A new energy vehicle power take-off misoperation suppression system and method

Abstract

This invention belongs to the technical field of new energy vehicle control and transmission systems, specifically relating to a system and method for suppressing power take-off (PTO) malfunctions in new energy vehicles. The system includes a motor controller, a vehicle controller, a PTO controller, and a multi-source information sensing array. The vehicle controller utilizes a physical information neural network model to establish a cross-domain dynamic mapping between current ripple characteristics and mechanical vibration spectrum, and implements a three-level progressive strategy including predictive soft suppression, cooperative locking, and adaptive evolution. In the predictive soft suppression stage, energy fluctuations are actively offset through dynamic compensation of motor torque to achieve shock-free suppression; in the cooperative locking stage, physical-level power path interruption is executed. This invention achieves proactive elimination of malfunctions beforehand, extending the lifespan of transmission components and ensuring the safety of the high-voltage system while improving the system's operational robustness and self-evolutionary predictive capabilities.

Description

Technical Field

[0001] This invention belongs to the technical field of new energy vehicle control and transmission systems, specifically relating to a system and method for suppressing malfunctions of the power take-off unit in new energy vehicles. Background Technology

[0002] With the iterative upgrade of new energy vehicle technology, the power take-off (PTO), as a key device for commercial and special-purpose vehicles to achieve external power output, has its control system stability and safety crucial to the overall vehicle's operational efficiency. Under a high-voltage electrification architecture, the PTO system needs to achieve precise power distribution in complex power transmission paths, and its operating status affects the energy matching efficiency between the electric drive system and the mechanical operating end.

[0003] PTO (Power Take-Off) malfunction suppression technology aims to prevent unintended engagement under illegal operating conditions by monitoring multi-source vehicle signals, thus ensuring the safety of the drivetrain and electrical circuits. In modern electronic control systems, this technology requires the system to be able to sense the dynamic coupling status of the drive motor, high-voltage battery pack, and transmission mechanism in real time, in order to construct a multi-dimensional safety operating boundary covering all operating conditions.

[0004] Existing suppression strategies largely rely on static physical constraints such as vehicle speed or parking status, lacking global perception capabilities for scenarios involving deep coupling between high-voltage electrical dynamics and mechanical transmission dynamics. This makes it difficult to identify false trigger commands caused by electromagnetic interference under complex operating conditions. Traditional solutions generally suffer from diagnostic lag, failing to implement predictive protection in the early stages of malfunction trends, leading to frequent damage to gear mechanisms from instantaneous mechanical shocks. Furthermore, the controllers in the power domain maintain only a single command transmission logic, lacking a collaborative arbitration mechanism to address electromechanical coupling risks, and thus failing to mitigate the risk of power backflow in the high-voltage system during abnormal interactions. Summary of the Invention

[0005] The purpose of this invention is to provide a power take-off (PTO) malfunction suppression system for new energy vehicles, which can solve the problems of instantaneous mechanical impact damage to the transmission system and safety risks of the high-voltage electrical system in the above-mentioned background art due to the single perception dimension, lagging diagnostic mechanism and insufficient coordination of various control nodes in the scenario of deep coupling between high-voltage electrical dynamics and mechanical transmission dynamics of new energy vehicles.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0007] A power take-off (PTO) malfunction suppression system for new energy vehicles includes a motor controller, a vehicle controller, a PTO controller, a PTO mechanical assembly, a drive axle housing assembly, a high-voltage power distribution unit, and a multi-source information sensing array, as follows:

[0008] The multi-source information sensing array is used to monitor the electrical physical quantity characteristics and mechanical dynamic characteristics in the power domain of new energy vehicles in real time, and transmits the collected raw physical signals to the vehicle controller after preprocessing.

[0009] The motor controller is used to receive the power demand command from the vehicle controller and control the operation of the drive motor. It uses its internally integrated electrical feature extraction unit to analyze the transient current fluctuations in the high-voltage circuit of the drive motor in order to identify potential illegal gear shifting electrical signal characteristics.

[0010] The vehicle controller is used to construct an electromechanical coupling collaborative arbitration architecture. It performs cross-domain fusion analysis on the data fed back by the multi-source information sensing array through the built-in physical information neural network model, establishes a dynamic mapping between electrical ripple characteristics and mechanical vibration spectrum, and outputs targeted multi-level suppression commands based on the mapping results.

[0011] The power take-off controller is used to control the action of the power take-off actuator according to the arbitration result of the vehicle controller, and to perform a physical lock-up operation when receiving a suppression command, so as to block the unexpected power coupling path.

[0012] The power take-off mechanical assembly is configured at the power output end of the drive motor or transmission system, and is used to engage and disengage power while meeting preset safety criteria.

[0013] The drive axle housing assembly serves as a carrier for mechanical vibration transmission, bearing the operating load of the transmission system and providing a mounting base for the vibration sensing element.

[0014] Preferably, the multi-source information sensing array includes a high-bandwidth current sampling sensor and a miniature broadband vibration sensor. The high-bandwidth current sampling sensor is deployed at the three-phase output terminal of the motor controller to capture transient current vector information of the drive motor under different load conditions; the miniature broadband vibration sensor is arranged on the housing surface of the power take-off mechanical assembly and the support part of the adjacent drive axle housing assembly to collect high-frequency vibration spectrum of the mechanical transmission chain in real time during engagement, disengagement and operation.

[0015] Furthermore, the motor controller integrates a high-bandwidth current ripple analysis unit. This unit not only monitors the average current intensity of the drive motor but also extracts the high-frequency ripple components from the three-phase current of the motor in real time. When the driver accidentally triggers the warning or abnormal electromagnetic interference in the system circuit causes the PTO to engage, the high-bandwidth current ripple analysis unit, based on the specific electromagnetic induction characteristics generated by the pre-drive circuit, identifies and predicts the upcoming electrical command trends that may lead to abnormal PTO engagement.

[0016] Furthermore, the vehicle controller is internally configured with a physical information neural network model. This model uses the current ripple characteristics of the electrical domain as the first input feature vector and the vibration spectrum characteristics of the mechanical domain as the second input feature vector. The physical information neural network model analyzes the spatiotemporal correlation between the first and second input feature vectors in the time and frequency dimensions by establishing a deep correlation logic based on physical constraints.

[0017] Furthermore, based on the output of the physical information neural network model, the vehicle controller implements a three-level progressive cooperative inhibition strategy. This three-level progressive cooperative inhibition strategy includes a predictive soft inhibition stage, a cooperative locking stage, and an adaptive evolution stage.

[0018] Furthermore, in the predictive soft suppression phase, when the physical information neural network model determines that the current operating condition is highly likely to be a misoperation trend, the vehicle controller does not immediately cut off the system power, but instead sends a torque dynamic compensation command to the motor controller. Based on this command, the motor controller dynamically adjusts the output torque of the drive motor at the microsecond level within a predetermined time. The phase of this dynamic adjustment is inversely correlated with the mechanical vibration characteristics collected by the miniature broadband vibration sensor. Through energy regulation at the electrical end, it actively counteracts the minor mechanical impacts caused by the mis-engaging gear trend, achieving soft contact protection of the mechanical interface.

[0019] Furthermore, during the cooperative interlocking phase, if, after implementing the predictive soft suppression, the vehicle controller detects a continuous increase in the intensity of the erroneous operation trend, and the amplitudes of the first and second input feature vectors synchronously exceed a preset safety boundary threshold, then the vehicle controller synchronously sends a forced interlocking command to the power take-off controller and instructs the motor controller to rapidly reduce the motor output torque to zero. The forced interlocking command will create an irreversible physical suppression state by cutting off the power supply to the solenoid valve of the power take-off actuator or locking the shift fork position, ensuring complete isolation between the high-voltage system and the mechanical transmission system.

[0020] Furthermore, in the adaptive evolution phase, the vehicle controller is equipped with an online learning unit. The online learning unit uses each successfully suppressed current ripple feature data pair and vibration feature data pair as positive guiding samples, feeding them back to the physical information neural network model for iterative update of parameter weights. The online learning unit uses feature events that did not trigger suppression but are in a critical safety state as negative reference samples to dynamically optimize the feature mapping accuracy and the response threshold of the suppression strategy in the next cycle.

[0021] Furthermore, when performing feature mapping, the physical information neural network model extracts the energy density distribution corresponding to the fork movement frequency, spline sleeve sliding frequency, and gear pre-meshing characteristic frequency from the vibration spectrum, and combines this with the harmonic distortion rate corresponding to the electromagnetic commutation frequency and inverter switching frequency in the current ripple to calculate the cross-domain feature overlap. When the cross-domain feature overlap is greater than a preset logical judgment probability, the system confirms the risk of misoperation.

[0022] Furthermore, the motor controller is also equipped with high-voltage backflow protection logic. When the power take-off mechanical assembly experiences an unexpected engagement that causes a sudden change in external load feedback energy, the motor controller, in conjunction with the high-voltage power distribution unit, adjusts the bus current path within a predetermined time to prevent instantaneous power backflow from causing electrical damage to the battery pack and high-voltage circuit.

[0023] Furthermore, the vehicle controller is equipped with a global perception matrix, which integrates vehicle speed signals, handbrake status signals, brake pedal travel signals, gear information, and high-voltage battery pack status parameters in real time, and uses them as boundary constraints for the physical information neural network model. When the vehicle is not stationary and the braking system is not activated, the global perception matrix automatically increases the sensitivity coefficient for detecting erroneous operations.

[0024] Furthermore, the power take-off (PTO) mechanical assembly includes a shift fork position sensor, and the signal from the shift fork position sensor is connected to the PTO controller. The PTO controller feeds back the actual shift fork position information to the vehicle controller to verify the execution effect of the three-level progressive cooperative inhibition strategy and to serve as the basis for closed-loop feedback control in the adaptive evolution stage.

[0025] Furthermore, the torque fluctuation amplitude generated by the torque dynamic compensation command is limited to a specific proportion of the current rated torque of the drive motor to ensure that the vehicle's driving state will not fluctuate drastically during the implementation of predictive soft suppression, thus ensuring both the safety of the transmission chain and driving smoothness.

[0026] Furthermore, the physical information neural network model is also equipped with an environmental adaptive compensation factor to correct for vibration spectrum drift caused by changes in lubricating oil temperature, fluctuations in ambient temperature, and mechanical wear. This compensation factor adjusts the gain of the second input feature vector by monitoring the value of the power take-off lubricating oil temperature sensor.

[0027] A method for suppressing malfunctions of the power take-off (PTO) in new energy vehicles is provided, which uses the aforementioned new energy vehicle PTO malfunction suppression system to suppress malfunctions of the PTO in new energy vehicles.

[0028] Compared with the prior art, the present invention has the following beneficial effects:

[0029] 1. This invention overturns the inevitability of mechanical shock in traditional technologies by constructing a collaborative suppression architecture based on the mapping of motor current and gear vibration characteristics. By introducing a predictive soft suppression strategy, before the impact energy generated by misoperation is transmitted to the core components of the power take-off transmission, the dynamic torque compensation function of the drive motor actively offsets energy fluctuations in the electrical domain, achieving shock-free suppression. This extends the fatigue life of key mechanical components such as transmission gears, shift forks, and spline sleeves, elevating the reliability of the transmission system to a whole new level.

[0030] 2. This invention achieves deep collaboration and harmonious operation between the high-voltage electrical system and the mechanical transmission system of new energy vehicles. This solution no longer treats the electromagnetic dynamics of the high-voltage drive system as an isolated source of interference, but rather transforms it into an active means of suppressing mechanical shock. By deeply analyzing the cross-domain correlation between current ripple and mechanical vibration, it breaks down the information silos between vehicle control, motor control, and power take-off control in the traditional architecture, constructing a dynamic safety boundary covering the entire electromechanical coupling process, and enhancing the system's operational robustness under complex operating conditions, strong electromagnetic interference, and extreme vibration environments.

[0031] 3. This invention endows the suppression system with superior predictive perception capabilities and self-evolutionary learning abilities. By integrating a physical information neural network model and an online learning unit, the system can identify and predict the initial characteristics of erroneous operations. With the continuous accumulation of operational data, the system can automatically adapt to new electromagnetic environments, different driver operating habits, and the aging characteristics of mechanical structures. Its prediction accuracy and response efficiency optimize exponentially with usage intensity, providing a solid technical guarantee for the intelligent safety protection of new energy commercial vehicles.

[0032] 4. This invention, through a three-level progressive strategy design, reduces the impact of system intervention on operational continuity while ensuring absolute safety. By organically combining soft inhibition, cooperative interlocking, and self-learning, it prevents hardware damage caused by accidental triggering and avoids the risk of frequent downtime caused by traditional hard-cutoff strategies, achieving an optimized balance between operational efficiency and equipment safety. Attached Figure Description

[0033] Figure 1 This is a schematic diagram of the overall technical solution architecture of the present invention;

[0034] Figure 2 This is a schematic diagram of the core principle framework of the present invention, which uses a physical information neural network to analyze the cross-domain spatiotemporal correlation between electrical ripple characteristics and mechanical vibration spectrum.

[0035] Figure 3 This is a schematic diagram of the multi-level interaction relationship and data flow between the multi-source information sensing array, the vehicle controller and the execution control node in this invention;

[0036] Figure 4 This is a logical flowchart of the three-level progressive cooperative inhibition strategy of predictive soft inhibition, cooperative locking and adaptive evolution implemented in this invention.

[0037] Figure 5 This is a flowchart illustrating the logical flow of the present invention, which uses dynamic fine-tuning of drive motor torque and phase compensation of mechanical vibration characteristics to achieve soft contact protection of mechanical interface.

[0038] Figure 6 This is a flowchart of the adaptive evolutionary stage logic process based on positive and negative sample backpropagation and iterative update of neural network parameter weights in this invention. Detailed Implementation

[0039] Example 1: Please refer to the appendix Figure 1 To be continued Figure 6 A new energy vehicle power take-off (PTO) malfunction suppression system includes a multi-source information sensing array, a motor controller, a vehicle controller, a PTO controller, a PTO mechanical assembly, a drive axle housing assembly, and a high-voltage power distribution unit.

[0040] The multi-source information sensing array is used to monitor the electrical and physical characteristics and mechanical dynamics characteristics within the power domain of new energy vehicles in real time, and transmits the pre-processed raw physical signals to the vehicle controller. The multi-source information sensing array includes a high-bandwidth current sampling sensor and a miniature wideband vibration sensor. The high-bandwidth current sampling sensor is deployed at the three-phase output of the motor controller, and its sampling accuracy can capture weak harmonic components during dynamic switching of the drive motor, providing a high-fidelity data source for subsequent electrical feature extraction. The miniature wideband vibration sensor uses a high-sensitivity piezoelectric accelerometer or a microelectromechanical system vibration chip, and is installed at key stress points on the power take-off mechanical assembly housing and near the bearing support of the drive axle housing assembly. This sensor can collect vibration signals in real time, covering a frequency range from low-frequency mechanical motion to high-frequency gear meshing impact, ensuring that the system can detect any minute physical displacement or collision precursors in the transmission chain.

[0041] The motor controller receives power demand commands from the vehicle controller and controls the operation of the drive motor. It utilizes its integrated electrical feature extraction unit to analyze transient current fluctuations in the high-voltage circuit of the drive motor to identify potential illegal gear engagement signals. The motor controller integrates a high-performance digital signal processing chip and a high-bandwidth current ripple analysis unit. This analysis unit does not merely monitor the macroscopic bus current or phase current RMS values, but extracts high-frequency ripple characteristics in the three-phase AC current in real time using Fast Fourier Transform or Wavelet Transform techniques. In the early stages of power take-off (PTO) malfunction, even before the drive command is fully established, specific electromagnetic disturbances may occur between the inverter's gate drive circuit and the power module due to the triggering of illegal commands. These disturbances couple into the motor stator windings in the form of weak current ripples. By identifying these specific ripple characteristics, the motor controller can predict potential malfunction risks within milliseconds before physical action occurs.

[0042] The vehicle controller is used to construct an electromechanical coupling collaborative arbitration architecture. It performs cross-domain fusion analysis on the data fed back from the multi-source information sensing array through a built-in physical information neural network model, establishing a dynamic mapping between electrical ripple characteristics and mechanical vibration spectrum, and outputs targeted multi-level suppression commands based on the mapping results. As the core computing node of the entire system, the vehicle controller adopts a multi-core parallel processing architecture to support the real-time operation of the physical information neural network model. This model not only possesses the data-driven capabilities of traditional neural networks but also introduces physical equations based on gear dynamics and motor electromagnetic constraints. By spatiotemporally aligning the feature vectors of the electrical domain with the spectral characteristics of the mechanical domain, the vehicle controller can determine whether the current signal disturbance is caused by normal load fluctuations or by a trend risk resulting from illegal gear shifting commands. This cross-domain collaborative arbitration mechanism breaks the isolation between electrical and mechanical systems in traditional control systems, improving the robustness of misoperation identification.

[0043] The power take-off (PTO) controller controls the movement of the PTO actuator based on the arbitration result of the vehicle controller, and performs a physical lock-up operation upon receiving a suppression command to block unintended power coupling paths. The PTO controller is connected to the vehicle controller via a dedicated communication bus and integrates a drive circuit to control the electromagnetic directional valve or servo motor in the PTO actuator. Under normal operating conditions, the PTO controller responds to the driver's gear-shifting request; however, upon detecting a tendency for misoperation, the controller immediately enters the highest-priority safety lock-up mode, ensuring that the mechanical shift fork cannot complete physical displacement by cutting off the drive circuit power or applying a reverse electromagnetic force, thus achieving absolute protection of the mechanical transmission chain.

[0044] The power take-off (PTO) mechanical assembly is located at the power output end of the drive motor or transmission system, and is used to engage and disengage power while meeting preset safety criteria. Internally, it includes a shift fork mechanism, a spline sleeve, a synchronizer, and a gear pair. The PTO mechanical assembly is also equipped with a high-precision position sensor to feed back the actual physical position of the mechanical components to the PTO controller in real time, forming a closed-loop control circuit.

[0045] The drive axle housing assembly, serving as a carrier for mechanical vibration transmission, bears the operating load of the transmission system and provides a mounting base for the vibration sensing element. The structural rigidity of the drive axle housing assembly is optimized to transmit weak vibration signals generated by gear contact. Its surface has a pre-drilled mounting plane for the sensor, ensuring good acoustic coupling between the miniature broadband vibration sensor and the housing, reducing signal attenuation and distortion.

[0046] During system operation, the multi-source information sensing array performs high-frequency data acquisition. The high-bandwidth current sampling sensor captures the three-phase vector components of the stator current of the drive motor in real time, and its data update frequency is synchronized with the pulse width modulation frequency of the motor controller. The miniature broadband vibration sensor acquires the housing vibration acceleration signal at a sampling rate of not less than 50 kHz. After low-noise amplification, filtering, and analog-to-digital conversion, the acquired signals are aggregated to the vehicle controller via in-vehicle high-speed Ethernet or a flexible controller area network bus.

[0047] The high-bandwidth current ripple analysis unit integrated within the motor controller possesses in-depth feature analysis capabilities. This unit not only monitors the average current intensity of the drive motor but also extracts the high-frequency ripple components from the three-phase current of the motor in real time. When the driver intentionally triggers the system or abnormal electromagnetic interference causes a PTO engagement warning, the high-bandwidth current ripple analysis unit, based on the specific electromagnetic induction characteristics generated by the pre-drive circuit, identifies and predicts the upcoming electrical command trends that may lead to abnormal PTO engagement. This prediction process, based on gradient analysis of the current envelope and identification of the energy distribution of specific harmonic orders, can complete risk warnings in a short time before the electrical command is converted into mechanical thrust.

[0048] The physical information neural network model configured inside the vehicle controller receives current ripple characteristics from the electrical domain as the first input feature vector and vibration spectrum characteristics from the mechanical domain as the second input feature vector. The physical information neural network model analyzes the spatiotemporal correlation between the first and second input feature vectors in the time and frequency dimensions by establishing deep correlation logic based on physical constraints. During the calculation process, the model extracts the energy density distribution corresponding to the shift fork movement frequency, spline sleeve sliding frequency, and gear pre-meshing characteristic frequency from the vibration spectrum, and combines this with the harmonic distortion rate corresponding to the electromagnetic commutation frequency and inverter switching frequency from the current ripple to calculate the cross-domain feature overlap. When the cross-domain feature overlap exceeds a preset logical judgment probability, the system confirms a risk of misoperation.

[0049] The vehicle controller implements a three-level progressive cooperative inhibition strategy based on the output of the physical information neural network model. The three-level progressive cooperative inhibition strategy includes a predictive soft inhibition stage, a cooperative locking stage, and an adaptive evolution stage.

[0050] In the predictive soft suppression phase, when the physical information neural network model determines that the current operating condition is highly likely to be a misoperation trend, the vehicle controller does not immediately cut off the system power, but instead sends a torque dynamic compensation command to the motor controller. Based on this command, the motor controller dynamically fine-tunes the output torque of the drive motor within a predetermined extremely short time. The phase of this dynamic fine-tuning is inversely correlated with the mechanical vibration characteristics collected by the miniature broadband vibration sensor. That is, by actively generating a weak torque fluctuation at the motor output end that is opposite to the mechanical impact trend, the principle of energy cancellation is used to actively counteract the minor mechanical impact caused by the mis-engagement trend. This strategy achieves soft contact protection of the mechanical interface, preventing violent physical impacts before the gears are fully engaged. The torque fluctuation amplitude generated by the torque dynamic compensation command is limited to a specific percentage range of the current rated torque of the drive motor, for example, between 3% and 8%, to ensure that the predictive soft suppression process does not cause violent fluctuations in the vehicle's driving state, ensuring both drivetrain safety and driving smoothness.

[0051] During the coordinated interlocking phase, if, after implementing the predictive soft suppression, the vehicle controller detects a continuous increase in the intensity of the erroneous operation trend, and the amplitudes of the first and second input feature vectors synchronously exceed a preset safety boundary threshold, then the vehicle controller synchronously sends a forced interlocking command to the power take-off controller and instructs the motor controller to rapidly reduce the motor output torque to zero. The forced interlocking command will create an irreversible physical suppression state by cutting off the power supply to the solenoid valve of the power take-off actuator or locking the shift fork position, ensuring complete isolation between the high-voltage system and the mechanical transmission system. The system will activate the alarm module on the instrument panel, alerting the driver to the risk of illegal operation and that the system has entered a protection state.

[0052] In the adaptive evolution phase, the vehicle controller is equipped with an online learning unit. This unit uses successfully suppressed current ripple feature data pairs and vibration feature data pairs as positive guidance samples, feeding them back to the physical information neural network model for iterative updates of parameter weights. This mechanism allows the model to self-adjust based on the specific vehicle's mechanical wear and sensor aging drift. The online learning unit uses feature events that did not trigger suppression but were in a critical safety state as negative reference samples to dynamically optimize the feature mapping accuracy and the suppression strategy's response threshold for the next cycle. Through this continuous self-learning process, the system can memorize new electromagnetic interference patterns or irregular driving habits, continuously improving the accuracy of predictions.

[0053] Furthermore, the vehicle controller is equipped with a global perception matrix, which integrates vehicle speed signals, handbrake status signals, brake pedal travel signals, gear information, and high-voltage battery pack status parameters in real time, and uses these as boundary constraints for the physical information neural network model. For example, when the global perception matrix detects that the vehicle speed is greater than a preset zero-speed threshold and the braking system is not activated, the system automatically increases the sensitivity coefficient for detecting erroneous operations. This global perception based on the vehicle's operating conditions provides additional logical boundaries for the physical information neural network model, preventing any form of power coupling attempt under high-speed driving conditions.

[0054] The motor controller is also equipped with high-voltage backflow protection logic. When the power take-off (PTO) mechanical assembly experiences an unexpected engagement that causes a sudden change in external load feedback energy, the motor controller, in conjunction with the high-voltage power distribution unit, adjusts the bus current path within a predetermined time. If the rise rate of the DC bus voltage exceeds a preset voltage fluctuation limit, the motor controller will immediately adjust the inverter's vector control parameters to convert excess kinetic energy into the motor's internal energy or absorb it through the discharge circuit in the high-voltage power distribution unit, preventing instantaneous power backflow from causing electrical damage to the battery pack and high-voltage circuit.

[0055] The physical information neural network model is also equipped with an environmental adaptive compensation factor to correct vibration spectrum drift caused by changes in lubricating oil temperature, fluctuations in ambient temperature, and mechanical wear. This compensation factor adjusts the gain of the second input feature vector by monitoring the values ​​from the power take-off lubricating oil temperature sensor. Since the viscosity difference of lubricating oil at different temperatures alters the vibration transmission characteristics during gear meshing, this compensation factor ensures that the system maintains consistent recognition accuracy even under extreme cold or prolonged high-temperature operating conditions.

[0056] The power take-off (PTO) controller also features communication link redundancy monitoring. When a communication interruption is detected on the main bus with the vehicle controller, or when the message loss rate exceeds a preset 1% threshold, the PTO controller will automatically trigger a preset fail-safe protection logic. This will force the PTO actuator to remain in its current disengaged state and cut off all automatic control permissions until communication is restored and the system self-test is passed. This function ensures that even in extremely harsh electromagnetic environments or when the wiring harness is damaged, the PTO will not malfunction due to missing or disordered control signals.

[0057] Example 2: Example 2 provides a new energy vehicle power take-off (PTO) malfunction suppression system based on a distributed edge computing architecture. Building upon Example 1, this system further optimizes the real-time performance of signal processing and the redundancy and reliability of the system, making it particularly suitable for heavy-duty new energy engineering vehicles with multiple power source couplings or multiple PTO output terminals.

[0058] The system includes edge-aware computing nodes, master control arbitration nodes, motor drive control units, distributed power take-off execution modules, and a multi-dimensional physical field perception network.

[0059] The multi-dimensional physical field sensing network is distributed across key nodes of the vehicle's powertrain. In addition to the three-phase current sensor and vibration sensor mentioned in Example 1, this example also adds an electromagnetic field strength sensing probe and an ultrasonic position sensing unit. The electromagnetic field strength sensing probe is positioned at the intersection of the high-voltage cable bundle and the power take-off solenoid valve, used to monitor the mutual inductance of the control circuit in real time. The ultrasonic position sensing unit is positioned laterally on the spline sleeve inside the power take-off, using non-contact high-frequency ultrasonic reflection ranging to monitor the minute axial displacement of the spline sleeve in real time, achieving a resolution down to the micrometer level, enabling it to detect mechanical displacement trends earlier than traditional shift fork position sensors.

[0060] The edge-sensing computing node, physically located near the power take-off (PTO) mechanical assembly, is used for local processing of the raw signals acquired by the multi-dimensional physical field sensing network. The edge-sensing computing node integrates a high-speed field-programmable gate array (FPGA) chip and is configured to execute a feature extraction algorithm based on streaming data. This node can convert complex, high-sampling-rate vibration spectra and electromagnetic pulse signals into low-dimensional feature descriptors in real time. Through rapid analysis at the edge, environmental background noise and conventional mechanical vibrations are eliminated, and only statistically significant anomalous feature data are reported to the main control arbitration node. This distributed design reduces the communication load on the vehicle backbone network and shortens the physical response delay from risk perception to issuing suppression commands.

[0061] The master arbitration node, acting as the central nervous system, encompasses all arbitration logic of the vehicle controller in Embodiment 1 and incorporates multimodal data alignment technology. The master arbitration node receives feature streams from multiple edge-aware computing nodes and ensures alignment of electrical and mechanical domain signals on the time axis using a high-precision timestamp synchronization protocol (such as the TSN protocol of automotive Ethernet), with alignment errors controlled to the microsecond level. The physical information neural network model built into this node adopts a hierarchical architecture: the first layer is a feature fusion layer, used for nonlinearly combining electrical, mechanical, and magnetic three-dimensional features; the second layer is a physical constraint mapping layer, which verifies the physical rationality of the feature combination through a preset dynamic transfer function; and the third layer is a decision output layer, which outputs multi-level suppression instructions based on the fused risk assessment value.

[0062] Upon receiving a suppression command from the main control arbitration node, the motor drive control unit, in addition to performing dynamic torque compensation, also possesses a harmonic active injection function. This function involves superimposing a harmonic current of a specific frequency and amplitude onto the fundamental current of the drive motor. This harmonic current generates tiny, high-frequency electromagnetic pulses between the motor stator and rotor. These pulses are transmitted to the power take-off via the drive shaft, producing a high-frequency, micro-amplitude chatter. This chatter disrupts the static friction between gears under erroneous operating conditions, making it easier for the spline sleeve to spring back to a safe position when subjected to unexpected thrust, thus assisting the mechanical locking mechanism in achieving the suppression effect.

[0063] The distributed power take-off (PTO) execution module includes multiple independent PTO controllers and corresponding mechanical actuators. For vehicles with a dual PTO architecture, this module can achieve independent arbitration and coordinated suppression. When the master arbitration node determines that only one PTO is at risk of malfunction, the distributed PTO execution module can perform targeted locking without affecting the normal operation of the other PTO. Each execution module has a built-in independent energy storage unit (such as a large-capacity capacitor) to ensure that even in extreme situations such as a vehicle power outage or loss of main communication, sufficient power can still be maintained to complete the final safety locking action.

[0064] In this embodiment, the predictive soft suppression strategy is further refined. The master arbitration node calculates the energy correlation coefficient in real time, which reflects the degree of matching between the compensation energy injected at the motor end and the impact energy sensed at the mechanical end. By adjusting the frequency response of the torque compensation command in a closed loop, the system can achieve adaptive damping adjustment. When the mechanical vibration exhibits a nonlinear increasing trend, the system automatically increases the gain coefficient of the torque compensation and introduces a predictive differential term to offset the phase difference caused by signal processing delay in advance, ensuring accurate offsetting of energy compensation at the physical level.

[0065] For the adaptive evolution phase, this embodiment introduces a cloud-based collaborative learning mechanism. The master arbitration node periodically uploads its locally extracted malfunction feature fingerprint database to the cloud server via a remote vehicle communication terminal (T-Box). The cloud server aggregates massive malfunction samples from different vehicle models and operating environments, and iteratively optimizes the global feature mapping model using deeper deep learning algorithms. The optimized model parameters are then distributed to each vehicle via over-the-air (OTA) updates, enabling individual vehicles' suppression systems to share the collective wisdom of the entire fleet, achieving proactive defense against rare malfunctions or extreme environmental interference.

[0066] The system also includes a thermal management coordination unit. Considering the potential heat accumulation in the power take-off (PTO) under continuous high-load operation or frequent suppression actions, the thermal management coordination unit monitors the temperature of relevant control circuits and actuators in real time. When the temperature exceeds a preset first-level thermal warning threshold, the system automatically adjusts the suppression strategy threshold, adopting a more conservative lockout strategy to reduce the thermal load on power electronic components. When the temperature exceeds a second-level thermal protection threshold, the system will force entry into a derating operation mode, limiting the motor's output response speed until the temperature returns to normal.

[0067] The distributed edge computing architecture also provides a fault tolerance mechanism. The master arbitration node monitors the health status of each edge sensing computing node in real time. If an edge node fails, the master arbitration node will immediately switch to a backup mode based on feature extraction from the motor drive control unit. Although the sensing dimension is reduced at this time, the system can still maintain basic malfunction suppression capabilities by relying on high-bandwidth ripple analysis at the motor end and logical constraints of the global sensing matrix, thus achieving deep defense and high availability of the system.

[0068] In terms of mechanical optimization, the power take-off (PTO) mechanical assembly employs a synchronization structure with variable stiffness characteristics. When receiving minute torque fluctuations during the soft-suppression phase, this synchronization structure utilizes its internal nonlinear spring elements to generate damping dissipation, converting minute mechanical energy into heat energy, further reducing the impact pressure on the tooth surfaces and splines. This electromechanical coupling energy-absorbing design reduces the instantaneous stress level of the mechanical transmission chain compared to traditional solutions when suppressing extreme misoperation scenarios.

[0069] Example 3: Example 3 details a new energy vehicle power take-off (PTO) malfunction suppression system with deep electrical topology monitoring and mechanical resonance suppression functions. This example focuses on improving robustness in complex electromagnetic compatibility environments and multi-frequency resonance conditions.

[0070] The system includes a comprehensive electrical signal analysis module, a mechanical dynamic compensation module, a machine-to-machine / electrical cross-domain arbitration module, and a self-learning security strategy library.

[0071] The integrated electrical signal analysis module is integrated into the front-end power distribution unit of the motor controller. It not only monitors the ripple of the drive current but also extends to real-time monitoring of the bus ripple voltage and the neutral point-to-ground potential. In the high-voltage systems of new energy vehicles, frequent switching of the inverter generates common-mode current, while malfunctioning commands from the power take-off (PTO) are often accompanied by voltage drops at specific frequencies in the control loop. The integrated electrical signal analysis module establishes a feature recognition model based on the power balance equation by analyzing the energy interaction characteristics of the DC and AC sides. This model can distinguish between power fluctuations caused by normal motor acceleration and energy anomalies caused by abnormal closure of the PTO drive circuit. The logic of this model can be described as follows: when the difference between the monitored DC bus power variation and the three-phase AC output power variation deviates from the preset loss compensation range, and the rate of change of this deviation conforms to the establishment characteristics of the PTO electromagnetic load, the system determines it as a malfunction symptom in the electrical domain.

[0072] The mechanical dynamic compensation module is deployed on the power take-off housing and drive shaft support. In addition to a high-bandwidth vibration sensor, the module is also equipped with an active damping actuator. This active damping actuator can generate localized high-frequency reverse vibrations according to the instructions of the electromechanical cross-domain arbitration module. During the predictive soft suppression phase, the module relies not only on the motor's torque compensation but also simultaneously initiates the damping action at the mechanical end. This dual-end energy cancellation mechanism can accurately cover all levels of resonance points in the transmission system. When the physical information neural network model identifies that the current vibration spectrum is mainly concentrated near the power take-off's first-level meshing frequency, the mechanical dynamic compensation module calculates the optimal compensation phase at that frequency and instructs the actuator to generate destructive interference, forming a dynamic vibration vacuum zone at the mechanical contact interface, mitigating the hard impact during gear shifting.

[0073] The electromechanical cross-domain arbitration module adopts an arbitration architecture based on a hybrid of fuzzy logic and neural networks. This module operates independently within the vehicle controller, and its core algorithm is configured to fuse the predicted probability of commands from the electrical domain with the probability of physical motion from the mechanical domain. This fusion process follows these principles: if the electrical domain predicts a high-probability erroneous command, the system will immediately enter a warning mode and limit the gradient of motor torque changes, even if the mechanical domain has not yet detected physical displacement; if the mechanical domain detects an unexpected physical displacement trend, even if the electrical domain command is normal, the system will suspect a failure of the mechanical interlock mechanism or external physical intervention and initiate an emergency interlock. This logic ensures that the system can respond promptly regardless of whether the fault originates from the electrical or mechanical side.

[0074] The self-learning safety strategy library is stored in non-volatile memory and connected to the online learning unit. In this embodiment, the strategy library has multi-dimensional scene classification capabilities. It provides different weight templates for the physical information neural network model based on the vehicle's current load, environmental slope, transmission gear position, and historical operation records. For example, under heavy-load hill start conditions, the background noise of mechanical vibration is extremely high. The strategy library automatically increases the weight of electrical features in arbitration and raises the filtering threshold for mechanical vibration to prevent false suppression actions triggered by normal heavy-load vibration. This scene-based dynamic weight adjustment enables the system to maintain an extremely low false trigger rate even under complex operating conditions.

[0075] Furthermore, the motor controller integrates an active suppression function based on virtual impedance. Upon receiving a malfunction warning, the motor controller adjusts the control parameters of the current loop to simulate a large moment of inertia and viscous damping coefficient in the motor's electrical characteristics. This causes the drive motor to exhibit motion resistance when faced with a sudden power take-off (PTO) engagement request. From a physical perspective, even if the PTO spline sleeve attempts to force engagement, the motor will maintain speed stability due to the existence of this virtual impedance, limiting the energy fluctuations generated during engagement to a very small range.

[0076] This embodiment also introduces a transient surge suppression module for the high-voltage power distribution unit. During the coordinated interlocking phase, if an emergency disconnection of the high-voltage contactor is involved, this module can absorb the inductive energy stored in the inductor due to the sudden change in current within microseconds through its built-in pre-charging circuit and residual voltage discharge logic. This protects the contactor contacts from arc erosion and also prevents electromagnetic impacts from high-voltage pulses on the vehicle's low-voltage control network. The protection logic can be described as follows: when the interlocking command is triggered and the bus current intensity is greater than the preset disconnection safety threshold, the system first performs a torque zeroing action, and then performs a physical disconnection after the current drops to a safe range, achieving a soft shutdown at the electrical level.

[0077] The adaptive evolution phase of the system also includes a predictive diagnostic function for mechanical wear. The online learning unit calculates the wear trends of the synchronizer and gears by tracking the vibration spectrum drift during the PTO engagement process over a long period. When the wear level approaches the preset maintenance limit, the system not only corrects the response parameters of the suppression strategy but also sends a maintenance warning to the backend via the vehicle diagnostic bus. This extension from misoperation suppression to full lifecycle health management further demonstrates the depth of the system's mechatronics design.

[0078] The vehicle controller is equipped with a high-priority hardware watchdog circuit. This circuit operates independently of the main processor and is specifically responsible for monitoring the operational status of the malfunction suppression function. If the main control algorithm crashes or enters a logic trap, the watchdog circuit bypasses the software control layer and forces the power take-off controller into a locked state via a hardwired hardware connection. This combined hardware and software defense system provides the final physical guarantee for the safe operation of the power take-off system in new energy vehicles.

[0079] Those skilled in the art should understand that the specific embodiments described above are only for illustrating the technical solutions of the present invention, and not for limiting it. Without departing from the spirit and scope of the present invention, various equivalent substitutions, combinations, or optimizations can be made to the hardware components, software algorithms, communication protocols, and physical parameters in the above embodiments. For example, the physical information neural network model in this embodiment can be replaced with other deep learning models with multi-physics constraints, or the ripple analysis algorithm of the motor controller can be applied to other types of electric drive systems. All such modifications, equivalent substitutions, or improvements should be included within the scope of the claims of the present invention. The module division described in the system is only a functional illustration; in actual applications, it can be merged or further refined according to the computing power distribution of the hardware platform. The preferred parameters and thresholds described in the specification are reference values ​​for engineering implementation; in specific applications, they need to be calibrated based on measured data from different vehicle models and different operating conditions.

Claims

1. A power take-off (PTO) misoperation suppression system for new energy vehicles, characterized in that, This includes the motor controller, vehicle controller, power take-off controller, power take-off mechanical assembly, drive axle housing assembly, high-voltage power distribution unit, and multi-source information sensing array, among which: The multi-source information sensing array is used to monitor the electrical physical quantity characteristics and mechanical dynamic characteristics in the power domain of new energy vehicles in real time, and transmits the collected raw physical signals to the vehicle controller after preprocessing. The motor controller is used to receive the power demand command from the vehicle controller and control the operation of the drive motor. At the same time, it uses its internally integrated electrical feature extraction unit to analyze the transient current fluctuations in the high-voltage circuit of the drive motor in order to identify potential illegal gear shifting electrical signal characteristics. The vehicle controller is used to construct an electromechanical coupling collaborative arbitration architecture. It performs cross-domain fusion analysis on the data fed back by the multi-source information sensing array through the built-in physical information neural network model, establishes a dynamic mapping between electrical ripple characteristics and mechanical vibration spectrum, and outputs targeted multi-level suppression commands based on the mapping results. The power take-off controller is used to control the action of the power take-off actuator according to the arbitration result of the vehicle controller, and to perform a physical lock-up operation when receiving a suppression command, so as to block the unexpected power coupling path. The power take-off mechanical assembly is configured at the power output end of the drive motor or transmission system, and is used to engage and disengage power while meeting preset safety criteria. The drive axle housing assembly serves as a carrier for mechanical vibration transmission, bearing the operating load of the transmission system and providing a mounting base for the vibration sensing element.

2. The new energy vehicle power take-off misoperation suppression system according to claim 1, characterized in that: The multi-source information sensing array includes a high-bandwidth current sampling sensor and a miniature broadband vibration sensor. The high-bandwidth current sampling sensor is deployed at the three-phase output of the motor controller to capture transient current vector information of the drive motor under different load conditions. Its sampling accuracy is configured to capture harmonic components of the drive motor during dynamic switching. The miniature broadband vibration sensor, employing a piezoelectric accelerometer or a microelectromechanical system vibration chip, is arranged on the housing surface of the power take-off mechanical assembly and the bearing support of the drive axle housing assembly. It is used to collect high-frequency vibration spectra of the mechanical transmission chain in real time during engagement, disengagement, and operation. The range of the high-frequency vibration spectrum collection covers low-frequency mechanical motion characteristics to high-frequency gear meshing impact characteristics.

3. The new energy vehicle power take-off misoperation suppression system according to claim 2, characterized in that: The motor controller integrates a high-bandwidth current ripple analysis unit, which is used to monitor the average current intensity of the drive motor and extract the high-frequency ripple component in the three-phase current of the motor in real time. When a power take-off (PTO) gear engagement warning signal is detected, the high-bandwidth current ripple analysis unit analyzes the electromagnetic induction characteristics generated by the pre-drive circuit to identify and predict the electrical command trend that will generate abnormal gear engagement action in advance. The electrical feature extraction unit uses fast Fourier transform or wavelet transform techniques to separate harmonic components of order from the three-phase alternating current. These harmonic components are correlated with the electromagnetic disturbances of the inverter's gate drive circuit and power module, and are used to determine potential malfunction risks within milliseconds before physical action occurs.

4. The new energy vehicle power take-off misoperation suppression system according to claim 3, characterized in that: The physical information neural network model configured inside the vehicle controller is used to take the current ripple characteristics of the electrical domain as the first input feature vector and the vibration spectrum characteristics of the mechanical domain as the second input feature vector. The physical information neural network model has a deep correlation logic based on physical constraints built inside, which is used to analyze the spatiotemporal correlation between the first input feature vector and the second input feature vector in the time dimension and frequency dimension. When performing feature mapping, the physical information neural network model extracts the energy density distribution corresponding to the fork movement frequency, spline sleeve sliding frequency and gear pre-meshing characteristic frequency in the vibration spectrum, and combines it with the harmonic distortion rate corresponding to the electromagnetic commutation frequency and inverter switching frequency in the current ripple to calculate the cross-domain feature overlap. When the overlap of the cross-domain features is greater than the preset logical judgment probability, the vehicle controller confirms the risk of misoperation.

5. The new energy vehicle power take-off misoperation suppression system according to claim 4, characterized in that: The vehicle controller is used to implement a three-level progressive cooperative suppression strategy, which includes a predictive soft suppression stage, a cooperative locking stage, and an adaptive evolution stage. The vehicle controller supports the real-time operation of the physical information neural network model through a multi-core parallel processing architecture, and establishes a physical equation verification mechanism based on gear dynamics and motor electromagnetic constraints when outputting commands, in order to determine whether the current signal disturbance attribute belongs to normal load fluctuation or trend risk caused by illegal gear shifting commands.

6. The new energy vehicle power take-off misoperation suppression system according to claim 5, characterized in that: During the predictive soft suppression phase, when the physical information neural network model determines that the current operating condition is in a trend of misoperation, the vehicle controller sends a torque dynamic compensation command to the motor controller. According to the torque dynamic compensation command, the motor controller performs microsecond-level dynamic fine-tuning of the output torque of the drive motor within a predetermined time. The phase of the dynamic fine-tuning is inversely correlated with the mechanical vibration characteristics collected by the miniature broadband vibration sensor. This is used to actively counteract the mechanical impact caused by the tendency of incorrect gear engagement through energy regulation at the electrical end, thereby achieving soft contact protection of the mechanical interface. The torque fluctuation amplitude generated by the torque dynamic compensation command is limited to a proportion of the current rated torque of the drive motor to ensure that the fluctuation of the vehicle's driving state is within a preset smoothness threshold during the implementation of predictive soft suppression.

7. The new energy vehicle power take-off misoperation suppression system according to claim 6, characterized in that: During the cooperative locking phase, if, after implementing the predictive soft suppression, the vehicle controller detects that the amplitudes of the first input feature vector and the second input feature vector synchronously exceed a preset safety boundary threshold, then the vehicle controller synchronously sends a forced locking command to the power take-off controller and instructs the motor controller to reduce the motor output torque to zero. The forced locking command creates an irreversible physical inhibition state by cutting off the power supply to the solenoid valve of the power take-off actuator or locking the shift fork position. The motor controller is also equipped with high-voltage backflow protection logic. When the power take-off mechanical assembly experiences an unexpected engagement that causes a sudden change in the external load feedback energy, the motor controller, in coordination with the high-voltage power distribution unit, adjusts the bus current path within a predetermined time. By adjusting the vector control parameters of the inverter, the excess kinetic energy is converted into the internal energy of the motor or absorbed through the discharge circuit.

8. The new energy vehicle power take-off misoperation suppression system according to claim 7, characterized in that: In the adaptive evolution stage, the vehicle controller is equipped with an online learning unit. The online learning unit is used to take the successfully suppressed current ripple feature data pairs and vibration feature data pairs as positive guidance samples and send them back to the physical information neural network model for iterative update of parameter weights. Meanwhile, the online learning unit uses feature events that have not triggered suppression but are in a critical safety state as negative reference samples to dynamically optimize the feature mapping accuracy and the response threshold of the suppression strategy. The physical information neural network model is also equipped with an environmental adaptive compensation factor, which is used to adjust the gain of the second input feature vector according to the value of the power take-off lubricating oil temperature sensor, so as to correct the vibration spectrum drift caused by changes in lubricating oil temperature, fluctuations in ambient temperature, and mechanical wear.

9. The new energy vehicle power take-off misoperation suppression system according to claim 8, characterized in that: The vehicle controller is equipped with a global perception matrix, which integrates vehicle speed signal, handbrake status signal, brake pedal travel signal, gear information and high-voltage battery pack status parameters in real time, and uses them as boundary constraints of the physical information neural network model. When the global perception matrix detects that the vehicle is not stationary and the braking system is not activated, the vehicle controller automatically increases the sensitivity coefficient of the erroneous operation judgment. The power take-off (PTO) mechanical assembly is equipped with a shift fork position sensor. The PTO controller feeds back the actual position information collected by the shift fork position sensor to the vehicle controller to verify the execution effect of the three-level progressive cooperative inhibition strategy and to serve as the basis for closed-loop feedback control.

10. A method for suppressing malfunction of the power take-off unit in a new energy vehicle, characterized in that, The new energy vehicle power take-off malfunction suppression system according to any one of claims 1-9 is used to suppress the malfunction of the power take-off in new energy vehicles.

Images