Intelligent temperature management method and system for motor of cutting machine based on load prediction

Abstract

The invention discloses a cutting machine motor intelligent temperature management method and system based on load prediction, relates to the technical field of cutting machine motor temperature management, and is used for solving the problems of temperature response lag and local overheating under the conditions of complex tracks and load fluctuation. In the operation process of the cutting machine, three-phase stator side current signals of the motor are continuously captured, a current vector input quantity is obtained through coordinate transformation and input into the sliding mode observer to estimate the rotor position and the motor rotating speed, a load trend quantity is generated, combined compensation modulation is carried out, and a compensation value is output to a power output link to dynamically adjust the duty ratio. According to path analysis, marking a hot stacking sensitive path section, calculating a power output change rate and shielding wind flow disturbance intensity, screening out a thermal response to-be-detected section, collecting data through a thermosensitive channel to construct a regional thermal risk index, executing differential air cooling adjustment and power modulation path adjustment, and recording state data in real time; and fine management of the full-working-condition temperature of the motor of the cutting machine is realized.

Description

Technical Field

[0001] This invention relates to the field of temperature management technology for cutting machine motors, and more specifically, to a method and system for intelligent temperature management of cutting machine motors based on load prediction. Background Technology

[0002] Existing methods for managing cutting machine motor temperature mostly rely on real-time readings from a single temperature sensor. They trigger air cooling or reduce output power when a preset threshold is exceeded. This passive response mechanism exhibits significant lag, especially under multi-segment cutting paths and complex load fluctuations, easily leading to localized overheating or insufficient air cooling. Traditional control strategies typically rely solely on instantaneous motor current or speed, lacking analysis of load variation trends and spatial cooling shielding, and failing to proactively identify and address heat accumulation. When the cutting path involves oblique entry, low-speed retraction, or clamp shielding, localized heat dissipation from the motor housing is limited, and conventional temperature sensors cannot provide comprehensive coverage, resulting in abnormal temperature accumulation in localized areas. Existing speed control logic typically only functions under steady-state or simple operating conditions. Under high-frequency load fluctuations and alternating path characteristics, the speed control response becomes disconnected from load changes. Furthermore, existing temperature control systems lack the ability to summarize and utilize historical operating conditions, failing to predict and optimize current adjustments based on past temperature and load records, and unable to establish mapping relationships between different path segments, resulting in a simplistic assessment of thermal risks.

[0003] To address the above problems, this invention proposes a solution. Summary of the Invention

[0004] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide a method and system for intelligent temperature management of cutting machine motors based on load prediction, in order to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: In a preferred embodiment, it includes: The three-phase stator current signal of the motor is continuously acquired through the current sampling path. The current vector input is obtained through coordinate transformation and then input into the sliding mode observer to estimate the rotor position and motor speed. The current speed deviation is calculated and the load trend is generated. The load trend is input into the compensation function for modulation, and the compensation value is output by combining the saturation function and the modulation gain. The compensation value is then sent to the power output link to dynamically adjust the PWM duty cycle. Analyze the path instructions of the CNC system, mark the path segments sensitive to heat accumulation, calculate the rate of change of power output and the intensity of shielding airflow disturbance, form the thermal response test area and establish the regional thermal hazard index; The temperature hysteresis adjustment function is calculated based on the regional thermal hazard index and historical data. Differentiated air-cooling regulation and power modulation path adjustment are performed, and relevant status data is recorded and written to the buffer.

[0006] In a preferred embodiment, during the operation of the cutting machine, the three-phase stator current signal of the motor is continuously captured and the coordinate transformation is completed to obtain the current vector input. The real-time differential operation between the speed command value and the current speed value is calculated to generate the current transient speed deviation of the motor. The current transient speed deviation is combined with the state variables constructed in the sliding diaphragm observer to generate the load trend.

[0007] In a preferred embodiment, the absolute value of the current transient speed deviation and the normalized state term in the sliding mode variable are used as input variables, and the output compensation value is sent to the power output link to dynamically adjust the duty cycle. Meanwhile, thermistor sensors are deployed at the stator core of the motor, in the area where the main drive transistor of the power converter is located, and on the outer surface area corresponding to the heat dissipation path of the motor housing. The thermistor channels are periodically scheduled to collect temperature values ​​and complete calibration correction.

[0008] In a preferred embodiment, the embedded path parsing unit is invoked to parse the task path segment by segment, and the tangential feed path segment, the oblique advance and retreat path segment, and the low-speed retreat path segment are marked as heat accumulation sensitive path segments based on the vector direction of the path segment, the trend of feed speed change and the tool entry method. This thermally sensitive path segment is included in the thermal response candidate segment set. The rate of change of power output and the rate of change of the intensity of shielding wind disturbance are calculated, fused into a risk scoring function and compared with the risk scoring threshold. Segments with significant risk status are extracted as thermal response test segments.

[0009] In a preferred embodiment, the sampling channel of the thermal matrix corresponding to the thermal response test section is activated, and the load and temperature rise comparison records of the same working condition task are called to construct a historical comparison dataset. The residual curve of the cooling process is generated and the temperature hysteresis rate and spatial shading ratio are extracted to construct the thermal inertia residual coefficient of the current thermal response test section.

[0010] In a preferred embodiment, the absolute value of the temperature rise rate collected in the current thermal response test section is used to determine the regional thermal hazard index, construct a thermal sensing curve, and obtain the temperature monitoring records and air-cooled duty cycle control logs of the corresponding historical tasks to form actual temperature control response time series data. The actual temperature control response time series data is compared with the thermal sensing curve to generate a temperature hysteresis adjustment function. The cooling control signal and power modulation path are adjusted according to the temperature hysteresis adjustment function. The temperature change trend curve during the cooling period is extracted and combined with the structural shielding intensity parameter and the air cooling duty cycle time series record to construct a thermal residual feature data vector group. The thermal residual feature data vector group is input into the thermal risk assessment function to generate risk level labels. All thermal response test sections are sorted and the section with the highest risk level is marked as the high residual risk thermal response test section.

[0011] In a preferred embodiment, the temperature value sequence, air-cooled duty cycle sequence, and timestamp sequence of the high residual risk thermal response test section are read. The cooling rate gradient parameter, air-cooled response suppression threshold, and power duty cycle slope change parameter are calculated. Temperature rise offset records and air-cooled response mismatch events are extracted to obtain thermal runaway trend assessment values ​​and determine risk intervals. Sections that meet the conditions are marked as severe thermal runaway trend states, and the upper limit gain of fan duty cycle, power modulation window shrinkage ratio, and temperature control start advance are output.

[0012] In a preferred embodiment, in the set of thermal response test areas, thermal response test areas that meet the conditions are selected as candidate sets of secondary thermal control early warning areas according to the risk level. The standard deviation of cooling rate, maximum temperature rise fluctuation and air-cooled response reversal delay of each secondary thermal control early warning area are calculated and merged into a thermal sensitivity parameter group and the temperature control confidence level is calculated. All secondary thermal control early warning sections are sorted according to temperature control confidence to form a secondary thermal control early warning pool. During the operation of the main control path section, the air-cooling regulation behavior and temperature rise feedback trend are monitored. When an abnormal trend is detected, the thermal response test section with the highest confidence is called from the secondary thermal control early warning pool and its corresponding control parameter set is loaded to replace the original main control path section.

[0013] In a preferred embodiment, the temperature values ​​of the motor stator core, the heat source area of ​​the power device, and the heat dissipation surface of the housing are compared to determine the stable temperature operating zone, the temperature warning zone, or the temperature over-limit zone, and the corresponding power and air cooling control are executed; the closed-loop speed regulation, the variable structure sliding mode observer, or the fuzzy parameter adjustment channel are selected according to the operating conditions, and the control path is switched in real time; the current control status data is recorded; the parallel enhancement structure and buffer circuit of the power output are enabled; and the high-frequency observation channel of the sliding mode observer is enabled.

[0014] In a preferred embodiment, the module includes: a current sampling and load trend generation module, a compensation value modulation and adjustment module, a path analysis and thermal response segment identification module, and a temperature hysteresis calculation and differential adjustment module, with signal connections between the modules. The current sampling and load trend generation module is mainly used to continuously acquire the three-phase stator current signal of the motor through the current sampling path, obtain the current vector input through coordinate transformation, and input it into the sliding mode observer to estimate the rotor position and motor speed, calculate the current speed deviation and generate the load trend. The compensation value modulation and adjustment module is mainly used to input the load trend into the compensation function for modulation, combine the saturation function and modulation gain to output the compensation value, and send the compensation value into the power output link to perform dynamic adjustment of the PWM duty cycle; The path analysis and thermal response section identification module is mainly used to analyze the path instructions of the CNC system, mark the thermal accumulation sensitive path sections, calculate the power output change rate and the intensity of shielding airflow disturbance, form the thermal response test section and establish the regional thermal hazard index. The temperature hysteresis calculation and differential adjustment module is mainly used to calculate the temperature hysteresis adjustment function based on the regional thermal hazard index and historical data, perform differential air-cooling adjustment and power modulation path adjustment, and record relevant status data into the buffer.

[0015] The technical effects and advantages of the intelligent temperature management method and system for cutting machine motors based on load prediction of this invention are as follows: This invention establishes a collaborative relationship between the current sampling path, sliding mode observer, and path analysis unit, enabling the motor to generate load trend quantities and perform pre-compensation modulation under different trajectories and load fluctuations. By marking heat accumulation-sensitive path segments and combining them with historical cooling residual curves, a regional thermal hazard index is constructed to achieve pre-identification of motor thermal response segments. Through differentiated air-cooling regulation and power modulation path adjustment, the temperature control link can adapt in real time during path switching and load changes. Compared with traditional schemes based on single-point temperature triggering, this invention introduces multi-point temperature acquisition, historical data comparison, and risk scoring functions, enabling specific adjustments and alternative path calls for high residual risk segments, ensuring synchronized management of power and temperature control even under complex path and uneven cooling conditions. Attached Figure Description

[0016] Figure 1 This is a timing diagram of the intelligent temperature management method and system for cutting machine motors based on load prediction, as presented in this invention.

[0017] Figure 2 This is a schematic diagram of the intelligent temperature management method and system module for cutting machine motors based on load prediction, as described in this invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Example This invention discloses a method for intelligent temperature management of cutting machine motors based on load prediction, such as... Figure 1 As shown, it includes: The three-phase stator current signal of the motor is continuously acquired through the current sampling path. The current vector input is obtained through coordinate transformation and input into the sliding mode observer to estimate the rotor position and motor speed. The current speed deviation is calculated and the load trend is generated.

[0020] Among them, the sliding mode observer adopts a simplified state equation based on the two-axis coordinate system of the motor. The method of its establishment is as follows: the components of the stator current in the direct axis and quadrature axis directions are used as state variables. A set of first-order differential equations is established based on the equivalent circuit parameters of the motor. Sliding mode observation is achieved by adding a switching gain term to the equations, and the estimated values ​​of rotor position and motor speed are output in real time.

[0021] The load trend is input into the compensation function for modulation, and the compensation value is output by combining the saturation function and the modulation gain. The compensation value is then sent to the power output link to dynamically adjust the PWM duty cycle.

[0022] The path instructions of the CNC system are analyzed, the path segments sensitive to heat accumulation are marked, the rate of change of power output and the intensity of shielding airflow disturbance are calculated, the thermal response test area is formed and the regional thermal hazard index is established.

[0023] The temperature hysteresis adjustment function is calculated based on the regional thermal hazard index and historical data. Differentiated air-cooling regulation and power modulation path adjustment are performed, and relevant status data is recorded and written to the buffer.

[0024] During the operation of the cutting machine, in order to establish a characteristic channel that can be used to dynamically characterize the actual load state changes of the motor, the current signal of the three-phase stator side of the motor is continuously captured through the current sampling path configured in the embedded sampling controller. The current sampling path is based on a shunt current sensor and a synchronous rectification circuit, and the sampled data is sent to the data receiving buffer of the control host according to a fixed refresh frequency.

[0025] After initial sampling, the three-phase stator current is mapped in the coordinate domain using the Clark and Park coordinate transformation algorithms to obtain the direct-axis and quadrature-axis current components, thereby constructing a current vector input that is easy for subsequent observer processing.

[0026] Subsequently, the obtained current vector input is input into the sliding mode observer. Based on the nonlinear state equation of the motor mathematical model, the sliding mode observer performs a fast estimation of the rotor flux state and derives the rotor position and motor speed information at the current moment.

[0027] The state approximation strategy used in the sliding mode observer relies on a switching gain adjustment mechanism to ensure that robust speed estimation results can still be output even under the conditions of modeling errors and disturbances.

[0028] Next, the estimated current speed value is compared with the speed command value issued by the CNC system in real time to generate the current transient speed deviation of the motor.

[0029] Based on the aforementioned current transient speed deviation, the current transient speed deviation is further combined with the state variables constructed in the sliding diaphragm observer to generate a load trend quantity characterizing the current load evolution trend of the motor. This load trend quantity not only considers the magnitude of the current speed error but also incorporates the response rate of the control state variables to sudden disturbances, reflecting whether the motor is in a dynamic critical stage before entering a sudden load increase or acceleration response.

[0030] Furthermore, to improve the response speed to sudden load disturbances and intervene in adjustment before significant temperature rise accumulates, a combined compensation function is used to modulate the load trend. Specifically, the absolute value of the current transient speed deviation and the normalized state term in the sliding mode variable are used as input variables and fed into a set of predefined saturation function models. This saturation function uses bilateral limiting to converge and control abnormal input changes. Simultaneously, a set of variable scaling factors is introduced into the compensation function as a dynamic modulation gain, which adaptively adjusts according to the rate of change of the current trend.

[0031] A saturation function model refers to a function structure that performs bidirectional limiting on the input variable in the compensation function. This is used to prevent the adjustment quantity from having excessive output when there are sudden load fluctuations or abnormal conditions. Specifically, based on preset upper and lower limits, the absolute value of the input variable is judged. When the input value is within the limit range, the output remains linear; when the input value exceeds the limit range, the output is fixed at the upper or lower limit.

[0032] The normalized state term refers to standardizing the current state variables obtained from the sliding mode observer according to a pre-set reference amplitude, thus limiting the numerical range of the state variables to a fixed proportion. This dimensionless processing ensures that the output characteristics of the compensation function remain consistent under different loads and current levels, avoiding imbalances in adjustment sensitivity caused by excessively large or small state values.

[0033] The compensation value output by the compensation expression is then fed into the voltage modulation stage in the power output link. By dynamically adjusting the PWM duty cycle, the actual output voltage amplitude is controlled, thereby completing the pre-adjustment of the output power reduction operation in advance.

[0034] After the load prediction control logic has stabilized, high-response-accuracy thermistor components are deployed at key thermal locations. Specifically, precision-grade thermistor sensors are selected and mounted at the motor stator core, the PCB area where the main drive transistor of the power converter is located, and the outer surface area corresponding to the heat dissipation path of the motor housing.

[0035] The three locations mentioned above represent the core area of ​​the heat source, the high-loss point of power conversion, and the heat dissipation path outlet area, respectively, which can cover the typical thermal path during motor operation. Then, an equally spaced timing sampling channel is established through on-chip timing control logic. The controller periodically schedules the three thermal channels to collect temperature values ​​and completes calibration correction, ultimately forming a basic dataset for three-point temperature monitoring, including the heat source, heat conduction path, and heat dissipation outlet.

[0036] It should be noted that when the cutting machine performs multi-segment oblique cutting and enters the retraction phase, the motor-driven load exhibits a non-periodic drop trend due to changes in the processing path direction and instantaneous feed speed adjustments. This causes the load prediction channel to continuously output low-power operation judgment signals. At this time, the control system assumes the motor is in a non-heating state, suspends the high-speed operation of the cooling fan, and maintains the output duty cycle within the normal range. However, in the oblique cutting path, the cooling air channel is blocked by the workpiece positioning structure, and heat diffusion is limited in the area of ​​the motor housing near the mounting bracket, forming a small area of ​​thermal inertia accumulation. Since no thermal sensing point is placed in this area, and the load prediction channel continuously judges it as a low-load state in the retraction path, it skips the power adjustment and fan strong cooling command. This causes local overheating in this thermal resistance area, and due to incomplete cooling, it accumulates into a high-temperature starting point, causing temperature control lag and over-temperature protection delay. Therefore, in this embodiment,

[0037] After the load prediction control logic has reached stable operation, when the cutting machine executes a composite trajectory task containing multiple oblique cutting and retraction paths, it first calls the embedded path parsing unit based on the known G-code trajectory information and tool strategy file of the CNC system, extracts the motion instructions in the task path segment by segment, and performs trajectory parsing.

[0038] Specifically, based on the vector direction of the path segment, the trend of feed rate change, and the tool entry method, the following three types of path segments with potential heat accumulation risks are identified: the first is the tangential feed path segment, characterized by continuous contact between the tool and the workpiece and the absence of a direct airflow channel; the second is the oblique advance and retreat path segment, characterized by the movement direction being deviated from the machine housing axis and the presence of shielding; and the third is the low-speed retreat path segment, characterized by a continuous decrease in feed rate, a reduction in cutting load, but a weakening of cooling.

[0039] Furthermore, the three types of path segments mentioned above are marked as heat accumulation sensitive path segments, and their relative azimuth information in the motor housing coordinate system and the spatial identification of the corresponding possible shielding structures such as mounting brackets, clamp positioning blocks, etc. are recorded.

[0040] Next, the marked heat accumulation sensitive path segments are included in the thermal response candidate segment set. For each thermal response candidate segment, two-dimensional index calculations are performed: the first dimension is the power output change rate, which is extracted by reading the power estimation value sequence output in real time in the load prediction and performing first-order difference calculation; the second dimension is the change rate of shielding airflow disturbance intensity, which is calculated by matching the shell azimuth angle of the thermal response candidate segment with the airflow shielding coefficient pre-stored in the cooling airflow simulation model and combining it with the fan operating conditions.

[0041] Subsequently, the aforementioned power output change rate and the rate of change of shielding airflow disturbance intensity are fused into a risk scoring function and normalized. A risk scoring threshold is set, and the risk scoring function value obtained after fusion is compared with the risk scoring threshold. When the risk scoring function value is less than the risk scoring threshold, it indicates that the risk status of the candidate thermal response segment is not significant; when the risk scoring function value is greater than or equal to the risk scoring threshold, it indicates that the risk status of the candidate thermal response segment is significant. This type of candidate thermal response segment is then extracted separately as the thermal response test segment, forming a set of thermal response test segments.

[0042] Meanwhile, to enhance the accuracy of temperature rise trend judgment, in the multi-point thermal matrix deployed outside the motor housing, only the sampling channels of the possible shading azimuth areas corresponding to each thermal response test section in the current thermal response test area set are activated. In the load prediction history associated with each thermal response test section, the load and temperature rise comparison records of the most recent N identical working conditions are called to construct a historical comparison dataset containing a three-dimensional structure of power output, air cooling duty cycle and local temperature sequence. The dataset is then aligned with a unified time axis to construct the cooling process residual curve experienced by the current path segment in the historical task.

[0043] Next, based on the generated historical cooling residual curves, the following two key indicators are extracted: one is the temperature hysteresis rate, which refers to the proportion of time offset between the temperature rise response and the load decrease in this path segment; the other is the spatial shading ratio, which refers to the cumulative proportion of time during which the cooling airflow is obstructed in this path segment during the mission. The spatial shading ratio and temperature hysteresis rate are then used as inputs to construct the thermal inertia residual coefficient of the current thermal response test segment.

[0044] Furthermore, by combining the absolute value of the temperature rise rate collected in real time by the thermal channel in the current thermal response test section, the thermal inertia residual coefficient is coupled with the absolute value of the temperature rise rate through a function mapping mechanism to determine the regional thermal hazard index, which is used to comprehensively describe the thermal runaway potential of the thermal response test section under the current operating conditions.

[0045] After calculating the regional thermal hazard index for the thermal response test section, to further identify the discrepancy between the predictive model of temperature control behavior and the actual execution behavior, thermal sensing curves are constructed for each thermal response test section based on the calculated regional thermal hazard index. The thermal sensing curve is composed of the residual thermal inertia coefficient and the current temperature rise rate, and is expressed in time series form as the potential temperature rise evolution trend of the path section under specific load and shading conditions.

[0046] Subsequently, to determine whether the current air-cooling control link effectively responds to the thermal risk area, temperature monitoring records from historical task cycles that share the same task path, load level, and air-cooling strategy as the thermal response test segment are retrieved. By analyzing the temperature values ​​recorded by the sampling control logic in conjunction with the air-cooling duty cycle control logs, an actual temperature control response timing data corresponding to the current path is constructed. Timing alignment is then performed based on a unified timestamp.

[0047] Furthermore, the weighted residual integral calculation and predicted response deviation fitting analysis are performed on the actual temperature control response time series data and the thermal sensing curve to form a temperature lag adjustment function, which reflects whether there is a delay in the current air-cooling strategy response and can quantitatively describe the degree of deviation between the delay and the thermal sensing expectation.

[0048] When the hysteresis adjustment function indicates that the temperature control strategy has a response delay, on the one hand, the regional thermal hazard index value of the current thermal response test section is read, and the preset fan duty cycle gain function model is matched according to the index level to increase the amplitude of the PWM control signal of the cooling system; on the other hand, the response window range of the power modulation path is adjusted according to the thermal residual superposition rate of the current thermal response test section and the same section in the historical task record. Specifically, the upper limit range of the output power duty cycle variation is narrowed to reduce the heat source expansion rate, thereby forming a differentiated air-cooling adjustment strategy based on thermal risk feedback.

[0049] Furthermore, for all thermal response test sections that have undergone differentiated air-cooling regulation, after the current cooling cycle ends, the temperature change trend curves of each section during the cooling period are extracted, and the corresponding structural shielding intensity parameters and air-cooling duty cycle time-series records are simultaneously called to jointly construct the thermal residual feature data vector group for that section. The thermal residual feature data vector group includes, but is not limited to, the cooling response delay representing the time difference between the occurrence of the temperature rise inflection point and the activation of air cooling; the shielding interference amplification factor defined as the reduction factor of the air-cooling effect in the structurally shielded area compared to the unshielded area; and the cooling rate normalization index calculated based on the unit time cooling rate under unit duty cycle.

[0050] Subsequently, the thermal residual feature data vector group is input into the thermal risk assessment function. The historical lag response factor, constructed by regression of response mismatch data of the same path segment in multiple historical tasks, is used for function fitting and threshold comparison to generate the risk level label of the thermal response test segment.

[0051] The thermal risk assessment function employs a weighted summation method based on eigenvectors. It assigns weights to characteristic parameters such as cooling response delay, shielding interference amplification factor, and cooling rate normalization index, and then performs a comprehensive calculation to obtain a risk score. The risk score is compared with a preset risk threshold; sections exceeding the threshold are identified as high-risk sections.

[0052] Meanwhile, all thermal response test sections were sorted according to risk level labels, and the section with the highest risk level was selected and marked as the high residual risk thermal response test section.

[0053] Furthermore, for the thermal response test sections marked as having high residual risk, after the current cooling cycle ends, the temperature sampling records and air-cooling duty cycle control logs corresponding to the thermal response test sections with high residual risk are called up. The temperature value sequence, air-cooling duty cycle sequence, and timestamp sequence data during the cooling process of the thermal response test sections with high residual risk are read. The first-order difference calculation is performed on the temperature value sequence, and the temperature drop rate change curve under unit duty cycle is fitted by combining the air-cooling duty cycle sequence to extract the cooling rate gradient parameter. Synchronously read the record of air-cooled response change in the high residual risk thermal response test section within the current cycle, compare the actual temperature response with the gain output amplitude according to the linear gain function set in the fan drive control logic, calculate the maximum ineffective gain range of air-cooled response, and set the air-cooled response suppression threshold for the current high residual risk thermal response test section. Next, the power duty cycle modulation sequence is invoked in the power modulation path, and its slope change value in the most recent two refresh cycles is calculated as the output power duty cycle slope change parameter. Then, the unique identifier key of the current high residual risk thermal response test section is used to match the historical execution records under the same task type, and the temperature rise offset records and air-cooled response mismatch events of the corresponding path segment are retrieved. All offset times and mismatch event descriptors are extracted as historical anomaly feature inputs.

[0054] Furthermore, the aforementioned cooling rate gradient parameters, air-cooled response suppression threshold, slope change value, offset time, and mismatch event descriptor are uniformly assembled into a structural response fusion feature vector. This feature vector is then input into a multivariate ridge regression model already loaded in the control host. The trained model outputs a thermal runaway trend assessment value, and a thermal runaway threshold is set. Based on the thermal runaway threshold, the risk interval to which the current high residual risk thermal response test section belongs is determined.

[0055] Specifically, when the thermal runaway trend assessment value is greater than or equal to the thermal runaway threshold, the current high-residual-risk thermal response test section is marked as a severe thermal runaway trend state. Combining the shading interference factor and cooling rate normalization index in the thermal residual feature vector group to which the current section belongs, bottleneck classification logic is executed to determine the cooling control bottleneck type label. Subsequently, the task load cycle, expected power recovery interval, and shading structure change sequence corresponding to the high-residual-risk thermal response test section are read in the path scheduling logic. This data, along with the cooling control bottleneck type label and the current temperature control priority level, is input into the differentiated temperature control generator. The differentiated temperature control generator outputs the following adjustment parameters based on the mapping rule base:

[0056] Fan duty cycle upper limit gain: The adjustment level is set according to the cooling control bottleneck type label and the current temperature control priority level; Power modulation window contraction ratio: Calculate the current upper limit compression ratio of the sliding window based on the thermal runaway trend assessment value; Temperature control activation lead time: Combine the task load cycle and the shading structure change sequence to predict the shading formation time and activate temperature control in advance; Subsequently, in the set of thermal response test areas, all sections that have undergone thermal runaway risk assessment are screened. Based on risk level, main control path sections marked as high residual risk are removed, and the top k thermal response test sections that meet the following conditions are selected as the candidate set of secondary thermal control early warning sections:

[0057] The risk level corresponding to the route segment belongs to the second-highest level range; The residual for assessing the thermal runaway trend is less than or equal to the risk score threshold; Furthermore, for each selected secondary thermal control early warning zone, the following parameter calculation process is performed: Within the most recent complete task cycle, extract the temperature drop curve of the secondary thermal control early warning section, calculate the standard deviation of its first derivative, and obtain the standard deviation of the cooling rate of the secondary thermal control early warning section. Extract the temperature difference between the highest and lowest points from the original temperature rise curve to obtain the maximum temperature rise fluctuation in the secondary thermal control early warning zone; Record the time interval between the air-cooling activation point and the first inflection point of the temperature rise curve to obtain the air-cooling response reversal delay in the secondary thermal control early warning section.

[0058] Furthermore, the cooling rate standard deviation, maximum temperature rise fluctuation, and air-cooled response reversal delay of the secondary thermal control early warning section are combined to form a thermal sensitivity parameter group. Each thermal sensitivity parameter group is then input into the thermal control confidence calculation function to calculate the temperature control confidence, which represents the effective response capability of the path segment as an alternative control path in the current control task.

[0059] Next, all the top k thermal response test sections are sorted in descending order according to their respective confidence levels to form a secondary thermal control early warning pool. During the operation of the main control thermal response link, the air-cooling regulation behavior and temperature rise feedback trend of the current main control path segment, i.e., the high residual risk thermal response test section, are monitored in real time. Specific detection indicators include: Air cooling adjustment is ineffective: The current duty cycle change exceeds the set gain threshold and the temperature rise curve continues to rise; Temperature rise continues to deviate: The temperature has not entered the stable region for NNN consecutive refresh cycles; Severely inaccurate load forecasting: The difference between the current load trend and the actual output power exceeds the set mismatch ratio.

[0060] When any of the above-mentioned abnormal trends are triggered, the thermal response test section with the highest confidence level is immediately called from the secondary thermal control early warning pool, and its corresponding set of control parameters is loaded, including control quantities such as air-cooled duty cycle gain, power adjustment window, and start-up advance. Its thermal control scheduling table is written into the current master control task table, overwriting the original master control path segment, ensuring that the temperature control link has continuous verification and failure replacement capabilities. At the same time, it realizes early identification of temperature rise changes, pre-activation of air-cooled adjustment, and adaptive evolution of temperature control execution path, ensuring that the accuracy and timeliness of temperature control response can still be maintained under conditions of uneven load fluctuation and spatial cooling imbalance.

[0061] After the temperature sampling channel completes the data acquisition for the current cycle, a unified comparison logic is executed on the collected temperature values ​​of the motor stator core, power device heat source area, and housing heat dissipation surface. These values ​​are compared with the lower and upper limits of the temperature threshold in the temperature grading judgment rule table to construct a three-segment graded response region: If the collected temperature values ​​of the motor stator core, the heat source area of ​​the power device, and the heat dissipation surface of the housing are all lower than the lower limit of the temperature threshold in the temperature classification judgment rule table, it is determined to be in the temperature stable operation zone. At this time, the output power channel is maintained in full output state, and the duty cycle of the air-cooled channel is maintained at the basic heat dissipation value. If any temperature value of the motor stator core, the heat source area of ​​the power device, or the heat dissipation surface of the housing is greater than or equal to the lower limit of the temperature threshold in the temperature classification judgment rule table and less than the upper limit of the temperature threshold in the temperature classification judgment rule table, it is judged as a temperature warning zone. The output power channel executes a linear decreasing strategy, and the output value is reduced relative to the degree of temperature deviation by a fixed proportion. At the same time, the fan drive channel increases the PWM duty cycle to match the heating speed. If any of the collected temperature values ​​of the motor stator core, the heat source area of ​​the power device, and the heat dissipation surface of the casing meets or exceeds the upper limit of the temperature threshold in the temperature grading judgment rule table, it is determined to be in the temperature over-limit zone. The motor PWM output is immediately disconnected, the active cooling strategy is activated, and the fan duty cycle is pushed to the maximum value until the temperature drops to the warning zone.

[0062] Simultaneously, to ensure precise matching between power output response and load conditions, a three-path condition discrimination mechanism is employed in the speed control logic. Specifically:

[0063] First path: When the motor enters the no-load stable operation state, the controller calls the static parameter setting table, reads the proportional coefficient and integral coefficient, and inputs the difference between the speed command and the actual estimated speed into the traditional PI closed-loop speed regulation to construct the output control quantity, which is used to adjust the power output; Second path: When the absolute value of the speed deviation of the controller shows an increasing trend after two consecutive cycles, and the load trend is judged to be in an upward state, the speed regulation path is immediately switched to the variable structure sliding diaphragm observer structure, the sliding mode control law is enabled to calculate the output regulation, and the current limit value is expanded to improve the high load transient response tolerance. The third path: When the motor is in the continuous cutting stage, the controller reads the current current fluctuation trend, calls the fuzzy parameter adjustment channel, and adjusts the instantaneous values ​​of the proportional coefficient and integral coefficient according to the preset fuzzy control rule set to construct a PI parameter pair that adjusts in real time with load changes, so that the speed regulation response curve always matches the current load characteristics.

[0064] Furthermore, based on the shape of the sampled current curve and the load trend judgment within the current cycle, a speed regulation path selection function is executed to switch the control path in real time, ensuring that the motor speed remains within the allowable error range under different load fluctuation conditions. Specifically, to ensure the traceability and subsequent debugging capability of the control strategy, a structured status data record structure is established in the control host, using the control cycle timestamp as the primary index key, and the following status points are encoded and written:

[0065] Current speed regulation path mode number; Current cycle output power value; Real-time speed error; Current temperature range level; Sliding mode state variables; Current limiting is active.

[0066] All data is recorded to a fixed-length buffer, and a circular overwrite mechanism is used to automatically replace the oldest record, while retaining the data from the most recent N cycles for export via the external debugging interface. The export interface is based on the standard UART communication protocol and supports asynchronous data readback triggered by host control commands.

[0067] At the power modulation and drive output level, the parallel enhancement structure of the power output is activated, switching the power channel to a parallel dual-drive bridge arm mode to improve the maximum output capability. A set of transient absorption units is connected in series at the output node of the drive stage. The structure consists of an RC snubber circuit composed of selected resistors and capacitors to absorb the peak energy formed by the sudden rise in bridge arm voltage, avoiding impact on the power devices.

[0068] Meanwhile, a high-frequency observation channel is enabled in the sliding mode observer to increase the internal update frequency to the level of synchronization with the control channel response frequency, so that the rotor position signal refresh cycle is precisely aligned with the controller execution cycle, realizing the real-time and phase accuracy synchronization of speed state estimation, and ensuring that the power modulation behavior is not mismatched due to signal lag.

[0069] This invention also proposes an intelligent temperature management system for cutting machine motors based on load prediction, such as... Figure 2 As shown, it includes: a current sampling and load trend generation module, a compensation value modulation and adjustment module, a path analysis and thermal response segment identification module, and a temperature hysteresis calculation and differential adjustment module, with signal connections between each module.

[0070] The current sampling and load trend generation module is mainly used to continuously acquire the three-phase stator current signal of the motor through the current sampling path, obtain the current vector input through coordinate transformation, and input it into the sliding mode observer to estimate the rotor position and motor speed, calculate the current speed deviation, and generate the load trend.

[0071] The compensation value modulation and adjustment module is mainly used to input the load trend into the compensation function for modulation, combine the saturation function and modulation gain to output the compensation value, and send the compensation value into the power output link to perform dynamic adjustment of the PWM duty cycle.

[0072] The path analysis and thermal response section identification module is mainly used to analyze the path instructions of the CNC system, mark the path segments sensitive to heat accumulation, calculate the rate of change of power output and the intensity of shielding airflow disturbance, form the thermal response test section and establish the regional thermal hazard index.

[0073] The temperature hysteresis calculation and differential adjustment module is mainly used to calculate the temperature hysteresis adjustment function based on the regional thermal hazard index and historical data, perform differential air-cooling adjustment and power modulation path adjustment, and record relevant status data into the buffer.

[0074] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0075] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0076] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and inventive constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0077] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0078] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0079] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for intelligent temperature management of cutting machine motors based on load prediction, characterized in that, include: The three-phase stator current signal of the motor is continuously acquired through the current sampling path. The current vector input is obtained through coordinate transformation and then input into the sliding mode observer to estimate the rotor position and motor speed. The current speed deviation is calculated and the load trend is generated. The load trend is input into the compensation function for modulation, and the compensation value is output by combining the saturation function and the modulation gain. The compensation value is then sent to the power output link to dynamically adjust the PWM duty cycle. Analyze the path instructions of the CNC system, mark the path segments sensitive to heat accumulation, calculate the rate of change of power output and the intensity of shielding airflow disturbance, form the thermal response test area and establish the regional thermal hazard index; The temperature hysteresis adjustment function is calculated based on the regional thermal hazard index and historical data. Differentiated air-cooling regulation and power modulation path adjustment are performed, and relevant status data is recorded and written to the buffer.

2. The intelligent temperature management method for cutting machine motors based on load prediction according to claim 1, characterized in that: During the operation of the cutting machine, the three-phase stator current signal of the motor is continuously captured and the coordinate transformation is completed to obtain the current vector input. The real-time differential operation between the speed command value and the current speed value is calculated to generate the current transient speed deviation of the motor. The current transient speed deviation is combined with the state variables constructed in the sliding diaphragm observer to generate the load trend.

3. The intelligent temperature management method for cutting machine motors based on load prediction according to claim 2, characterized in that: The absolute value of the current transient speed deviation and the normalized state term in the sliding mode variable are used as input variables, and the output compensation value is sent to the power output link to dynamically adjust the duty cycle. Meanwhile, thermistor sensors are deployed at the stator core of the motor, in the area where the main drive transistor of the power converter is located, and on the outer surface area corresponding to the heat dissipation path of the motor housing. The thermistor channels are periodically scheduled to collect temperature values ​​and complete calibration correction.

4. The intelligent temperature management method for cutting machine motors based on load prediction according to claim 3, characterized in that: The embedded path parsing unit is invoked to parse the task path segment by segment. Based on the vector direction of the path segment, the trend of feed rate change and the tool entry method, the tangential feed path segment, the oblique advance and retreat path segment and the low-speed retreat path segment are marked as heat accumulation sensitive path segments. This thermally sensitive path segment is included in the thermal response candidate segment set. The rate of change of power output and the rate of change of the intensity of shielding wind disturbance are calculated, fused into a risk scoring function and compared with the risk scoring threshold. Segments with significant risk status are extracted as thermal response test segments.

5. The intelligent temperature management method for cutting machine motors based on load prediction according to claim 4, characterized in that: Activate the sampling channel of the thermal matrix corresponding to the thermal response test section, call the load and temperature rise comparison records of the same working condition task to build a historical comparison dataset, generate the cooling process residual curve and extract the temperature hysteresis rate and spatial shading ratio, and construct the thermal inertia residual coefficient of the current thermal response test section.

6. The intelligent temperature management method for cutting machine motors based on load prediction according to claim 5, characterized in that: By combining the absolute value of the temperature rise rate collected in the current thermal response test section, the regional thermal hazard index is determined, a thermal sensing curve is constructed, and the temperature monitoring records and air-cooled duty cycle control logs of the corresponding historical tasks are obtained to form actual temperature control response time series data. The actual temperature control response time series data is compared with the thermal sensing curve to generate a temperature hysteresis adjustment function. The cooling control signal and power modulation path are adjusted according to the temperature hysteresis adjustment function. The temperature change trend curve during the cooling period is extracted and combined with the structural shielding intensity parameter and the air cooling duty cycle time series record to construct a thermal residual feature data vector group. The thermal residual feature data vector group is input into the thermal risk assessment function to generate risk level labels. All thermal response test sections are sorted and the section with the highest risk level is marked as the high residual risk thermal response test section.

7. The intelligent temperature management method for cutting machine motors based on load prediction according to claim 5, characterized in that: Read the temperature value sequence, air-cooled duty cycle sequence, and timestamp sequence of the high residual risk thermal response test section, calculate the cooling rate gradient parameter, air-cooled response suppression threshold, and power duty cycle slope change parameter, extract temperature rise offset records and air-cooled response mismatch events, obtain thermal runaway trend assessment values ​​and determine risk intervals, mark the sections that meet the conditions as severe thermal runaway trend states, and output the fan duty cycle upper limit gain, power modulation window shrinkage ratio, and temperature control start advance.

8. The intelligent temperature management method for cutting machine motors based on load prediction according to claim 7, characterized in that; In the set of thermal response test areas, thermal response test areas that meet the conditions are selected as candidate sets of secondary thermal control early warning areas according to the risk level. The standard deviation of cooling rate, maximum temperature rise fluctuation and air-cooled response reversal delay of each secondary thermal control early warning area are calculated and merged into a thermal sensitivity parameter group and the temperature control confidence is calculated. All secondary thermal control early warning sections are sorted according to temperature control confidence to form a secondary thermal control early warning pool. During the operation of the main control path section, the air-cooling regulation behavior and temperature rise feedback trend are monitored. When an abnormal trend is detected, the thermal response test section with the highest confidence is called from the secondary thermal control early warning pool and its corresponding control parameter set is loaded to replace the original main control path section.

9. The intelligent temperature management method for cutting machine motors based on load prediction according to claim 8, characterized in that: The temperature values ​​of the motor stator core, the heat source area of ​​the power device, and the heat dissipation surface of the housing are compared to determine the stable operating temperature zone, the temperature warning zone, or the temperature over-limit zone, and the corresponding power and air cooling control are executed; the closed-loop speed regulation, variable structure sliding mode observer, or fuzzy parameter adjustment channel are selected according to the operating conditions and the control path is switched in real time; the current control status data is recorded; the parallel enhancement structure and buffer circuit of power output are enabled; and the high-frequency observation channel of the sliding mode observer is enabled.

10. A cutting machine motor intelligent temperature management system based on load prediction, characterized in that, include: The module includes a current sampling and load trend generation module, a compensation value modulation and adjustment module, a path analysis and thermal response section identification module, and a temperature hysteresis calculation and differential adjustment module, with signal connections between the modules. The current sampling and load trend generation module is mainly used to continuously acquire the three-phase stator current signal of the motor through the current sampling path, obtain the current vector input through coordinate transformation, and input it into the sliding mode observer to estimate the rotor position and motor speed, calculate the current speed deviation and generate the load trend. The compensation value modulation and adjustment module is mainly used to input the load trend into the compensation function for modulation, combine the saturation function and modulation gain to output the compensation value, and send the compensation value into the power output link to perform dynamic adjustment of the PWM duty cycle; The path analysis and thermal response section identification module is mainly used to analyze the path instructions of the CNC system, mark the thermal accumulation sensitive path sections, calculate the power output change rate and the intensity of shielding airflow disturbance, form the thermal response test section and establish the regional thermal hazard index. The temperature hysteresis calculation and differential adjustment module is mainly used to calculate the temperature hysteresis adjustment function based on the regional thermal hazard index and historical data, perform differential air-cooling adjustment and power modulation path adjustment, and record relevant status data into the buffer.

Images