obtain the processing priority, target cumulative effective dose D*(x), local step size p(x), and single impact energy limit E for each spatial unit. max(x) ; Generate control commands for adjusting the impact frequency, stroke, and local step distance of the motion unit, as well as control commands for the variable compliant actuator damping; The motion unit and variable compliant actuator are adjusted by control commands; Save the processed data for reproducing the process and updating the process fingerprint of the risk sensitivity coefficient.

Images