A logic control system for intelligent assembly of a heat sink

Abstract

The application relates to the field of intelligent assembly control and discloses a logic control system for intelligent assembly of a radiator, which comprises an industrial field bus interface unit, a control state observation unit and a core logic scheduling unit, and the core lies in that the core logic scheduling unit is used for monitoring statistical variance of force feedback signals to determine a contact transient state, starting a contact transient state logic masking program, cutting off a deviation accumulation path through latching integral branch state variables, and simultaneously switching a control architecture to an open-loop feedforward compensation topology based on a comprehensive disturbance estimator, the application uses a back electromotive force second-order differential residual to construct a masking mechanism, avoids an instruction step risk induced by sensor feedback lag of a physical contact transient state, converts uncontrollable mechanical impact into a determined logic dissipation process, and realizes smooth control right handover when a controlled assembly end crosses a physical boundary.

Description

Technical Field

[0001] This invention belongs to the field of intelligent assembly control technology, and in particular relates to a logic control system for intelligent assembly of radiators. Background Technology

[0002] The stability of the actuator during the transition from a high-speed approximation state in free space to a precise pressing state in a confined space is a core factor determining the assembly accuracy of the production line. Current mainstream solutions typically employ a control architecture based on force-position hybrid feedback, utilizing end-effector force sensors to collect the interaction forces at the contact interface and coordinating with position closed-loop feedback to adjust the actuator's motion trajectory. However, heat sink components are mostly non-standard thin-walled structures with non-uniform fin stiffness distribution. Under the dynamic conditions of high-speed assembly, a strong nonlinear impact is generated at the moment the mechanical actuator makes rigid contact with the heat sink. Due to the time delay effect of mechanical wave transmission and the inherent time constant of the sensor's internal filtering circuit, the external force feedback signal has a physical phase lag relative to the actual physical contact cross-section. If the control system drives the weighted smoothing algorithm based on such feedback data with phase lag, it will cause the integral term in the multivariable decoupling loop to accumulate errors and induce integral saturation, thereby generating step oscillations in the control command and increasing the risk of heat sink structure damage.

[0003] To resolve the aforementioned conflicts, conventional improvement approaches typically focus on increasing the sensor's sampling frequency or introducing a compliant buffer mechanism at the mechanical end. While increasing the sampling frequency can increase data density, it cannot overcome the inherent physical delay caused by the establishment and filtering of electrical signals. Introducing a compliant mechanism, on the other hand, reduces the system's positioning stiffness, increases the complexity of the control algorithm, and raises hardware maintenance costs. In addition to hardware physical constraints, the software control logic also has shortcomings. For example, Chinese invention patent application CN119781318A discloses a heat sink assembly control system, electronic device, and control method. By setting a pressure detection component between the heat sink and the chip, the motor is controlled to rotate the screw based on pressure feedback to adjust the gap between the chip and the heat sink. This type of solution is essentially a steady-state compensation mechanism based on static or quasi-static pressure feedback. The control closed loop highly depends on the real-time data feedback from the sensor. In high-speed dynamic assembly tasks, this logic architecture cannot overcome the physical phase offset of the sensor feedback path. When the execution end crosses the physical boundary, the lagging pressure signal causes the control logic to produce instruction mismatch, making it impossible to achieve transient impact dissipation. The delayed adjustment of the feedback instruction induces secondary mechanical oscillations.

[0004] Therefore, the technical problem to be solved by this invention is how to solve the problem of control loop command divergence caused by sensor feedback signal lag during physical contact transients by logically reconstructing the underlying topology of the control system without adding external mechanical compliance hardware. Summary of the Invention

[0005] In this technical solution, a logic control system for intelligent assembly of a radiator includes: The industrial fieldbus interface unit is used to connect to the controlled assembly execution end to synchronously acquire force feedback signals, displacement feedback signals, and phase current data that characterize the operating status of the controlled assembly execution end. The control state observation unit, connected to the industrial fieldbus interface unit, is used to calculate the comprehensive disturbance estimate of the controlled assembly execution end based on the force feedback signal and displacement feedback signal. The core logic scheduling unit is connected to both the industrial fieldbus interface unit and the control status observation unit. It monitors the statistical variance of the force feedback signal within a sliding time window and determines a contact transient when the statistical variance exceeds a preset fluctuation threshold, thereby initiating a contact transient logic masking procedure. This procedure includes: using a variable latch to maintain the cumulative error value of the integral branch in the sampling period preceding the contact transient determination time; simultaneously switching the control architecture from a closed-loop feedback topology to an open-loop feedforward compensation topology based on the comprehensive disturbance estimate; and using a preset command change rate constraint operator to limit the rate of change of the control command within a preset slope range. After the current operating clock crosses the preset masking time window boundary, the core logic scheduling unit restores the external feedback channel of the controlled assembly execution end.

[0006] Preferably, the core logic scheduling unit has a built-in back EMF observer, which is used to calculate the second-order differential residual of the back EMF, which characterizes the rate of change of mechanical impedance, based on the phase current data. When the core logic scheduling unit determines that the absolute value of the second-order differential residual of the back EMF is greater than the preset physical contact limit parameter, it triggers the contact transient logic masking program and switches the proportional gain coefficient to the preset damping control parameter to adaptively offset the drive frequency of the controlled assembly execution end.

[0007] Preferably, the control state observation unit adopts a linear state prediction model to convert the nonlinear characteristics of the controlled assembly execution end into a comprehensive disturbance estimate. Under the open-loop feedforward compensation topology, the comprehensive disturbance estimate is superimposed on the control output of the core logic scheduling unit in real time to achieve logical decoupling between spatial motion parameters and mechanical feedback parameters.

[0008] Preferably, the core logic scheduling unit uses a variable structure control algorithm to generate control commands. The variable structure control algorithm includes a time-varying gain nonlinear scheduling law. During the process of the controlled assembly execution end approaching the heat sink component, the gain is switched at a preset slope to suppress the pulsation of the output command.

[0009] Preferably, the industrial fieldbus interface unit uses the EtherCAT protocol for periodic synchronous communication. It reads the underlying drive data at a sampling period of no more than 1ms and uses a message check operator to filter out the transition noise in the displacement feedback signal. The servo driver baseboard includes an independent oscillator hardware sampling circuit to acquire the phase current oversampling sequence at a rate three orders of magnitude higher than the bus frequency. It uses a high-speed comparator to extract the edge transition characteristics of the sequence and transmits them across the clock domain to the main memory buffer so that the core logic scheduling unit can capture the state transition signal in microseconds.

[0010] Preferably, the control state observation unit includes a residual audit module, which is used to calculate the Euclidean distance between the estimated state value and the sensor feedback value, and output a state reset command to the core logic scheduling unit when the Euclidean distance exceeds a preset deviation threshold.

[0011] Preferably, the core logic scheduling unit includes a finite state machine architecture for logical arbitration between high-speed displacement mode, contact transient logic masking procedure and constant force contact mode, and reinitializes the control architecture when a state reset command is received.

[0012] Preferably, after exiting the contact transient logic masking procedure, the core logic scheduling unit guides the system state variables back to the closed-loop feedback topology through exponential approach logic, and uses the linear release of the integral branch state to compensate for the initial deformation deviation of the physical contact interface.

[0013] Preferably, the logic control system interacts concurrently with multiple controlled assembly execution terminals through an industrial fieldbus interface unit to achieve high-frequency collaborative assembly of the heat sink assembly across multiple task positions.

[0014] Compared with existing technologies, the logic control system for intelligent assembly of radiators in this invention has the following advantages: 1. In the logic control of intelligent radiator assembly, the inherent back electromotive force data in the current loop of the drive unit is used to calculate the stiffness distortion residual reflecting the transient state of physical contact, and a zero-delay prior identification mechanism for collision behavior is established. By avoiding the physical phase lag generated by traditional force sensors in the process of mechanical transmission and electrical signal filtering, the control system can capture the state transition signal within microseconds of the contact occurrence, eliminating the risk of command mismatch caused by feedback data distortion.

[0015] 2. By activating the masking mechanism during the physical contact transient, the input of distorted feedback data is directly cut off from the signal link, causing the control loop to enter the preset dimensionality reduction topology state. By using the state register freezing and integral branch bypass technology, the uncontrollable nonlinear impact is transformed into a deterministic static logic dissipation process, avoiding the integral saturation phenomenon induced by the phase lag of the feedback signal, and ensuring the smoothness of the handover of control and the continuity of command output when the system crosses the physical contact boundary.

[0016] 3. By utilizing the collaborative architecture of a third-order linear extended state observer and a reduced-order observer, comprehensive monitoring of unmodeled dynamics, external high-frequency noise, and lumped disturbances of the system is achieved. The observation results are fed forward to the control input through a preset time sliding mode reaching law algorithm, so that multi-source variables converge to the sliding surface within a finite time. This achieves dynamic decoupling of spatial motion parameters and contact mechanical parameters, and improves the system's ability to suppress high-frequency fluctuations and disturbances under complex non-standard working conditions. Attached Figure Description

[0017] Figure 1 This is a flowchart of the intelligent assembly contact transient logic control process for the radiator of this invention; Figure 2 This is a block diagram of the hardware architecture of the intelligent assembly logic control system of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0019] It should be noted that all directional and positional terms used in this invention, such as: up, down, left, right, front, back, vertical, horizontal, inner, outer, top, bottom, transverse, longitudinal, center, etc., are only used to explain the relative positional relationship and connection between components in a specific state (as shown in the accompanying drawings). They are only for the convenience of describing this invention and do not require that this invention be constructed and operated in a specific orientation. Therefore, they should not be construed as limiting this invention. In addition, the descriptions of "first," "second," etc., in this invention are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated.

[0020] In the description of this invention, unless otherwise explicitly specified and limited, the terms installation, connection, and linking should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; they can refer to mechanical connections; they can refer to direct connections or indirect connections through an intermediate medium; they can refer to the internal connection of two components. For those skilled in the art, the specific meaning of the above terms in this invention can be understood according to the specific circumstances.

[0021] In the description of this specification, references to the terms "an embodiment," "some embodiments," "illustrative embodiments," "examples," "specific examples," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example, and the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0022] A logic control system for intelligent assembly of radiators, comprising: The industrial fieldbus interface unit is used to connect to the controlled assembly execution end to synchronously acquire force feedback signals, displacement feedback signals, and phase current data that characterize the operating status of the controlled assembly execution end. The control state observation unit, connected to the industrial fieldbus interface unit, is used to calculate the comprehensive disturbance estimate of the controlled assembly execution end based on the force feedback signal and displacement feedback signal. The core logic scheduling unit is connected to both the industrial fieldbus interface unit and the control status observation unit. It monitors the statistical variance of the force feedback signal within a sliding time window and determines a contact transient when the statistical variance exceeds a preset fluctuation threshold, thereby initiating a contact transient logic masking procedure. This procedure includes: using a variable latch to maintain the cumulative error value of the integral branch in the sampling period preceding the contact transient determination time; simultaneously switching the control architecture from a closed-loop feedback topology to an open-loop feedforward compensation topology based on the comprehensive disturbance estimate; and using a preset command change rate constraint operator to limit the rate of change of the control command within a preset slope range. After the current operating clock crosses the preset masking time window boundary, the core logic scheduling unit restores the external feedback channel of the controlled assembly execution end.

[0023] Preferably, the core logic scheduling unit has a built-in back EMF observer, which is used to calculate the second-order differential residual of the back EMF, which characterizes the rate of change of mechanical impedance, based on the phase current data. When the core logic scheduling unit determines that the absolute value of the second-order differential residual of the back EMF is greater than the preset physical contact limit parameter, it triggers the contact transient logic masking program and switches the proportional gain coefficient to the preset damping control parameter to adaptively offset the drive frequency of the controlled assembly execution end.

[0024] Preferably, the control state observation unit adopts a linear state prediction model to convert the nonlinear characteristics of the controlled assembly execution end into a comprehensive disturbance estimate. Under the open-loop feedforward compensation topology, the comprehensive disturbance estimate is superimposed on the control output of the core logic scheduling unit in real time to achieve logical decoupling between spatial motion parameters and mechanical feedback parameters.

[0025] Preferably, the core logic scheduling unit uses a variable structure control algorithm to generate control commands. The variable structure control algorithm includes a time-varying gain nonlinear scheduling law. During the process of the controlled assembly execution end approaching the heat sink component, the gain is switched at a preset slope to suppress the pulsation of the output command.

[0026] Preferably, the industrial fieldbus interface unit uses the EtherCAT protocol for periodic synchronous communication. It reads the underlying drive data at a sampling period of no more than 1ms and uses a message check operator to filter out the transition noise in the displacement feedback signal. The servo driver baseboard includes an independent oscillator hardware sampling circuit to acquire the phase current oversampling sequence at a rate three orders of magnitude higher than the bus frequency. It uses a high-speed comparator to extract the edge transition characteristics of the sequence and transmits them across the clock domain to the main memory buffer so that the core logic scheduling unit can capture the state transition signal in microseconds.

[0027] Preferably, the control state observation unit includes a residual audit module, which is used to calculate the Euclidean distance between the estimated state value and the sensor feedback value, and output a state reset command to the core logic scheduling unit when the Euclidean distance exceeds a preset deviation threshold.

[0028] Preferably, the core logic scheduling unit includes a finite state machine architecture for logical arbitration between high-speed displacement mode, contact transient logic masking procedure and constant force contact mode, and reinitializes the control architecture when a state reset command is received.

[0029] Preferably, after exiting the contact transient logic masking procedure, the core logic scheduling unit guides the system state variables back to the closed-loop feedback topology through exponential approach logic, and uses the linear release of the integral branch state to compensate for the initial deformation deviation of the physical contact interface.

[0030] Preferably, the logic control system interacts concurrently with multiple controlled assembly execution terminals through an industrial fieldbus interface unit to achieve high-frequency collaborative assembly of the heat sink assembly across multiple task positions.

[0031] Example 1: In a scenario involving the high-frequency concurrent assembly of multiple heat sink components, the controlled assembly actuator performs a high-speed position approximation action in free space and needs to switch to a high-precision force-controlled contact state instantaneously upon contact with the heat sink fins. Due to the time delay in the transmission of mechanical structure waves and the sensor electrical signal... During the filtering process, the external force feedback signal lags behind the physical contact section by 3ms to 5ms. If the control system continues to drive the weighted smoothing algorithm based on this feedback signal with phase lag, the integral term in the closed loop will accumulate deviations, triggering command steps and increasing the risk of crushing of the thin-walled heat sink structure. The industrial fieldbus interface unit synchronously reads the phase current data and rotor position data in the servo driver of the controlled assembly execution end according to a sampling period of 1ms. The servo driver baseboard contains an independent oscillator hardware sampling circuit to acquire the phase current oversampling sequence at a rate three orders of magnitude higher than the bus frequency. The high-speed comparator extracts the edge transition characteristics of the sequence and transmits them across the clock domain to the main memory buffer. The core logic scheduling unit uses the built-in back EMF observer. The system calculates the second-order difference residual of the back electromotive force (EMF) in real time, which characterizes the abrupt change rate of mechanical impedance. This residual serves as an identification index reflecting collision behavior and captures state transition signals within microseconds, thus avoiding the physical time delay caused by external sensors during mechanical transmission. In the internal data conversion process of the back EMF observer, the system extracts the current terminal voltage control command based on the drive bridge duty cycle. It then subtracts the product of the measured phase current and the stator internal resistance from this terminal voltage, and further subtracts the product of the stator inductance constant and the first time derivative of the phase current. This deconstructs a single-valued estimate of the transient back EMF. The system performs two consecutive subtraction operations on the back EMF history sequence output within three consecutive oversampling clock cycles, with adjacent step sizes, and finally outputs the second-order difference residual of the back EMF characterizing the rapid velocity drop.

[0032] When the two-stage adaptive transition module determines that the absolute value of the second-order differential residual of the back electromotive force is greater than the preset physical contact limit parameter, it outputs a topology blocking command to the global scheduling core. In response to this command, the global scheduling core cuts off the input port of the control logic receiving external force sensor and displacement feedback signals in the data flow link, and simultaneously sends a third-order linear expansion state observer to the control state observation unit. A register freeze instruction is issued, suspending the displacement observations, velocity observations, and comprehensive disturbance estimates within the register and maintaining the static cross-sectional values ​​at the moment before triggering. Simultaneously, in the control instruction operation path, a variable latch maintains the cumulative error value of the integral branch at the moment of contact transient determination in the previous sampling period. This two-stage adaptive transition module is directly embedded in the operation flow of the core logic scheduling unit in the form of software logic. Its first stage is the hard dead zone comparison stage, which is specifically responsible for locking electrical high-frequency mutations exceeding the reference threshold within microseconds and triggering unidirectional safety blocking. Its second stage is the slope tracking and unlimiting stage, which is specifically responsible for confirming the completion of the mechanical oscillation dissipation trend by continuously comparing the convergence slope of the residual envelope after the blocking action takes effect, thereby independently providing an arbitration flag for exiting the masking procedure, thus completing the intermediate mechanism for the system to switch from mutation protection to steady-state tracking.

[0033] During the 5ms time window defined by the safety dead zone time window register, the core logic scheduling unit switches the control weight to the preset damping control parameters and outputs an open-loop constant control command. It extracts the fluctuation component of the initial displacement feedback signal during the contact transient, extracts the mechanical resonant frequency constituting the fluctuation, and synthesizes a counter-impact frequency signal with a phase delay of half a cycle based on the resonant frequency. This counter-impact frequency signal is then positively superimposed onto the open-loop constant control command based on the damping control parameters to drive the controlled assembly execution end. Once the clock counter value reaches the upper limit of the time window and the physical oscillation has passed the dead zone, the global scheduling core cancels the topology blocking command and reopens the external feedback channel. Using the observer state calculated through full closed-loop iteration as the initial value, it guides the system state variables to return to the closed-loop feedback topology through exponential convergence logic, allowing the controlled assembly execution end to smoothly enter the force-controlled bonding condition. The resonant frequency here is not calculated in real-time within the microsecond-level blind zone where contact occurs. Instead, the control system performs a pre-calibration test of contact surface impedance frequency sweep before assembly and operation. It uses vibration analysis to extract the inherent mechanical resonance frequency of the radiator fins of this specific model under step impact and stores it in the non-volatile memory of the scheduling unit for a long time. When the contact transient logic masking program is triggered during the actual dynamic assembly process, the system avoids the computational cost of real-time frequency domain analysis and directly calls the preset main frequency parameter from the memory. It calculates the half-cycle delay and directly synthesizes the impulse waveform in sequence during the transient. The control system uses the instantaneous response characteristics of back electromotive force to rotor micro speed distortion to enhance the adaptability of the control loop to sensor phase lag. Through the cooperative logic of forced switching of the control architecture and state variable latching, it eliminates the dependence of multivariable control on external sensor feedback path during state transition. Without adding external compliant mechanical hardware, it improves the operational stability of the system when dealing with complex variable coupling tasks.

[0034] Example 2: In the verification test of the precision pressing process of heat sink fins, the purpose of this test is to verify the suppression effectiveness of the logic masking program based on back electromotive force characteristics on the force overshoot induced by sensor hysteresis. The test platform is built on a mechanical assembly unit. The controlled assembly execution end is driven by a permanent magnet synchronous motor with a rated torque of 2.4 Nm. The data source is generated from the internal sampling register of the matching servo driver. The phase current data acquisition resolution is 16 bits, and the sampling frequency is set to 10 kHz. The external force sensor is connected to the system through an analog-to-digital converter. Affected by the RC filtering process in the signal conditioning circuit, the effective sampling frequency is maintained at 500 Hz.

[0035] Regarding the sampling period of core parameters The setting is influenced by factors such as the discretization error of the control algorithm, the processing load of the logic scheduling unit, and the sampling period. The setting is used to balance the real-time performance of data acquisition with the system communication bandwidth load. The decision logic follows the sampling theorem constraint. When the bandwidth requirement of the back EMF observer reaches 500Hz or more, in order to maintain the numerical stability of the observer's state variables and avoid signal aliasing, the sampling period is... The sampling period was determined to be 1ms in this experiment because the sampling rate tends to be 1ms. At the same time, in order to verify the reliability of the system in an industrial electromagnetic environment, Gaussian white noise with a root mean square value of 0.3N was superimposed on the force feedback signal and 50Hz power frequency harmonic interference was simulated. During the test, the controlled assembly execution end moved towards the heat sink base at a speed of 150mm / s. The performance of the force feedback signal in the original input data stream during the contact transient was recorded. The control group as a reference generated a peak impact force of 28.6N at the moment of contact. The corresponding control command generated an oscillating step within 6ms after contact. The 2ms delay of the force feedback signal caused the multivariate decoupling algorithm to maintain the high-speed movement command after the physical contact occurred, which caused the deviation accumulation of the integral branch circuit.

[0036] The core logic scheduling unit of this invention, 3.5ms before detecting variance fluctuations in the force feedback signal, captures the second-order difference residual caused by the rotor speed drop through a back electromotive force observer. When its value reaches... When the physical contact limit parameters are reached, the system initiates a contact transient logic masking procedure, locks the cumulative error value of the integral branch at 0.0084 at the trigger time using a variable latch, and switches the control topology to one based on the comprehensive disturbance estimate. To verify the synergistic effect of the feedforward compensation mode, a partially missing control group was set up in the experiment, which lacked prior back EMF determination. This group only activated the latching action after detecting that the variance of the force feedback signal exceeded the standard. Data showed that the peak impact force of the sample group of this invention was 12.3N and the pressure curve tended to smooth after entering the masking window. The partially missing control group generated an impact of 19.8N. The data confirmed that there is a synergistic relationship between the back EMF characteristics and the variable latching mechanism, that is, the former provides the trigger advance for the latter. In the boundary test for the reasonableness of the numerical range, out-of-range control groups were set up with masking window durations of 1ms and 10ms. When the time window was set below 1ms, the masking procedure resumed feedback before the mechanical oscillation disappeared, causing the system to reintroduce disturbance and generate 22.1N. The impact of N; when the time window is set to more than 10ms, the system is in an open-loop state, causing the position tracking error to shift to 0.85mm, which exceeds the assembly accuracy requirements of the heat sink. However, the prototype of this invention, within the range of 3ms to 8ms, achieves impact suppression while controlling the position deviation within 0.05mm, confirming that the defined parameter range is a window that takes into account both dynamic response and assembly accuracy. The gradient correlation data obtained from the experiment shows that as the moving speed increases from 50mm / s to 300mm / s, the force overshoot rate of the prototype of this invention remains within the controlled range, without force overshoot variation. This confirms that the logic control system improves the operational reliability of the assembly task under varying working conditions by transforming the nonlinear impact of the physical boundary into a logic dissipation process determined within the control architecture.

[0037] Example 3: This example combines Figures 1 to 2 A description of a logic control system for intelligent assembly of a radiator, such as... Figure 1 As shown, the controlled assembly execution end outputs operating status and mechanical parameters, which are decomposed into force feedback signal, displacement feedback signal, and phase current data. The force feedback signal represents the force at the contact interface, the displacement feedback signal represents the end motion trajectory, and the phase current data is used to extract stiffness distortion electrical parameters reflecting the physical contact transient. The force feedback signal and displacement feedback signal are jointly input to the comprehensive disturbance estimate based on the nonlinear characteristic transformation link, and the output of this link is laterally directed to the control architecture switching to the open-loop feedforward compensation topology link to provide feedforward compensation components. Simultaneously, the system performs real-time monitoring in the control state monitoring force feedback signal statistical variance link, and after confirming the trigger condition in the out-of-bounds transient judgment statistical variance exceeding the preset fluctuation threshold link, the system sequentially starts the variable latch to hold the accumulated error value of the locked product branch, the control architecture to switch to the open-loop feedforward compensation topology, and the command change rate constraint operator to limit the preset slope of the control command. Finally, after the external feedback channel recovers and crosses the preset masking time window boundary, the reset link completes the re-handover of the closed-loop control of the system.

[0038] like Figure 2As shown, the logic control system includes a controlled assembly execution end, an industrial fieldbus interface unit, a control state observation unit, and a core logic scheduling unit in its hardware architecture. The controlled assembly execution end and the industrial fieldbus interface unit are connected through a bidirectional signal link to realize the synchronous acquisition of operating data and the real-time issuance of control commands. The industrial fieldbus interface unit outputs low-level sampled data to the control state observation unit and the core logic scheduling unit respectively. The output end of the control state observation unit is connected to the core logic scheduling unit to transmit the calculated system state observation results and the comprehensive disturbance estimate to the core logic scheduling logic for arbitration decision-making.

[0039] Example 4: In the flexible assembly of heat sink components involving multiple batches and fluctuating material stiffness, the controlled assembly execution end faces the technical problem of inconsistent physical impedance characteristics of fins in different batches. If fixed physical contact limit parameters are used, the core logic scheduling unit may miss or falsely trigger during contact determination, preventing the contact transient logic masking program from starting at the precise physical contact section. This weakens the control system's ability to handle phase lag in force feedback signals. Therefore, it is necessary to determine physical contact limit parameters suitable for the current controlled assembly execution end. The control system operates under quasi-static load calibration procedures, driving the controlled assembly actuators at a preset calibration speed under no-load conditions. Run and record The back EMF observer output for each sampling period is used to extract phase current data using an industrial fieldbus interface unit. The root mean square value of the background noise is calculated based on the second-order difference residual sequence of the back EMF. Physical contact limit parameters are then used. The calibration procedure sets the threshold to be 3 to 5 times the root mean square value of the background noise calculated based on the second-order difference residual sequence of the back electromotive force. This threshold is anchored above the electrical background noise level of the controlled assembly execution end itself.

[0040] The core logic scheduling unit integrates discretized observation logic at each sampling time. At this point, the system reads the current displacement feedback signal. The system state estimate is updated using a third-order linear extended state observer. The specific observer state transition equation is as follows: ,in, This is the observation error; These are displacement observations. For velocity observations, This is a comprehensive disturbance estimate that includes external disturbances. The sampling period is The observer gain coefficient, The control gain constant, To provide control commands to the driver, in this iterative formula system, the comprehensive disturbance estimate does not simply rely on the implicit derivation of the input-output residuals, but directly introduces the force feedback signal as an observable deterministic external disturbance term for dimensionality reduction and fusion. The system uses the industrial fieldbus to obtain the measured value of the force feedback signal, divides it by the equivalent mass of the controlled assembly execution end to obtain the torque acceleration equivalent, and in the third-order variable update cycle of the state transition equation, multiplies the torque acceleration equivalent by the sampling period and directly uses it as a feedforward compensation component to be superimposed into the estimated disturbance value, thereby explicitly transforming the dynamic force of the physical contact interface into a mathematical benchmark participating in the closed-loop iteration of the observer.

[0041] Based on this, the core logic scheduling unit synchronously runs the sliding time window variance calculation program, by maintaining a length of The force feedback signal is in a first-in-first-out queue at each sampling time. Remove the old data from the head of the queue and push in the latest measured value of the force feedback signal. By using an incremental calculation formula to update the statistical variance, the computational complexity does not change with the length of the time window. The value increases with the increase of the second-order difference residual of the captured back electromotive force, and rises when the instantaneous value of the second-order difference residual of the captured back electromotive force exceeds the physical contact limit parameter obtained by the aforementioned calibration. At that time, the core logic scheduling unit triggers a register locking action, causing the state transition equation to... , and Forced Inheritance The value is unaffected by feedback error. The correction mechanism avoids the observer divergence risk caused by sensor phase lag by cutting off the bias injection channel in the feedback loop.

[0042] Once the clock counter in the safe dead-time window register reaches the preset boundary, the system removes the register lock and utilizes the recovered comprehensive disturbance estimate. As a feedforward component injected into the sliding mode control loop, it guides the controlled assembly execution end to complete the heat sink pressing under a defined damping constraint. This method, based on a discrete state transition model and dynamic calibration threshold, enables the system to maintain stable control handover logic when facing heat sink components with different impedance characteristics. It resolves the contradiction that control accuracy is limited by environmental changes under complex non-standard working conditions. By disclosing the discrete iterative equations and threshold calibration procedures of the third-order linear extended state observer, this system reduces the functional implementation logic of the control unit from an abstract algorithm name to a defined mathematical recursive process, ensuring the portability of the technical solution on different computing platforms and ensuring that those skilled in the art can reproduce the transient masking and disturbance compensation functions for heat sink assembly based on the description.

[0043] Example 5: In a scenario involving the replacement of a drive motor at a controlled assembly execution end, the winding resistance of different batches of drive motors varies. With stator inductor Due to differences in physical parameters, the logic control system operates according to a parameter identification procedure. The core logic scheduling unit outputs a high-frequency pulse voltage excitation signal with an amplitude of 10% of the rated current to the servo driver. The measured resistance value is obtained by reading the phase current data and using a least-squares iterative algorithm. Compared with the measured inductance value The identified values ​​are transmitted to the back EMF observer register. The system obtains the actual impedance time parameter by dividing the measured inductance value by the measured resistance value, extracts the reference time parameter from the drive motor nameplate, calculates the quotient obtained by dividing the actual impedance time parameter by the reference time parameter, generates a calibration factor reflecting electrical characteristic drift, multiplies the default gain coefficient by the calibration factor, and corrects the gain coefficient according to the physical parameter deviation. and This allows the convergence speed of the system's state variables to be adapted to the physical characteristics of the drive mechanism.

[0044] In scenarios involving material switching of the radiator base, considering the differences in crush load for different materials, the system executes a slope constraint calibration process. The controlled assembly execution end performs a quasi-static displacement feed action on the sample. A mechanical impedance characteristic model is constructed using force feedback signals and displacement feedback signals. The characteristic model's operating procedure drives the controlled assembly execution end to perform a bonding action at a constant speed. The changes in force feedback signals and displacement signals within five consecutive sampling clock cycles are extracted. The two are divided to output the local stiffness of the contact surface. The system's built-in radiator ultimate crush force is called, and the product of the local stiffness and the current pressing depth is subtracted to output the remaining safe load. The remaining safe load is divided by the total inertia and mass of the moving parts. Based on the mechanical impedance characteristic model, the maximum acceleration vector of the controlled assembly execution end under the boundary condition of not causing damage to the radiator structure is calculated. The core logic scheduling unit determines the slope parameter of the instruction change rate constraint operator based on this vector. During the transient logic masking program, the physical impact energy is attenuated by limiting the rate of change of control commands. Once the increment of commands in the command buffer stack is lower than the preset safety threshold, the controlled assembly execution end enters the steady-state force-controlled bonding condition. In the underlying mathematical chain of calculating the local stiffness of the contact surface by dividing by the displacement change, in order to avoid the fatal arithmetic singularity caused by the physical assembly terminal stopping instantaneously on the rigid contact surface, a lower limit shielding condition is forcibly implanted inside the controller. Once it is confirmed in real time that the absolute value of the displacement change within five sampling clocks has not exceeded the physical recognition lower limit corresponding to two pulse steps of the position encoder, the system immediately blocks the execution command of this round of division and automatically bypasses the extraction of the local stiffness history backup of the previous effective sampling cycle as the background replacement, thereby eliminating the potential calculation risk of control output divergence induced by the divisor being zero.

[0045] Example 6: In a working condition where the accuracy of feedback data from the controlled assembly execution end is limited due to timing fluctuations in the distributed network, in order to eliminate the impact of sampling jitter generated by the industrial fieldbus interface unit in the data exchange link on the calculation of the statistical variance of the force feedback signal, the control system performs a timing calibration procedure before system deployment. This is achieved by recording the clock synchronization residual of the industrial fieldbus under no-load communication conditions. According to the sampling period Calculate the bus jitter factor and use it to determine the length of the sliding time window. ,length The value of is the smallest positive integer that allows the observation window to cover at least 5 consecutive communication cycles. This calibration procedure is used to anchor the observation depth of the statistical criterion to the physical characteristics of the communication link.

[0046] In the initial deployment phase of the control system, during the topology switch from the contact transient logic masking procedure to the steady-state pressing condition, in order to adapt to the structural damping characteristics of the controlled assembly execution end and avoid transient overshoot during control handover, the core logic scheduling unit uses the comprehensive disturbance estimate latched by the third-order linear extended state observer. Inertia parameters of the controlled assembly actuator Calculate the damping weight factor ,in, For inertia parameter, As a damping weighting factor, the system uses the command change rate constraint operator to weightedly allocate control power from the open-loop feedforward compensation topology to the closed-loop feedback topology. After continuously comparing the displacement feedback signal with the distance sequence estimated by the observer and determining that the step component of the distance sequence has disappeared, the register locking action is canceled to complete the restoration of the external feedback channel. In the process of executing the above-mentioned damping weighting factor calculation, the logic control system retrieves the absolute value of the latched comprehensive disturbance estimate and divides it by the inertia parameter to extract the acceleration mapping parameter characterizing the instantaneous mechanical impact intensity. The arithmetic unit performs division normalization processing on this parameter and the internally preset system critical damping constant. The percentile coefficient of the result is directly based on the damping weighting factor, which exists in the form of a single-valued scalar. Then, through this quantitative scalar, the system intervenes in real time and regulates the mixed proportion of control quantities between the feedforward compensation channel and the closed-loop proportional channel during the closed-loop recovery process.

[0047] In the sensitivity mapping operation of the intelligent assembly logic control system for radiators, in order to determine the preset fluctuation threshold used to judge contact transients... The control system operates under a dynamic benchmark calibration procedure. The core logic scheduling unit drives the controlled assembly execution end to perform multi-rate reciprocating approximation actions under no-load conditions, utilizing the industrial fieldbus interface unit according to the sampling period. The system acquires the force feedback signal and stores it in a circular buffer. The system uses a sliding time window variance calculation program to obtain the background noise fluctuation of the force feedback signal under high-speed movement. Based on the amount of background noise fluctuation The probability distribution characteristics of the time series identify the upper limit of the inertial disturbance of the actuator, and a preset fluctuation threshold is set. Set as the upper limit of inertial disturbance and the preset gain coefficient The product value, this calibration procedure aligns the logic trigger point of the contact determination with the physical noise floor of the assembly mechanism in situ. In the initial stage of deploying the control system on the non-standard heat sink assembly line, in order to match the structural dissipation characteristics of fins of different materials and determine the preset masking window boundary. The system operates a transient energy decay identification procedure. After triggering the contact transient judgment, the controlled assembly execution end maintains the locked state of each register in the control state observation unit. The core logic scheduling unit monitors the envelope change trend of the second-order differential residual of the back electromotive force in real time and calculates the time step of the residual signal returning from the peak to the steady-state background noise range. Determine the boundaries of the preset masking window Preset masking window boundaries The calculation formula is as follows: ,in, This represents the number of sampling steps for residual attenuation. For the sampling period, this procedure couples and maps the masking duration of the logic control system with the physical energy dissipation process of the assembly interface.

[0048] The embodiments of this application have been described above with reference to the accompanying drawings. Unless otherwise specified, the embodiments and features in the embodiments of this application can be combined with each other. This application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit of this application and the scope of protection of this invention, and all of these forms are within the protection scope of this application.

Claims

1. A logic control system for intelligent assembly of a radiator, characterized in that, include: The industrial fieldbus interface unit is used to connect to the controlled assembly execution end to synchronously acquire force feedback signals, displacement feedback signals, and phase current data that characterize the operating status of the controlled assembly execution end. The control state observation unit, connected to the industrial fieldbus interface unit, is used to calculate the comprehensive disturbance estimate of the controlled assembly execution end based on the force feedback signal and displacement feedback signal. The core logic scheduling unit is connected to the industrial fieldbus interface unit and the control status observation unit respectively. The core logic scheduling unit is used to monitor the statistical variance of the force feedback signal within the sliding time window, and when the statistical variance exceeds the preset fluctuation threshold, it is determined to be a contact transient, and then the contact transient logic masking program is started. The contact transient logic masking procedure includes: using variable latches to maintain the cumulative error value of the integral branch in the previous sampling period before the contact transient determination time, while switching the control architecture from a closed-loop feedback topology to an open-loop feedforward compensation topology based on the comprehensive disturbance estimate, and using a preset command change rate constraint operator to limit the change rate of the control command within a preset slope range; after the current running clock crosses the preset masking time window boundary, the core logic scheduling unit is used to restore the external feedback channel of the controlled assembly execution end.

2. The logic control system for intelligent assembly of a radiator according to claim 1, characterized in that, The core logic scheduling unit has a built-in back EMF observer, which is used to calculate the second-order differential residual of back EMF, which characterizes the rate of change of mechanical impedance, based on the phase current data. When the core logic scheduling unit determines that the absolute value of the second-order differential residual of back EMF is greater than the preset physical contact limit parameter, it triggers the contact transient logic masking program and switches the proportional gain coefficient to the preset damping control parameter to adaptively offset the drive frequency of the controlled assembly execution end.

3. The logic control system for intelligent assembly of a radiator according to claim 2, characterized in that, The control state observation unit adopts a linear state prediction model to convert the nonlinear characteristics of the controlled assembly execution end into a comprehensive disturbance estimate. Under the open-loop feedforward compensation topology, the comprehensive disturbance estimate is superimposed on the control output of the core logic scheduling unit in real time, thereby realizing the logical decoupling of spatial motion parameters and mechanical feedback parameters.

4. The logic control system for intelligent assembly of a radiator according to claim 1, characterized in that, The core logic scheduling unit uses a variable structure control algorithm to generate control commands. The variable structure control algorithm includes a time-varying gain nonlinear scheduling law. During the process of the controlled assembly execution end approaching the heat sink component, the gain is switched at a preset slope to suppress the pulsation of the output command.

5. The logic control system for intelligent assembly of a radiator according to claim 1, characterized in that, The industrial fieldbus interface unit uses the EtherCAT protocol for periodic synchronous communication. It reads the underlying drive data at a sampling period of no more than 1ms and uses a message check operator to filter out the transition noise in the displacement feedback signal. The servo driver baseboard contains an independent oscillator hardware sampling circuit to obtain the phase current oversampling sequence at a rate three orders of magnitude higher than the bus frequency. It uses a high-speed comparator to extract the edge transition characteristics of the sequence and transmits them across the clock domain to the main memory buffer so that the core logic scheduling unit can capture the state transition signal in microseconds.

6. The logic control system for intelligent assembly of a radiator according to claim 1, characterized in that, The control state observation unit includes a residual audit module, which is used to calculate the Euclidean distance between the estimated state value and the sensor feedback value, and output a state reset command to the core logic scheduling unit when the Euclidean distance exceeds a preset deviation threshold.

7. The logic control system for intelligent assembly of a radiator according to claim 6, characterized in that, The core logic scheduling unit contains a finite state machine architecture for logical arbitration between high-speed displacement mode, contact transient logic masking procedure and constant force contact mode, and reinitializes the control architecture when a state reset command is received.

8. The logic control system for intelligent assembly of a radiator according to claim 1, characterized in that, After exiting the contact transient logic masking procedure, the core logic scheduling unit guides the system state variables back to the closed-loop feedback topology through exponential approach logic, and uses the linear release of the integral branch state to compensate for the initial deformation deviation of the physical contact interface.

9. The logic control system for intelligent assembly of a radiator according to claim 1, characterized in that, The logic control system interacts concurrently with multiple controlled assembly execution terminals through an industrial fieldbus interface unit to achieve high-frequency collaborative assembly of radiator components across multiple task positions.

Images