Motor thermal overload prediction control system based on multi-source data fusion

Abstract

This invention relates to the field of motor overload predictive control technology, specifically disclosing a motor thermal overload predictive control system based on multi-source data fusion, applicable to multiple motors. The system includes: multiple end-side fast / slow dual-track protection units, a side-side production line-level collaborative control unit, and a cloud-side global health management platform. The multiple end-side fast / slow dual-track protection units are respectively configured on the multiple motors. In practical applications on continuous industrial production lines, this invention effectively balances the technical conflict between protection response speed and production continuity through a multi-level collaborative architecture. When a production line motor encounters a sudden high-load impact, the end-side pure hardware architecture can intercept the destructive impact at the lowest level with extremely low latency and dynamically adjust the judgment threshold based on real-time thermal capacity margin, effectively shielding against false triggering interference from fluctuations in normal operating conditions and significantly improving the safety of the underlying control.

Description

Technical Field

[0001] This invention relates to the field of motor overload predictive control technology, specifically a motor thermal overload predictive control system based on multi-source data fusion. Background Technology

[0002] In modern continuous industrial production lines, motors are the core driving equipment. Existing thermal overload protection solutions mainly rely on single-node software algorithms or traditional thermal relays to simulate temperatures and determine fixed thresholds. Such existing systems typically only focus on the independent operating state of a single motor, and their data acquisition and processing dimensions are singular, with relatively fixed and mechanical execution logic.

[0003] When faced with complex operating conditions of high-frequency dynamic fluctuations at the production line level, if a strict fixed trigger threshold is set to prevent the motor from burning out instantly, the inherent processing delay of conventional software algorithms is very likely to cause malfunctions under transient high load impacts, directly cutting off the power supply of a single motor, thereby causing unplanned shutdowns and serious capacity losses of the entire continuous production line.

[0004] However, if the protection threshold is relaxed in order to ensure continuous production, a single static thermodynamic model cannot accurately track the actual deep heat accumulation state of the motor. Long-term operation with defects can easily lead to irreversible aging and failure of insulation materials or even direct burnout of the motor.

[0005] This isolated protection architecture not only fails to handle both millisecond-level sudden impacts and long-term steady-state thermal evolution, but also fails to proactively mitigate crises through load coordination across the entire production line when a single device approaches its thermal limit. Summary of the Invention

[0006] This invention aims to at least partially solve one of the technical problems in related technologies. Therefore, the objective of this invention is to propose a motor thermal overload predictive control system based on multi-source data fusion to achieve long-term reliable operation of the motor.

[0007] To achieve the above objectives, a first aspect of the present invention proposes a motor thermal overload predictive control system based on multi-source data fusion, applicable to multiple motors, comprising: multiple end-side fast and slow dual-track protection units, a side-side production line-level collaborative control unit, and a cloud-side global health management platform; the multiple end-side fast and slow dual-track protection units are respectively configured on the multiple motors;

[0008] The end-side fast and slow dual-rail protection unit includes a Serdes high-speed transmission module, a hardware fast rail protection control unit, and a software slow rail prediction control unit; the Serdes high-speed transmission module collects multi-source operating data of the corresponding motor, and the multi-source operating data includes at least three-phase current data, three-phase voltage data, and vibration data.

[0009] The hardware fast track protection control unit calculates the motor winding temperature and thermal capacity margin in real time based on the multi-source operating data, and dynamically adjusts the transient current transient change rate threshold and peak current threshold of the transient protection according to the thermal capacity margin. When it is determined that the change rate or peak value of the three-phase current data exceeds the transient current transient change rate threshold or the peak current threshold, transient overload protection is executed.

[0010] The software slow track prediction control unit generates steady-state thermal prediction results by performing full-system thermodynamic modeling based on the multi-source operating data, and forms a two-way closed-loop calibration mechanism with the hardware fast track protection control unit to correct the calculation parameters of the hardware fast track protection control unit in real time.

[0011] The side-side production line-level collaborative control unit performs dynamic load scheduling for the multiple motors based on the thermal capacity margin reported by each of the end-side fast and slow dual-track protection units.

[0012] The cloud-based global health management platform performs insulation aging assessment and global parameter optimization based on the operating data of the multiple motors throughout their entire life cycle.

[0013] To achieve the above objectives, a second aspect of the present invention proposes a motor thermal overload predictive control method based on multi-source data fusion, applied to a control system comprising multiple end-side fast and slow dual-track protection units, a side-side production line-level collaborative control unit, and a cloud-side global health management platform. The multiple end-side fast and slow dual-track protection units are respectively configured for multiple motors. The method includes the following steps:

[0014] The Serdes high-speed transmission module in the end-side fast and slow dual-rail protection unit collects multi-source operating data of the corresponding motor. The multi-source operating data includes at least three-phase current data, three-phase voltage data and vibration data.

[0015] The hardware fast rail protection control unit in the end-side fast and slow dual rail protection unit calculates the motor winding temperature and thermal capacity margin in real time based on the multi-source operating data, and dynamically adjusts the transient current transient change rate threshold and peak current threshold of the transient protection according to the thermal capacity margin. When it is determined that the change rate or peak value of the three-phase current data exceeds the transient current transient change rate threshold or the peak current threshold, transient overload protection is executed.

[0016] The software slow track prediction control unit in the end-side fast and slow dual track protection unit generates steady-state thermal prediction results by performing full-system thermodynamic modeling based on the multi-source operating data, and forms a two-way closed-loop calibration mechanism with the hardware fast track protection control unit to correct the calculation parameters of the hardware fast track protection control unit in real time.

[0017] The side-side production line-level collaborative control unit performs dynamic load scheduling for the multiple motors based on the thermal capacity margin reported by each of the end-side fast and slow dual-track protection units.

[0018] The cloud-based global health management platform is used to perform insulation aging assessment and global parameter optimization based on the operating data of the multiple motors throughout their entire life cycle.

[0019] To achieve the above objectives, a third aspect of the present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory. When the computer program is executed by the processor, it implements the above-described motor thermal overload predictive control method based on multi-source data fusion.

[0020] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0021] In practical applications on continuous industrial production lines, this invention effectively balances the technical conflict between protection response speed and production continuity through a multi-level collaborative architecture. When the production line motor encounters a sudden high-load impact, the end-side pure hardware architecture can intercept the destructive impact at the lowest level with extremely low latency and dynamically adjust the judgment threshold based on real-time thermal capacity margin, effectively shielding against false triggering interference from fluctuations in normal operating conditions and significantly improving the safety of the underlying control. At the same time, the steady-state thermodynamic model at the software level provides high-precision closed-loop self-calibration for the hardware, effectively eliminating the risk of thermal model drift under long-cycle operation.

[0022] During the overall operation of the production line, the edge system utilizes the precise thermal margin data reported in real time from the end side to proactively and dynamically transfer the total load of the production line to other healthy motors in proportion to their thermal carrying capacity before any single motor approaches its overload limit. This greatly avoids the overall paralysis of the production line caused by power failure due to a single machine overload. Combined with the cloud-based macroscopic insulation aging assessment and global parameter distribution based on full lifecycle operation data, this technical solution achieves seamless integration from underlying transient hardware protection and steady-state precise thermal tracking to macroscopic lifecycle scheduling of the production line. While ensuring the physical safety of each piece of equipment, it maximizes the efficient and continuous operation of the entire industrial production line. Attached Figure Description

[0023] The disclosure of this invention is illustrated with reference to the accompanying drawings. It should be understood that the drawings are for illustrative purposes only and are not intended to limit the scope of protection of this invention. In the drawings, the same reference numerals are used to refer to the same parts. Wherein:

[0024] Figure 1 This is a schematic diagram illustrating the implementation of the motor thermal overload predictive control system based on multi-source data fusion provided by the present invention.

[0025] Figure 2This is a flowchart illustrating the motor thermal overload predictive control method based on multi-source data fusion provided by the present invention.

[0026] Figure 3 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0027] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0028] The following description, with reference to the accompanying drawings, outlines a method, system, and electronic device for predictive control of motor thermal overload based on multi-source data fusion, according to embodiments of the present invention.

[0029] Example 1:

[0030] like Figure 1 As shown in the figure, this embodiment discloses a motor thermal overload prediction and control system based on multi-source data fusion. This system is applied to a complex industrial continuous production line scenario containing multiple motors.

[0031] In actual industrial settings such as metallurgy, papermaking, or large material conveying, multiple motors typically work together to drive the same mechanical transmission network. The system provided in this embodiment is divided into three layers in terms of both physical and logical architecture: multiple end-side fast / slow dual-track protection units, a side-side production line-level collaborative control unit, and a cloud-side global health management platform. The multiple end-side fast / slow dual-track protection units are respectively configured for each of the multiple motors, serving as the lowest-level physical data sensing and transient execution mechanism; the side-side production line-level collaborative control unit is deployed in the field control cabinet or edge computing server of the production line, responsible for coordinating its subordinate multiple end-side units; the cloud-side global health management platform is deployed in a remote data center, responsible for long-term computation and baseline distribution of massive historical data.

[0032] Specifically, the end-side fast and slow dual-rail protection unit includes a SerDes high-speed transmission module, a hardware fast-rail protection control unit, and a software slow-rail prediction control unit in its hardware entity. The SerDes high-speed transmission module is directly connected to the front-end current transformer, voltage sensor, and vibration sensor located at the motor bearing, for collecting multi-source operating data of the corresponding motor. This multi-source operating data includes at least three-phase current data, three-phase voltage data, and vibration data. The significance of using the SerDes high-speed transmission module lies in its ability to complete the analog-to-digital conversion and serial transmission and reception of massive amounts of industrial analog signals with extremely low physical layer latency without consuming central processing unit resources.

[0033] Optionally, to balance the trade-offs between system communication bus bandwidth, processor computational load, and data fidelity, the SerDes high-speed transmission module incorporates an adaptive intelligent load balancing engine. The operation of this engine relies on the real-time status of the motor's thermal capacity margin. The thermal capacity margin is defined as follows. This is a dimensionless scalar representing the remaining usable thermal energy space of the motor. The thermal capacity margin... The defining formula is:

[0034] ;

[0035] In the formula, This indicates the maximum allowable temperature of the motor insulation material. This value is uniquely determined by the motor's insulation class. For example, the maximum allowable temperature for Class F insulation is set at 155 degrees Celsius. This indicates the current temperature of the motor windings; This indicates the ambient temperature of the physical space where the multiple motors are located.

[0036] Specifically, the adaptive intelligent flow splitting engine calculates the heat capacity margin in real time. A level 3 dynamic traffic splitting strategy is implemented, and the data flow reconstruction mechanism of this strategy is as follows:

[0037] Level 1: When the calculation yields... When the current sampling rate is low, it indicates that the motor's thermal capacity is in a relatively sufficient and safe state. At this time, the adaptive intelligent current shunting engine actively reduces the sampling rate of the three-phase current and three-phase voltage data. The purpose of reducing the sampling rate is to significantly reduce the redundant invalid data on the system bus when the motor is in a low-load safe state. At this time, the system only sends the downsampled low-frequency operating data to the software slow-track prediction control unit for steady-state trend recording, while the original high-frequency raw operating data is cached locally. The significance of local caching is that if a sudden anomaly occurs later, the system can restore the complete high-frequency waveform before the fault by retrieving the cached data.

[0038] Level 2: When the calculation yields... When this occurs, it indicates that the motor has entered the warning operating range of medium thermal load. At this time, the adaptive intelligent shunt engine stops downsampling, maintains the initial sampling rate of the analog-to-digital converter, and simultaneously sends all the high-frequency raw operating data to the hardware fast-track protection control unit and the software slow-track prediction control unit. This dual-track concurrent routing strategy ensures that the hardware circuitry can capture transient glitches, while the software algorithm can acquire high-precision data for thermodynamic deviation calculation.

[0039] Level 3: When the calculation yields... When the motor's thermal capacity is about to be exhausted, it is in an extremely dangerous critical overload state. At this time, the adaptive intelligent current shunting engine not only resumes full-speed sampling but also further increases the sampling rate of the three-phase current and three-phase voltage data, unconditionally sending all the high-frequency raw operating data to the hardware fast-track protection control unit and the software slow-track prediction control unit. Simultaneously, this state triggers a synchronous increase in the sampling rate of the vibration data. The reason for increasing the vibration sampling rate is that when the motor approaches its thermal limit, the softening of the insulation material or the thermal expansion of the rotor can easily induce high-frequency mechanical resonance, necessitating an increased sampling rate to capture high-frequency vibration characteristics.

[0040] It is also important to note that within the dual-rail protection unit on the end side, the hardware fast rail protection control unit is the core underlying physical protection mechanism for the entire system to cope with sudden high load impacts. To ensure deterministic response time, the hardware fast rail protection control unit is implemented using a field-programmable gate array (FPGA) – a pure hardware circuit – without any microprocessor core or instruction set execution mechanism. Internally, it is configured with a hardware current RMS value calculation subunit, a hardware first-order thermodynamic temperature estimation subunit, and a thermal capacity margin calculation subunit, all connected sequentially via a hardware description language.

[0041] Specifically, the hardware current RMS value calculation subunit operates independently according to a preset calculation cycle. Within each preset calculation cycle, the subunit performs pipelined squaring, integral accumulation, and square root operations on the input three-phase current data to calculate the instantaneous RMS value of the three-phase current data. Due to the pipelined hardware design, the output update delay of this RMS value is strictly controlled at the nanosecond level, which is far higher than the software calculation speed of traditional digital signal processors.

[0042] Subsequently, the hardware first-order thermodynamic temperature estimation subunit receives the instantaneous effective value. Based on the law of conservation of energy and Newton's law of cooling, the temperature change of the motor windings is determined by both Joule heating and convective heat dissipation. This subunit calculates the hardware-estimated winding temperature in real time based on a preset recursive formula combined with the instantaneous effective value. The preset recursive formula is as follows:

[0043] ;

[0044] In the formula, Indicates the current calculation time. The hardware-estimated winding temperature output by the subunit; Indicates the time at the previous calculation time. The hardware-estimated winding temperature already output by the sub-unit is stored in a hardware register for use in this iteration calculation; Indicates ambient temperature; This indicates the calculation cycle of the hardware fast track protection control unit, which is a fixed constant; This represents the current thermal time constant of the motor, which characterizes the physical inertia of the motor's heating rate. Indicates the current calculation time. The instantaneous effective value is input from the hardware current effective value calculation subunit; This indicates the current winding resistance of the motor; represents the base constant of the natural logarithm. In this formula, the sum of the first and second terms represents the residual heat generated by the motor in the previous state during natural cooling in the current cycle, and the third term represents the additional Joule heat generated by the current in the current cycle.

[0045] Next, the heat capacity margin calculation subunit estimates the winding temperature based on the hardware, inputs the aforementioned heat capacity margin formula, and outputs the current heat capacity margin in real time. .

[0046] For example, to ensure that the threshold determination at the hardware layer is no longer a static and inflexible constant value, the hardware fast track protection control unit is equipped with a dedicated dynamic threshold adjustment submodule. The significance of this module is that when the motor is cold, it can withstand extremely large transient inrush currents without damage; however, when the motor is hot, even small current fluctuations can be the final blow that breaks down the insulation. Therefore, the dynamic threshold adjustment submodule adjusts the threshold based on the thermal capacity margin. Threshold for transient change rate of current and peak current threshold Perform nonlinear dynamic adjustment.

[0047] Its specific nonlinear adjustment logic is as follows:

[0048] when At that time, the system directly assigns the highest tolerance and executes the equation. as well as ;

[0049] when Since the heat generated by the current exhibits a square relationship, a simple linear adjustment cannot accurately reflect the decrease in the thermal tolerance of the insulation. Therefore, a square root nonlinear attenuation model is introduced, and the equation is applied. ,as well as ;

[0050] when At this point, the risk of thermal collapse increases exponentially, at which point the threshold contraction strategy is transformed into a more stringent steep linear contraction, executing the equation... ,as well as .

[0051] The physical meaning of each constant parameter is defined as follows: This indicates the maximum rate of change of the input current over time that the motor is allowed to operate in a fully cooled state. This represents the minimum rate of change of current over time that the motor can be supplied under rated maximum temperature rise conditions. This indicates the maximum peak overload current that the motor can be input in a fully cooled state. This indicates the minimum peak overload current that the motor is allowed to input under its rated maximum temperature rise condition.

[0052] Specifically, based on dynamically set thresholds, the hardware fast track protection control unit also incorporates a multi-dimensional fault feature parallel detection module and an anti-false triggering logic module. In industrial settings, severe electromagnetic interference makes it highly susceptible to misjudgments from single-dimensional electrical feature exceedances. Therefore, the multi-dimensional fault feature parallel detection module simultaneously extracts five key features from the multi-source operating data: three-phase current transient change rate, three-phase current peak value, motor input power change rate, bearing vibration impact pulse value, and the aforementioned calculated heat capacity margin. .

[0053] The anti-false triggering logic module executes a hierarchical cross-validation judgment mechanism: when... Furthermore, if any single key feature exceeds its corresponding adjusted threshold, the protection condition is directly triggered, because the extremely fragile motor cannot tolerate any anomalies at this point; when The protection condition can only be triggered when at least two of the aforementioned key feature quantities simultaneously exceed their corresponding thresholds within the same hardware clock cycle; when The protection condition is triggered only when at least three of the aforementioned key characteristic quantities simultaneously exceed their corresponding thresholds within the same hardware clock cycle. Furthermore, to filter out occasional electromagnetic spikes, after the protection condition is triggered, the logic determination state must remain unchanged for a preset time window before the anti-false triggering logic module finally issues the action execution command to implement the transient overload protection. The transient overload protection is executed by directly cutting off the gate pulse signal of the motor driver, thus cutting off the energy source.

[0054] It should also be noted that although the hardware-based fast track protection control unit has an extremely fast response speed, its internal motor's current thermal time constant... With the current winding resistance of the motor After prolonged operation, deviations may occur due to the slow drift of material physical properties. To resolve this inherent contradiction, the system is designed with the aforementioned two-way closed-loop calibration mechanism.

[0055] Specifically, the workflow of the bidirectional closed-loop calibration mechanism relies on the software slow-track prediction and control unit. This unit runs on a high-performance processor and is not limited by hardware gate resources. It employs a complex multi-scale spatiotemporal fusion thermodynamic model that considers not only first-order electrothermal conversion but also the spatial structure of the motor, the three-dimensional heat dissipation equations of the cooling airflow, and the partial differential matrix of heat conduction from the rotor to the stator. This allows it to periodically output highly accurate software-estimated winding temperatures. .

[0056] Subsequently, the software slow-track prediction control unit calculates the software-estimated winding temperature. The aforementioned hardware estimation of winding temperature The deviation is calculated. Based on this deviation, the system determines that key parameters in the hardware model have experienced physical drift and corrects the current thermal time constant of the motor within the hardware fast track protection control unit in real time. and the current winding resistance of the motor .

[0057] Specifically, the parameter correction formula in the bidirectional closed-loop calibration mechanism is designed based on the proportional feedback principle, as follows:

[0058] ;

[0059] ;

[0060] In the formula: This indicates the current calibration time at which the software slow-track prediction control unit synchronously sends out execution parameters; This indicates the last calibration time when the unit last executed parameter synchronization; This represents the corrected current thermal time constant of the motor calculated at this current calibration time. This represents the corrected current winding resistance of the motor calculated at this current calibration time; This represents the current thermal time constant of the motor before correction, which is used from the previous calibration time to the present. This indicates the current winding resistance of the motor before correction, which is used from the previous calibration time. This indicates the software-estimated winding temperature output at the current calibration time; This indicates the hardware-estimated winding temperature captured at the current calibration moment; Indicates ambient temperature. The new correction value will be used. and The underlying registers of the field-programmable gate array are rewritten to complete a closed-loop calibration cycle, ensuring that the hardware algorithm always approximates the real physical state.

[0061] For example, when multiple motors are operating on their respective trajectories, traditional independent protection often causes a chain reaction of production line shutdowns. The side-side production line-level collaborative control unit in this system effectively resolves this problem by taking over global scheduling authority. When a motor's thermal capacity margin continuously decreases due to localized material accumulation or other reasons, but has not yet reached the mandatory shutdown threshold, the side-side production line-level collaborative control unit will intervene in advance.

[0062] The specific dynamic load scheduling process is as follows: First, the side-side production line-level collaborative control unit obtains the thermal capacity margin of each motor in the multiple motors in real time through the industrial fieldbus. and the factory-set rated power of a single motor Since the size and current load conditions of each motor differ, a unified quantitative indicator must be introduced to measure its physical capacity to accept new loads.

[0063] Therefore, this unit calculates the remaining heat load capacity of each motor. The mathematical definition of this indicator is: Based on the remaining heat-bearing capacity calculated for each individual motor, their respective weighting ratios are calculated. While maintaining the total mechanical load that the entire production line needs to overcome unchanged, the total load of the production line is... Perform a global reallocation so that the first The updated target power allocated to the TE Connectivity Follow the following equation principle: In the formula, the parameters This represents the sum of the remaining thermal carrying capacity of all controlled motors within the corresponding production line. The essence of this scheduling mechanism is to sacrifice the micro-torque margin of some lightly loaded motors in their mechanical branches in exchange for physical cooling space for overloaded motors on the verge of thermal exhaustion, thus ensuring the macro-level continuity of the entire industrial system's transport.

[0064] It is also important to note that after several months or years of system operation, the insulation material itself will undergo irreversible aging due to the thermal cracking effect of polymers. This will cause the original insulation tolerance limit of the motor to gradually decrease. The cloud-based global health management platform is responsible for handling this extremely long-term macroscopic physical degradation process and performing motor aging assessments for the multiple motors.

[0065] The specific process is as follows: The cloud-based global health management platform uses a time-series database to continuously collect massive amounts of historical operating data of the motor throughout its entire lifecycle, and extracts the corresponding extraction step size from it. Absolute temperature of motor windings inside This differs from the Celsius temperature scale used in calculating the heat margin mentioned above. To meet the dimensional requirements of the Arrhenius chemical reaction rate equation, the... The Kelvin (K) absolute thermodynamic temperature scale must be used to ensure the cumulative insulation aging rate. The physical reality of the calculation.

[0066] Subsequently, the cumulative insulation aging of the motor was calculated using the Arrhenius model from the field of physical chemistry. The mathematical formula for calculating this cumulative degree is:

[0067] ;

[0068] In the formula: This indicates the cumulative degree of insulation aging, which reflects the quantitative index of dielectric strength damage to insulating materials. This constant represents the activation energy of the chemical reaction of a specific motor insulation material; it is usually provided by the insulation varnish manufacturer. This represents the universal ideal gas constant. This summation operation is performed throughout every start-stop cycle of the motor.

[0069] When the cloud data server determines the cumulative degree of insulation aging When the preset benchmark threshold, which characterizes the later stage of insulation life, is reached, it indicates that the motor's shock resistance has substantially deteriorated. At this point, the cloud-based global health management platform proactively sends synchronization signals to each underlying network node and issues correction commands to the end-side fast and slow dual-track protection unit, proactively and permanently lowering the current transient rate threshold or the peak current threshold in the benchmark equation of the dynamic threshold adjustment submodule. This allows older motors at the end of their life cycle to receive more stringent and sensitive underlying protection for their remaining service life, effectively preventing catastrophic short-circuit burnout during service with defects.

[0070] Compared to existing motor protection technologies that rely on a single bimetallic thermal relay or a fixed-time overload software calculation model, this solution constructs a data processing and control chain covering the micro to macro levels. Existing single hardware solutions lack flexibility and adaptability, while simple microprocessor software models cannot overcome the latency limitations imposed by system bus communication.

[0071] This embodiment delegates the control of low-level protection to a field-programmable gate array (FPGA) that achieves ultra-fast response through hard-wired pure digital circuitry. Simultaneously, it offloads the computationally intensive steady-state thermal field calculations to an upper-level processor. While ensuring the safety of individual devices, edge scheduling effectively prevents unplanned downtime of the entire production line due to single-device failures. Finally, the cloud platform's global adjustments based on a long-term thermodynamic model overcome the technical challenges of long-cycle data evaluation. This three-layer collaborative architecture endows multiple motors in the entire industrial production line with both instantaneous physical immunity to sudden abnormal fluctuations and macroscopic self-healing capabilities for maintaining a healthy state throughout their entire lifecycle.

[0072] Example 2:

[0073] This embodiment further discloses an advanced protection and control mechanism for interference from complex industrial physical environments. In real-world heavy-duty continuous industrial production line applications such as metallurgical plants, cement manufacturing plants, large paper mills, or mine material conveyor belts, motors not only face high-frequency dynamic fluctuations in their internal electrical loads, but also encounter extremely harsh external physical environments. As the equipment's operating cycle extends, irregular dust accumulation, oil adhesion, or clogging of the independent cooling fan filter on the motor casing surface can easily occur. These environmental disturbances cause a dramatic nonlinear increase in the convective heat transfer resistance between the motor casing and the surrounding air, preventing the heat generated inside the motor from being effectively dissipated into the ambient air at the designed rate.

[0074] If the system relies solely on the slow-track model in Implementation Example 1 to detect temperature anomalies and mechanically triggers a production line-level load transfer action via the side-side production line-level collaborative control unit, it will unnecessarily reduce the mechanical capacity of the entire production line when the electrical health of the motor itself is actually good. To address this technical defect of misjudging "external physical heat dissipation obstruction" as "internal electrical load overload," this embodiment expands the information dimensions and scheduling logic of the entire control system.

[0075] It is important to note that, in order to achieve physical isolation of the heat source, the system must introduce physical parameters that can directly characterize the external heat transfer boundary conditions, in addition to the basic electrical data. Therefore, the multi-source operating data also includes the real-time temperature of the motor housing surface. At the engineering implementation level, obtaining this temperature data is not simply a matter of arbitrarily attaching sensors to the outside of the housing; rather, it requires strict site selection and installation process constraints. Typically, a high-precision platinum resistance temperature sensor is used, and it is permanently bonded to the base of the heat dissipation fins on the outer surface of the motor stator housing using epoxy resin with high thermal conductivity. This avoids direct forced airflow from external cooling fans, ensuring that the sensor readings accurately reflect the steady-state heat level of the stator core reaching the housing surface through conduction. This temperature signal is also connected to the SerDes high-speed transmission module on the end side. After analog-to-digital conversion, it is strictly clock-synchronized with the aforementioned three-phase current and three-phase voltage data for sampling and packet transmission.

[0076] Optionally, to perform in-depth processing of the aforementioned multi-source operating data and implement intervention strategies, the side-side production line-level collaborative control unit integrates an external heat dissipation degradation decoupling and active intervention module. This module is used to isolate the causes of heat sources before executing the dynamic load scheduling. Its specific control logic is as follows: The core physical idea behind this heat source isolation is to decompose the driving source of the overall motor temperature rise into two independent parts: one is the increase in internal Joule heating caused by the increase in internal current due to the increase in torque output; the other is the heat accumulation caused by the increase in thermal resistance due to external physical coverings. The existence of this module allows the control system to transition from a single passive defensive electrical scheduling to proactive environmental intervention-based comprehensive management.

[0077] For example, the specific operation of this module is based on discrete time series decomposition. It operates within a preset side-side scheduling time step. Next, extract the current step size. The real-time temperature of the motor housing surface And calculate the motor at the current step size. Average internal comprehensive heating power During this process, the side-side scheduling time step The step size setting must fully consider the enormous thermal inertia of the motor housing. Since the heating process of the metal housing is a slow integral process, if the step size is set too short, the temperature change will be drowned out by the sensor's measurement noise; if the step size is set too long, sudden deterioration in heat dissipation cannot be detected in time. Typically, the step size is set in the range of several minutes to tens of minutes.

[0078] Specifically, regarding the average internal comprehensive heating power... The calculation, in its physical essence, involves integrating and averaging all energy losses converted into heat within the motor within the given step window. To ensure accuracy and adapt to the data stream format of field-programmable gate arrays or digital signal processors, this embodiment uses the following formula to calculate this parameter:

[0079] ;

[0080] In the formula, This indicates that the motor is at the current step size. The average internal comprehensive heating power; Indicates the current step size Within the time window, the equivalent stator current mean value is extracted from the three-phase current data after processing by root mean square integration. This represents the equivalent internal resistance constant of the stator winding after taking temperature compensation into account. This represents the fixed physical losses of the motor under constant excitation level, including the hysteresis loss and eddy current loss of the stator core, and the sum of mechanical friction loss and wind resistance loss at the rotor bearings. The first term of this equation accurately quantifies the copper loss heat generation that varies drastically with load fluctuations, while the second term compensates for the basic heat source under constant operating conditions. The sum of the two objectively reflects the true heat generation rate inside the motor.

[0081] It should also be noted that after acquiring the heat generation data and boundary temperature data, the external heat dissipation degradation decoupling and active intervention module is based on the average internal comprehensive heat generation power. Construct an internal and external thermal resistance decoupling model and calculate the current step size. External heat dissipation degradation characteristic factor The calculation formula is:

[0082] ;

[0083] This formula is based on the theory of steady-state heat conduction networks. Among them, Ambient temperature; This is the standard convective heat transfer thermal resistance constant for a pre-calibrated motor in a clean, unobstructed state. Under thermodynamic equilibrium, the temperature difference between the motor casing surface and the ambient air should be equal to the product of the internal heat generation power and the external convective heat transfer thermal resistance. Therefore, the numerator in the formula represents the actual temperature rise between the current motor casing surface and the ambient air; the product term in the denominator represents the theoretical temperature rise that should occur at the current heat generation power level if the motor surface were in an ideally clean state. Thus, it can be seen that... The physical meaning of this dimensionless factor is the ratio of the actual external thermal resistance of the current motor to the standard ideal thermal resistance.

[0084] Optionally, regarding parameters Obtaining this constant typically requires calibration during the initial deployment phase of the system, when the motor body is clean and free of contaminants, and the cooling air ducts are unobstructed. The motor is then continuously run under rated load until it reaches thermal stability. The steady-state casing temperature, ambient temperature, and heat output are recorded at this point. This constant can then be solidified using a reverse division method and stored in the non-volatile memory of the side-side production line-level collaborative control unit.

[0085] Specifically, after extracting the feature factors, the system enters the crucial decision execution branch. The external heat dissipation degradation decoupling and active intervention module decouples the external heat dissipation degradation feature factors. With the preset external physical obstruction alarm threshold A comparison is performed. This threshold is typically set in the range of 1.2 to 1.5, indicating that when the actual thermal resistance increases to a specific multiple of the standard condition, an environmental anomaly is considered to have occurred.

[0086] In the first case of this comparison logic: if If the rate of change of the heat capacity margin is less than 0, then the dynamic load scheduling for the corresponding motor is suspended to maintain the load distribution ratio of the motor, and an active environmental intervention command is output to the external physical cooling equipment. In this logic, the condition "the rate of change of the heat capacity margin is less than 0" is used to confirm that the motor is indeed in a deteriorating process of evolving towards thermal overload, in order to prevent the ambient temperature from rising. The sudden drop caused a erroneous triggering of mathematical calculations; when calculating the rate of change, the system... Smoothing filtering was performed to eliminate normal static temperature fluctuations and external cold shock interference.

[0087] When both conditions are met, the system's diagnostic conclusion is very clear: although the motor's heating leads to a decrease in thermal margin, the underlying reason is that external deposits (such as thick dust) increase the surface convection heat transfer resistance, hindering heat dissipation.

[0088] At this point, the system decisively suspended the load transfer command originally planned in Implementation Example 1. The reason for not reducing the load is that since electrical load is not inherently problematic, blindly reducing the load not only fails to immediately remove surface dirt but also directly causes a decrease in output at the production line station where the motor is located. By maintaining the load distribution ratio, the continuity of the production process is ensured. Simultaneously, the system outputs an active environmental intervention command to external physical cooling equipment. In actual industrial settings, this typically manifests as: triggering a high-frequency pulse pneumatic purging solenoid valve installed above the motor casing to forcibly blow away dust and carbon deposits from the casing surface using high-pressure compressed air; or sending an acceleration command to an external independent forced air cooling fan inverter to forcibly reduce external convective heat transfer resistance by increasing the airflow velocity. Through this active intervention in the external environment, the motor's heat dissipation capacity is restored from the root without sacrificing any mechanical output.

[0089] For example, in the second case of this comparison logic: if If the rate of change of the heat capacity margin is less than 0, then the dynamic scheduling action of redistributing the total load of the production line according to the proportion of the remaining heat carrying capacity will resume. When the system falls into this branch judgment, it indicates that the actual external thermal resistance of the motor has not significantly exceeded the standard, and the heat exchange channels on the surface of the casing are in a healthy and unobstructed state. However, the overall heat capacity margin of the motor is still decreasing, which leads to the main physical cause: the motor is experiencing a real and severe electrical or mechanical overload, and the rate at which heat is generated inside it has far exceeded the limit that a healthy cooling system can dissipate.

[0090] Faced with this internally caused overload, external environmental cleaning or airflow is insufficient to effectively reduce the system temperature. Therefore, the side-line level collaborative control unit immediately releases the suspended state and resumes dynamic scheduling. Based on the remaining heat capacity reported by each motor in real time, the system quickly reduces the load allocation weight of the faulty motor, smoothly transferring its portion of the load to other healthy motors with sufficient thermal margin in the production line. In this way, the operating current of the faulty motor is forcibly reduced to within a safe range, preventing its insulation system from being broken down by high temperature.

[0091] Traditional overload protection mechanisms often attribute the rise in stator winding temperature solely to excessive equipment load. When faced with environmental degradation such as dust accumulation, their only response is to forcibly reduce the motor speed or torque, which can lead to significant economic losses in heavy industry.

[0092] This system innovatively incorporates the casing temperature dimension, combining it with real-time estimation of internal heat generation to construct a robust mathematical model for thermal resistance decoupling. By distinguishing between two distinct physical causes—physical heat dissipation obstruction and electrical overload—the system can precisely issue control commands. This enables condition-based maintenance and self-cleaning in the early stages of environmental degradation, while also ensuring smooth load transfer under actual overload conditions. This mechanism effectively improves the capacity utilization of industrial equipment in harsh environments, achieving a more reasonable balance between safety and economic efficiency in the overall system operation.

[0093] Example 3:

[0094] This embodiment further discloses a comprehensive solution for addressing heat generation caused by ineffective electrical losses within a motor. In practical industrial operations such as modern precision machining, high-speed rotating machinery, or heavy-duty traction, motors are typically driven by frequency converters (VFDs) or servo drives. Because the pulse-width modulation (PWM) voltage output by the frequency converter contains a large number of high-order harmonics, these harmonic currents do not generate effective torque in the motor stator windings. However, due to the skin effect and proximity effect, they significantly increase the high-frequency equivalent resistance of the windings, thereby triggering severe harmonic Joule heating.

[0095] It is important to note that existing thermal protection mechanisms often cannot distinguish whether the temperature rise is caused by excessive mechanical load (increased fundamental current) or by deterioration of the driver output waveform quality (excessive harmonic content). Indiscriminately performing load transfer or shutdown actions will result in unnecessary losses of production efficiency.

[0096] Therefore, this embodiment, based on the high-frequency raw operating data collected by the Serdes high-speed transmission module, achieves an electrical heating suppression mechanism without reducing mechanical output by using the harmonic thermal feature separation and carrier frequency dynamic optimization module built into the software slow track prediction control unit.

[0097] For example, the execution logic of this module does not continuously occupy computing resources, but instead adopts an on-demand computing mode triggered by the deviation rate. Its control steps are as follows: First, the system continuously monitors the thermal capacity margin calculated in real time by the hardware fast track protection control unit. When the software slow-track prediction control unit determines the thermal capacity margin... The rate of change is less than At that time, it is determined This means that the motor is currently in a worsening trend of continuous heat accumulation and shrinking available thermal space.

[0098] Specifically, once the above-mentioned rate of change determination condition is triggered, the system immediately initiates a discrete analysis step-size-based process. The harmonic decoupling process is analyzed here. The step size is also analyzed. The software slow-track prediction control unit extracts the high-frequency raw operating data, particularly the three-phase current data, within the step-size window from the buffer of the Serdes high-speed transmission module. To ensure the accuracy of the time-domain signal to frequency-domain feature conversion, the analysis step size... The sampling length is usually set to an integer multiple of the motor's fundamental frequency period to avoid spectrum leakage.

[0099] It is also important to note that after acquiring high-fidelity current waveform data, the system performs a Fast Fourier Transform (FFT) to decompose the complex time-domain current signal into amplitude and phase characteristics of different frequency components. By performing energy integration and feature filtering on the spectrum, the current step size is separated. RMS value of fundamental current and the effective value of total harmonic current The effective value of the total harmonic current. It integrates the current energy at all discrete frequency points except the fundamental frequency, and it represents the electrical energy component in the driver output current that cannot be converted into mechanical energy.

[0100] For example, to quantify the contribution of these harmonic components to the temperature rise of the motor windings, this embodiment introduces a harmonic thermal contribution ratio evaluation model. This module calculates the current step size. The proportion of harmonic heat contribution The calculation formula is:

[0101] ;

[0102] In the formula: Indicates the analysis step size The proportion of heat loss generated by the lower harmonics to the total Joule loss of the stator winding; Indicates the current step size The effective value of total harmonic current extracted internally; Indicates the current step size The effective value of the fundamental current extracted internally; This represents the equivalent amplification factor of the skin effect of the winding. It is a physical constant used to characterize the increase in equivalent resistance caused by the uneven distribution of high-frequency harmonic current across the conductor cross-section. Its value is related to the current carrier frequency of the inverter. This is also related to the material and geometric parameters of the winding conductors. The denominator of the formula reflects the total copper loss energy level of the stator winding, while the numerator accurately extracts the reactive power loss contributed by harmonic currents.

[0103] Specifically, the system will calculate the proportion of harmonic thermal contribution. With the preset harmonic heating alarm threshold Perform a comparison. This threshold... The setting is based on the rated harmonic tolerance limit of the motor design. In practical applications, if... Setting it too low will cause frequent carrier frequency adjustments, increasing the switching losses of the driver; setting it too high may cause the motor windings to be in a state of localized high heat for a long time. This threshold is usually set in the range of 0.15 to 0.25.

[0104] If the judgment result is This indicates the current thermal capacity margin of the motor. The temperature drop is primarily caused not by excessive mechanical load, but by ineffective electrical harmonic heating. In this case, the system deems cooling through "reducing mechanical load" or "shifting load during production line scheduling" uneconomical and unnecessary. Therefore, the harmonic thermal characteristic separation and carrier frequency dynamic optimization module performs an intervention interception operation: it sends a "suspend" command to the side-side production line-level collaborative control unit, preventing it from performing dynamic load scheduling based on mechanical derating. Simultaneously, the system initiates an active thermal suppression process based on energy spectrum optimization.

[0105] It should also be noted that the carrier frequency dynamic optimization module will adjust the frequency based on the current step size. Based on the spectral characteristics and the motor's loss mapping matrix, a pre-set energy spectrum optimization algorithm is used to output an optimized carrier frequency. The physical logic of this energy spectrum optimization algorithm lies in finding a specific switching frequency that minimizes the sum of the inverter's switching losses and the motor's harmonic losses at the current operating point.

[0106] Specifically, in order to achieve this optimization process, this embodiment adopts the following loss evaluation objective function:

[0107] ;

[0108] In the formula: The optimal carrier frequency to be solved; This indicates that the frequency converter operates at a carrier frequency of The power switching device losses at that time, which typically increase approximately linearly with increasing carrier frequency; This indicates that at a carrier frequency of The first time generated The amplitude of the second harmonic current; Indicates correspondence The equivalent AC resistance of the winding at subharmonic frequencies is different from the low-frequency resistance calibrated by the slow-track closed-loop system. The AC resistor The skin effect and proximity effect induced by the high-frequency alternating magnetic field are taken into account, and its resistance varies with the harmonic order. The increase is nonlinear and asymmetric, resulting in a significant increase. This indicates the highest harmonic order set by the software slow-track prediction control unit (e.g., a value of 50). The optimization algorithm traverses the driver's allowed carrier frequency band (e.g., 2kHz to 16kHz) to select the frequency that can significantly reduce the effective harmonic current of the motor without causing the driver to overheat. value.

[0109] For example, this optimized carrier frequency Once determined, the software-based slow-track prediction control unit sends a command to the corresponding motor's drive front end via internal bus or industrial Ethernet. Upon receiving this command, the front-end frequency converter performs a seamless frequency switching. It is important to emphasize that the frequency converter dynamically switches the pulse-width modulation switching frequency to the optimized carrier frequency. During the process, the fundamental frequency and amplitude of its output voltage remain constant, thereby indicating that the motor operates under the condition of maintaining the target output mechanical torque.

[0110] This control behavior has significant implications in industrial practice: by altering the spectral distribution of the PWM signal, it reduces the energy of specific-order harmonic currents, thereby suppressing the additional copper and iron losses derived from high-frequency harmonics. From a physical perspective, this is equivalent to achieving a thermal capacity margin for the motor by optimizing the quality of the electrical power supply without changing the motor's output. Its self-recovery mechanism provides highly reliable technical support for ensuring the continuous and stable operation of precision production lines under heavy load and high-speed conditions.

[0111] This embodiment constructs a multi-dimensional, collaborative motor thermal safety defense system. Compared with existing technologies that rely on a single temperature control switch or simple current limiting, this invention has significant advantages:

[0112] First, in terms of time dimension, the system achieves cross-scale coverage from nanosecond-level hardware triggering (fast track unit) to second-level software modeling (slow track unit), and then to full lifecycle aging management (cloud platform). This not only ensures physical safety in response to extreme situations such as lightning strikes and sudden stalls, but also solves the accuracy problem of model drift over time.

[0113] Secondly, in terms of physical cause identification, this solution overcomes the limitation of traditional technologies that treat temperature rise as a single load problem. By introducing environmental thermal resistance decoupling in Example 2 and harmonic loss decoupling in Example 3, the system possesses deep fault tracing capabilities. Active purging is triggered when environmental heat dissipation is obstructed, and carrier optimization is triggered when harmonics cause heating. This targeted intervention strategy allows the system to maximize the potential of the motor through the deep integration of electrical and mechanical means, without reducing production line capacity or incurring detrimental shutdowns.

[0114] Finally, in terms of cluster collaboration, the edge-side collaborative control unit has changed the previous model of independent protection for individual motors. Through the global proportional allocation of thermal load capacity, the system achieves production line-level load balancing. When a single motor faces a real mechanical overload that cannot be resolved through carrier optimization or environmental intervention, the system ensures that the entire production line's process flow remains uninterrupted through a smooth load transfer mechanism. This three-level closed-loop linkage scheme greatly improves the operational reliability and overall economic benefits of large-scale automated production lines, providing a clear and sufficient technical implementation path for the operation and maintenance of motor systems in intelligent manufacturing.

[0115] Example 4:

[0116] like Figure 2 As shown in the figure, this embodiment four provides a motor thermal overload predictive control method based on multi-source data fusion.

[0117] In practical continuous industrial production lines, such as steel rolling mills, large paper machines, or long-distance mineral conveyor belts with multiple drive systems, existing motor protection schemes mainly rely on single-node software algorithms or traditional thermal relays for temperature simulation and fixed threshold determination. These existing technologies face significant conflicting technical problems in specific application scenarios:

[0118] On the one hand, if a strict fixed trigger threshold is set to prevent the motor from burning out instantly, the processing delay of conventional software algorithms is very likely to cause malfunctions under transient impacts of industrial loads, directly cutting off the power supply to a single motor, resulting in unplanned shutdowns and capacity losses for the entire continuous production line.

[0119] On the other hand, if the protection threshold is relaxed in order to ensure continuous production of the production line, a single static thermodynamic model cannot accurately track the actual deep heat accumulation state of the motor. Long-term operation with defects can easily lead to accelerated aging and failure of insulation materials or even motor burnout.

[0120] To objectively and effectively resolve the technical conflict between protection response speed and production continuity, this embodiment discloses a motor thermal overload predictive control method based on multi-source data fusion. This method is applied to a control system comprising multiple end-side fast / slow dual-track protection units, a side-side production line-level collaborative control unit, and a cloud-side global health management platform. The multiple end-side fast / slow dual-track protection units are respectively configured for multiple motors. The method includes the following steps:

[0121] First, the Serdes high-speed transmission module in the end-side fast and slow dual-rail protection unit collects multi-source operating data of the corresponding motor. The multi-source operating data includes at least three-phase current data, three-phase voltage data and vibration data.

[0122] In actual hardware deployment, this step relies on a series of high-precision physical sensors installed on the underlying electrical circuit. Specifically, one set of high-frequency current transformers and one set of voltage transmitters are directly connected to the power supply cable from the frequency converter to the motor stator. Simultaneously, a piezoelectric high-frequency accelerometer is rigidly bolted to the bearing housing on the non-drive end of the motor. The continuous analog signals captured by these physical sensors are processed by a high-bit analog-to-digital converter (ADC) and then input to the SerDes high-speed transmission module. The SerDes high-speed transmission module converts the parallel multi-channel digital signals into a high-speed serial bit stream. In industrial environments with strong electromagnetic interference, the SerDes protocol enables the transmission and reception of massive amounts of data with extremely low physical layer latency. This not only significantly reduces the number of cables inside the end-side equipment but also ensures the timeliness and fidelity of the original operating data when transmitted to the processing core.

[0123] Secondly, the hardware fast rail protection control unit in the end-side fast and slow dual-rail protection unit calculates the motor winding temperature and thermal capacity margin in real time based on the multi-source operating data, and dynamically adjusts the transient current transient change rate threshold and peak current threshold of the transient protection according to the thermal capacity margin. When it is determined that the change rate or peak value of the three-phase current data exceeds the transient current transient change rate threshold or the peak current threshold, transient overload protection is executed.

[0124] In its physical implementation, this hardware fast-track protection control unit is composed of a Field-Programmable Gate Array (FPGA) chip without a microprocessor core. The FPGA utilizes its internal hardware multipliers and logic gate arrays to perform clock-level parallel pipelined operations on the data stream from the SerDes high-speed transmission module, directly deriving the winding temperature and thermal capacity margin. Based on the calculated margin indicators, the hardware logic dynamically adjusts two dynamic thresholds used for comparison (current transient rate threshold and peak current threshold), avoiding the inconsistency of static thresholds under cold and hot motor conditions. When an overload is detected, the FPGA physical circuitry bypasses all software communication protocol stacks and directly outputs a hardware-level interrupt signal. This signal directly cuts off the drive circuit pulses of the insulated-gate bipolar transistors (IGBTs) inside the front-end inverter, cutting off the three-phase energy input to the motor within a few microseconds, thereby protecting the motor windings from irreversible overcurrent thermal damage under harsh operating conditions.

[0125] Subsequently, the software slow track prediction control unit in the end-side fast and slow dual-track protection unit generates steady-state thermal prediction results by performing full-system thermodynamic modeling based on the multi-source operating data, and forms a two-way closed-loop calibration mechanism with the hardware fast track protection control unit to correct the calculation parameters of the hardware fast track protection control unit in real time.

[0126] In conjunction with the actual equipment control architecture, the software slow-track predictive control unit is typically equipped with a high-performance 32-bit or 64-bit digital signal processor (DSP) chip. Because the underlying silicon-based hardware circuits, such as FPGAs, experience temperature drift or aging shifts in their internally mapped stator resistance and thermal time constant due to physical laws during continuous operation over months, the DSP utilizes its floating-point arithmetic capabilities to run complex thermodynamic partial differential equations involving three-dimensional spatial convection coefficients and historical load integrals. The calculated high-precision steady-state thermal prediction results are used as a benchmark and compared with the FPGA's rapid calculation results. If the deviation exceeds a set range, the DSP directly rewrites the register parameter values ​​of the FPGA's underlying calculation module via its internal Serial Peripheral Interface (SPI) bus. This step eliminates the risk of thermal model drift during long-cycle operation without interfering with the nanosecond-level execution of the hardware fast track, ensuring the accuracy of the underlying assessment.

[0127] Furthermore, the side-side production line-level collaborative control unit performs dynamic load scheduling for the multiple motors based on the thermal capacity margin reported by each of the end-side fast and slow dual-track protection units.

[0128] In industrial automation networks, edge-side production line-level collaborative control units typically consist of industrial programmable logic controllers (PLCs) or edge computing control hosts deployed in field rooms. These units aggregate data from all edge nodes on the production line via industrial real-time Ethernet such as PROFINET or EtherCAT. For example, in a material handling assembly containing five drive motors, when the second motor experiences a continuous decrease in its thermal capacity margin and approaches the critical protection threshold due to localized mechanical jamming or ventilation blockage, the edge computing host, upon receiving the reported data, will send feedforward commands to the inverters of the other four healthy motors on the production line to increase frequency or torque, while simultaneously appropriately reducing the load reference value of the second motor. Through the physical coupling effect of the production line's mechanical transmission mechanisms (such as coupling networks or continuous conveyor belts), the mechanical work demand that the second motor cannot handle is smoothly transferred according to the proportion of the remaining thermal capacity of each motor. This scheduling intervention at the equipment level prevents a single motor from suddenly tripping due to reaching the underlying overload threshold, ensuring the continuous operation of the industrial process.

[0129] Finally, the cloud-based global health management platform is used to perform insulation aging assessment and global parameter optimization based on the operating data of the multiple motors throughout their entire life cycle.

[0130] In practical applications, the cloud-based global health management platform is deployed in the enterprise's remote data center. The edge gateway on-site uses 5G networks or industrial fiber optics to upload historical data such as the motor's daily average operating temperature and overload frequency to the cloud. The cloud server runs an aging assessment algorithm (such as the Arrhenius life equation) that includes chemical reaction activation energy parameters to calculate the cumulative fatigue damage of the stator insulation system for each motor. When it is determined that the insulation material of a motor is at the end of its life cycle degradation period, the cloud platform will proactively generate a global parameter optimization command and send it to the field. This command will correspondingly lower the upper and lower limits of the transient protection threshold in the hardware fast rail protection control unit for that aging motor, enabling the system to control this equipment with reduced shock resistance with a more stringent safety margin.

[0131] Combining existing technologies with the challenges faced in specific application scenarios, the method provided in this embodiment achieves significant and objectively beneficial effects. In practical use, this solution improves the accuracy and reliability of the underlying control system in the face of transient changes through closed-loop verification of a nanosecond-level pure hardware architecture and a steady-state software model. More importantly, before a single device approaches its thermal limit and triggers a power outage, this method actively and dynamically transfers the total load of the production line proportionally through the side-side system, mitigating the electrical overload crisis at the source of mechanical power distribution. Combined with the cloud-side's adaptive adjustment function for insulation aging status, this technical solution effectively avoids the overall production line paralysis caused by blind tripping of a single device, while ensuring accurate physical protection for every specific electrical device on the production line, thus maintaining the overall operating efficiency of the industrial continuous production line.

[0132] Example 5:

[0133] Corresponding to the above embodiments, the present invention also proposes an electronic device.

[0134] like Figure 3 The diagram shows a structural schematic of an electronic device according to the present invention. The electronic device 100 includes a processor 101 and a memory 103. The processor 101 and the memory 103 are connected, for example, via a bus 102. Optionally, the electronic device 100 may further include a transceiver 104. It should be noted that in practical applications, the transceiver 104 is not limited to one unit, and the structure of this electronic device 100 does not constitute a limitation on the embodiments of the present invention.

[0135] Processor 101 may be a CPU, a general-purpose processor, a DSP, an ASIC, an FPGA, or other programmable logic device, transistor logic device, hardware component, or any combination thereof. It may implement or execute the various exemplary logic blocks, modules, and circuits described in connection with this disclosure. Processor 101 may also be a combination that implements computational functions, such as including one or more microprocessor combinations, a combination of a DSP and a microprocessor, etc.

[0136] Bus 102 may include a pathway for transmitting information between the aforementioned components. Bus 102 may be a PCI bus or an EISA bus, etc. Bus 102 may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 3 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0137] The memory 103 stores a computer program corresponding to the motor thermal overload predictive control method based on multi-source data fusion in the above embodiments of the present invention. This computer program is executed by the processor 101. The processor 101 executes the computer program stored in the memory 103 to implement the content shown in the aforementioned method embodiments.

[0138] Among them, electronic devices 100 include, but are not limited to: mobile terminals such as laptops and PADs (tablet computers) and fixed terminals such as desktop computers. Figure 3 The electronic device 100 shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments of the present invention.

[0139] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A motor thermal overload predictive control system based on multi-source data fusion, applied to multiple motors, characterized in that, include: Multiple end-side fast and slow dual-track protection units, side-side production line-level collaborative control unit, and cloud-side global health management platform; The multiple end-side fast and slow dual-rail protection units are respectively configured on the multiple motors; The end-side fast and slow dual-rail protection unit includes a Serdes high-speed transmission module, a hardware fast rail protection control unit, and a software slow rail prediction control unit; the Serdes high-speed transmission module collects multi-source operating data of the corresponding motor, and the multi-source operating data includes at least three-phase current data, three-phase voltage data, and vibration data. The hardware fast track protection control unit calculates the motor winding temperature and thermal capacity margin in real time based on the multi-source operating data, and dynamically adjusts the transient current transient change rate threshold and peak current threshold of the transient protection according to the thermal capacity margin. When it is determined that the change rate or peak value of the three-phase current data exceeds the transient current transient change rate threshold or the peak current threshold, transient overload protection is executed. The software slow track prediction control unit generates steady-state thermal prediction results by performing full-system thermodynamic modeling based on the multi-source operating data, and forms a two-way closed-loop calibration mechanism with the hardware fast track protection control unit to correct the calculation parameters of the hardware fast track protection control unit in real time. The side-side production line-level collaborative control unit performs dynamic load scheduling for the multiple motors based on the thermal capacity margin reported by each of the end-side fast and slow dual-track protection units. The cloud-based global health management platform performs insulation aging assessment and global parameter optimization based on the operating data of the multiple motors throughout their entire life cycle.

2. The system according to claim 1, characterized in that, The Serdes high-speed transmission module has a built-in adaptive intelligent load balancing engine, which adjusts the load based on the current thermal capacity margin. Implement a level 3 dynamic traffic splitting strategy: when At the same time, the sampling rate of the three-phase current data and the three-phase voltage data is reduced, and the downsampled low-frequency operating data is sent to the software slow track prediction control unit, while the high-frequency original operating data is cached locally; when At the same time, the initial sampling rate is maintained, and all the high-frequency raw operating data are simultaneously sent to the hardware fast track protection control unit and the software slow track prediction control unit; when At the same time, the sampling rate of the three-phase current data and the three-phase voltage data is increased, and all the high-frequency raw operating data are sent to the hardware fast track protection control unit and the software slow track prediction control unit, while triggering the sampling rate of the vibration data to be increased synchronously. Wherein, the heat capacity margin The defining formula is: ； In the formula This refers to the maximum permissible temperature of the motor insulation material. This is the current temperature of the motor windings. The ambient temperature.

3. The system according to claim 1, characterized in that, The hardware fast track protection control unit is implemented by a field programmable gate array (FPGA) pure hardware circuit, which is connected in sequence to a hardware current effective value calculation subunit, a hardware first-order thermodynamic temperature estimation subunit, and a heat capacity margin calculation subunit. The hardware current RMS value calculation subunit calculates the instantaneous RMS value of the three-phase current data according to a preset calculation period; The hardware first-order thermodynamic temperature estimation subunit calculates the hardware estimated winding temperature in real time based on a preset recursive formula and the instantaneous effective value. The heat capacity margin calculation subunit outputs the current heat capacity margin in real time based on the hardware-estimated winding temperature. .

4. The system according to claim 3, characterized in that, The preset recursive formula is specifically as follows: ； in, For the current calculation time The hardware estimates the winding temperature. For the previous calculation time Hardware estimation of winding temperature, For ambient temperature, For the calculation period, This is the current thermal time constant of the motor. For the current calculation time The instantaneous effective value, The current winding resistance of the motor. It is a natural constant.

5. The system according to claim 4, characterized in that, The hardware fast track protection control unit is equipped with a dynamic threshold adjustment submodule, which adjusts the threshold based on the thermal capacity margin. Threshold for transient change rate of current and peak current threshold Non-linear dynamic adjustment is performed, and the adjustment logic is as follows: when hour, , ; when hour, , ; when hour, , ; in, This represents the maximum permissible rate of change of current in the cold state. The minimum permissible rate of change of current in hot state. This represents the maximum permissible peak current in the cold state. This is the minimum permissible peak current in hot condition.

6. The system according to claim 5, characterized in that, The hardware fast track protection control unit also has a built-in multi-dimensional fault feature parallel detection module and an anti-false triggering logic module. The multi-dimensional fault feature parallel detection module synchronously extracts five key features from the multi-source operating data: three-phase current transient change rate, three-phase current peak value, motor input power change rate, bearing vibration and shock pulse value, and heat capacity margin. ; The anti-false triggering logic module performs a tiered judgment: when... Furthermore, the protection condition is triggered when any single key feature quantity exceeds the limit; when And the protection condition is triggered when at least two of the aforementioned key feature quantities exceed the limit simultaneously; when Furthermore, the protection condition is triggered when at least three of the aforementioned key feature quantities exceed their limits simultaneously; The transient overload protection can only be executed after the protection condition is triggered and a preset time window is maintained.

7. The system according to claim 4, characterized in that, The specific workflow of the bidirectional closed-loop calibration mechanism includes: The software-based slow-track prediction and control unit employs a multi-scale spatiotemporal fusion thermodynamic model to periodically output software estimates of winding temperature. ; The software slow-track prediction control unit calculates the software-estimated winding temperature. The hardware estimates the winding temperature. The deviation is used to correct the current thermal time constant of the motor inside the hardware fast track protection control unit in real time. and the current winding resistance of the motor .

8. The system according to claim 7, characterized in that, The parameter correction formula in the bidirectional closed-loop calibration mechanism is as follows: ； ； in, For the current calibration time, For the previous calibration time, and These are the corrected current thermal time constant of the motor and the current winding resistance of the motor, respectively. and These are the motor's current thermal time constant and current winding resistance before correction, respectively. The ambient temperature.

9. The system according to claim 1, characterized in that, The load dynamic scheduling process within the side-side production line-level collaborative control unit is as follows: Real-time acquisition of the thermal capacity margin of individual motors among the multiple motors and rated power ; Calculate the remaining heat load capacity of each motor. ; The weighting ratio is calculated based on the remaining heat-bearing capacity, and the total load of the production line is... Redistribute so that the first The target power allocated to the trolley motor ,in This is the sum of the remaining thermal load capacity of all the aforementioned motors within the corresponding production line.

10. The system according to claim 1, characterized in that, The specific process for motor aging assessment on the cloud-based global health management platform is as follows: Collect historical operating data of the motor throughout its entire life cycle and extract the corresponding extraction step size. Absolute temperature of motor windings inside ; Calculate the cumulative insulation aging of the motor using the Arrhenius model. : ； In the formula The activation energy of insulating materials, It is the gas constant; When the cumulative degree of insulation aging When the preset benchmark threshold is reached, an instruction is sent to the side fast and slow dual-rail protection unit to actively lower the current transient change rate threshold or the peak current threshold.

Images