A joint control method for power allocation of a motor train unit

Abstract

The present application relates to the technical field of joint control of different types or different functions of vehicle subsystems of a motor train unit, and discloses a joint control method for power allocation of a motor train unit, comprising: a central control unit acquiring intermediate DC bus voltage data and grid-side power supply voltage data, calculating a voltage change rate and filtering out noise through dynamic dead zone filtering logic, and adjusting a first determination boundary value according to grid-side voltage fluctuation; when the voltage change rate exceeds the first determination boundary value, a power resting period is started and a torque boost instruction is suspended; if the power of an auxiliary energy storage unit reaches a maximum power threshold value and the voltage change rate does not fall back, an instruction current upper limit is cut off in proportion; when the voltage change rate falls back to within a convergence safety boundary, the suspension is removed, the action timing is reconstructed to hedge communication time delay, the risk of overvoltage is eliminated, and the transient voltage stability is effectively improved.

Description

Technical Field

[0001] This invention relates to a joint control method for power distribution of high-speed trains, belonging to the field of joint control technology for vehicle subsystems of different types or functions in high-speed trains. Background Technology

[0002] During the operation of current high-speed trains, the traction drive subsystem and the auxiliary power supply system respectively undertake the functions of traction power output and onboard power supply. The central control unit typically coordinates the power distribution and operating status of these two systems. Under normal operating conditions, each subsystem completes inverter control and energy conversion according to control commands. The intermediate DC bus transmits electrical energy between the traction side and the auxiliary power supply side, and also plays a certain role in energy buffering to maintain the balance of power flow within the vehicle. When the high-speed train is under heavy load acceleration conditions, or when the traction power increases rapidly in a short period of time, the power demand of the traction drive subsystem will increase significantly. Meanwhile... The energy storage conversion action in the auxiliary power supply system will also affect the intermediate DC bus. Since the train communication network usually has a transmission delay of 10ms to 15ms, and the electrical circuit itself also has energy transfer lag, the traditional control method that relies on the independent local adjustment of each subsystem is difficult to coordinate the sequence of actions of the traction side and the auxiliary power supply side in a timely manner, which can easily cause significant voltage overshoot or voltage drop in the intermediate DC bus. To suppress such voltage fluctuations, a common solution is to adopt a fixed-duration delay waiting strategy, that is, after detecting a voltage abnormality, the torque boosting command is temporarily suspended, and control is restored after the timer expires.

[0003] However, this open-loop delay control method has significant shortcomings in engineering applications. If the waiting time is set too short, it is difficult to offset the lag caused by communication transmission and electrical response; if the waiting time is set too long, it will affect the vehicle's power response. More importantly, the fixed timing method cannot understand the real-time energy absorption of the auxiliary load during the waiting period. If torque boost is resumed directly after the timing ends, before the auxiliary energy storage has completed its absorption and adjustment, it may again impact the intermediate DC bus voltage, thereby triggering traction blocking or system degradation protection. However, the above suppression strategy is not only limited by the hardware physical form such as the topology of the on-board power supply equipment, but also has shortcomings at the software control level. For example, the authorization announcement number CN113 Chinese invention patent 147417B discloses a pantograph tripping handling method, system, and train power supply system. This solution implicitly relies on the single external fault premise of a strong power interruption on the grid source side at its underlying mechanism, which is a typical passive energy compensation mechanism. However, under the complex actual operating conditions of EMU trains, bus voltage fluctuations are more caused by the timing misalignment caused by the sudden power change across subsystems and the 10ms~15ms network transmission delay during heavy load acceleration. Due to the lack of timing reconstruction compensation for communication delay and closed-loop constraints on the power recovery trajectory at the load end, this solution is prone to control oscillation and secondary voltage overshoot at the moment of torque recovery, making it difficult to meet the high-precision closed-loop coordinated allocation requirements of transient power flow between multiple subsystems.

[0004] Therefore, the technical problem to be solved by this invention is how to overcome the transient imbalance of DC bus voltage caused by cross-subsystem communication delay and power change, so that the traction drive subsystem which undertakes the traction function and the auxiliary power supply system which undertakes the power supply and energy storage functions can form a closed loop coordination in terms of action timing and energy absorption capacity, thereby realizing precise control of the speed and torque of the traction motor of the electric drive propulsion vehicle and improving the operational stability of the traction motor control device and manufacturing parts of electric vehicles or rail vehicles. Summary of the Invention

[0005] To address the problems in the background art, the technical solution of the present invention is as follows: A joint control method for power allocation of high-speed trains, comprising the following steps:

[0006] Step S1: The central control unit acquires intermediate DC bus voltage data and grid-side power supply voltage data, which characterize the transient power state between the traction drive subsystem and the auxiliary power supply system.

[0007] Step S2: Calculate the voltage change rate based on the intermediate DC bus voltage data, and filter out the high-frequency voltage noise signal in the voltage change rate through dynamic dead-time filtering logic;

[0008] Step S3: Extract the low-frequency fluctuation characteristics of the grid-side power supply voltage data, and adjust the first judgment boundary value of the trigger power timing adjustment according to the step direction of the low-frequency fluctuation characteristics.

[0009] Step S4: When the voltage change rate after dynamic dead-zone filtering logic processing exceeds the first judgment boundary value after translation adjustment, the torque boost command is suspended, the power rest period is started, and the available power boundary of the auxiliary energy storage unit is read.

[0010] Step S5: If, during the power rest period, the current adjustable power of the auxiliary energy storage unit within the available power boundary reaches the maximum power threshold and the voltage change rate does not fall back to the convergence safety boundary, then the upper limit of the command current of the traction drive subsystem is cut off according to the ratio of the transient overvoltage difference value of the intermediate DC bus voltage data.

[0011] Step S6: When the voltage change rate falls back to within the convergence safety boundary, the suspension state of the torque boost command is released, and the rising slope of the command current of the traction drive subsystem is discretized and constrained according to the current voltage change rate.

[0012] Preferably, step S2, which filters out high-frequency voltage noise signals in the voltage change rate through dynamic dead-zone filtering logic, includes: step S21, acquiring intermediate DC bus voltage data within multiple consecutive control cycles; step S22, calculating the voltage difference sequence between adjacent control cycles; step S23, extracting the amplitude of high-frequency components from the voltage difference sequence and setting the amplitude of high-frequency components as the dynamic dead-zone threshold of the dynamic dead-zone filtering logic; step S24, setting the voltage change rate to 0 and outputting it when the absolute value of the voltage change rate in the current control cycle is less than the dynamic dead-zone threshold; and step S25, outputting the actual voltage change rate when the absolute value of the voltage change rate in the current control cycle is greater than or equal to the dynamic dead-zone threshold.

[0013] Preferably, the first determination boundary value for shifting and adjusting the trigger power timing adjustment in step S3 includes: step S31, low-pass filtering the grid-side power supply voltage data to obtain a grid voltage effective value reference sequence characterizing low-frequency fluctuations; step S32, comparing the grid voltage effective value reference sequence at the current moment with the grid voltage effective value reference sequence at the previous historical moment to generate a voltage step direction identifier; step S33, when the voltage step direction identifier is a positive jump and the jump amplitude exceeds a given grid voltage fluctuation threshold, shifting the first determination boundary value upward by a fixed step increment; step S34, when the voltage step direction identifier is a reverse jump and the jump amplitude exceeds a given grid voltage fluctuation threshold, shifting the first determination boundary value downward by a fixed step increment.

[0014] Preferably, the system further includes a degradation control step S7 for communication anomalies. Step S7 includes: Step S71, acquiring the total network communication delay data between the central control unit, the traction drive subsystem, and the auxiliary power supply system in real time; Step S72, determining that the system has entered a communication congestion state when the total network communication delay data exceeds the rigid communication threshold; Step S73, in the communication congestion state, the central control unit stops adjusting the translation of the first determination boundary value and cutting off the upper limit of the command current of the traction drive subsystem; Step S74, the traction drive subsystem and the auxiliary power supply system switch to independent limiting control state and independently limit the local power output command according to their respective fixed transformer drop power coefficients.

[0015] Preferably, step S4, reading the available power boundary of the auxiliary energy storage unit, includes: step S41, obtaining the current state of charge data and health characteristic variables of the auxiliary energy storage unit; step S42, calculating the maximum allowable charge / discharge rate of the auxiliary energy storage unit at the current ambient temperature based on the current state of charge data and health characteristic variables; and step S43, multiplying the maximum allowable charge / discharge rate by the rated capacity of the auxiliary energy storage unit to obtain the real-time available power boundary.

[0016] Preferably, step S5, which cuts off the upper limit of the command current of the traction drive subsystem according to the transient overvoltage difference value of the intermediate DC bus voltage data, includes: step S51, calculating the difference between the current intermediate DC bus voltage data and the standard rated bus voltage data to obtain the transient overvoltage difference value; step S52, inputting the transient overvoltage difference value into the limiting ratio mapping table to find the corresponding overvoltage cutoff coefficient; step S53, multiplying the overvoltage cutoff coefficient by the rated command current upper limit of the traction drive subsystem to generate the cutoff command current upper limit; and step S54, limiting the actual output command current of the traction drive subsystem within the cutoff command current upper limit.

[0017] Preferably, step S6 involves discretizing the rise slope of the command current in the traction drive subsystem based on the current voltage change rate, including: step S61, obtaining a pre-calibrated reference slope adjustment factor, and calculating the torque boost slope gating coefficient based on the reference slope adjustment factor, the current voltage change rate, and the first judgment boundary value after translation adjustment; step S62, using the torque boost slope gating coefficient to discretize the rise slope of the command current; wherein the torque boost slope gating coefficient is calculated according to the following formula: ,in, This is the torque boost slope gating coefficient; The baseline slope adjustment factor; This represents the current rate of voltage change. This is the first decision boundary value after translation adjustment.

[0018] Preferably, step S62 uses the torque boosting slope gating coefficient to discretize the command current rising slope, including: step S621, obtaining the nominal current rising slope of the traction drive subsystem; step S622, multiplying the nominal current rising slope by the torque boosting slope gating coefficient to generate the target rising slope; step S623, increasing the given current value of the traction drive subsystem according to the target rising slope until the given current value reaches the current required for the target torque.

[0019] Preferably, the criteria for determining that the voltage change rate falls back to the convergence safety boundary include: the absolute value of the voltage change rate is lower than the safety recovery threshold for three consecutive discrete control cycles; and the intermediate DC bus voltage data shows a monotonic trend of convergence to the rated voltage range for three consecutive discrete control cycles.

[0020] Preferably, the method further includes a pulse modulation drive step S8, which includes: step S81, where the central control unit converts the rising slope of the command current after discretization constraint into a drive level signal; step S82, where the drive level signal is transmitted back to the traction converter control board inside the traction drive subsystem through the train communication network; and step S83, where the traction converter control board adjusts the pulse width modulation duty cycle of the inverter switching transistor according to the drive level signal.

[0021] Compared with the prior art, the beneficial effects of the present invention are:

[0022] 1. In the joint control of EMU power allocation, to address the issue of voltage fluctuations in the intermediate DC bus caused by a short-term surge in EMU traction power and synchronous energy regulation by the auxiliary power supply side, this invention suspends torque boost commands when needed and sets a power rest period. This staggers the power increase on the traction side with the energy regulation on the auxiliary power supply side in time, thus mitigating the impact of communication delays and electrical response lags. This allows the intermediate DC bus voltage to return to a safer range of variation before traction is restored, thereby improving the coordination between the traction drive subsystem and the auxiliary power supply system, optimizing the speed and torque control response of the electric drive vehicle's traction motor, enhancing the coordinated speed regulation performance of the traction control device and manufacturing components, and strengthening the transient voltage stability of the entire vehicle.

[0023] 2. This invention, by combining the drop in intermediate DC bus voltage during the power rest period, constrains the rise rate of the command current when the torque is restored, so that the traction power restoration process no longer relies solely on a fixed timer. As a result, the speed at which the traction power is released again can better adapt to the energy absorption capacity of the auxiliary energy storage unit at that time, reducing the risk of the intermediate DC bus voltage being raised again during torque restoration, avoiding traction blockage or degraded protection caused by secondary overshoot, and helping to maintain the coordinated and stable traction and auxiliary power supply of the EMU.

[0024] 3. This invention processes high-frequency voltage noise signals in the voltage change rate through dynamic dead-zone filtering logic, which can reduce the interference of small noise on control judgment and avoid frequent false triggering of field control. At the same time, it adjusts the first judgment boundary value by combining the low-frequency fluctuation of the grid-side power supply voltage, so that the trigger judgment can be adjusted accordingly with the change of power supply environment. In this way, under the condition that grid voltage fluctuation and high-frequency interference coexist, it can still accurately distinguish between real power impact and noise disturbance, reduce the possibility of control oscillation, and improve the stability of cross-subsystem power coordination control. Attached Figure Description

[0025] Figure 1 This is a flowchart of the power timing adjustment steps of the central control unit of the present invention;

[0026] Figure 2 This is the transient voltage stability control logic diagram of the present invention.

[0027] The objectives, features, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0028] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0029] A joint control method for power distribution of high-speed trains includes the following steps:

[0030] Step S1: The central control unit acquires intermediate DC bus voltage data and grid-side power supply voltage data, which characterize the transient power state between the traction drive subsystem and the auxiliary power supply system.

[0031] Step S2: Calculate the voltage change rate based on the intermediate DC bus voltage data, and filter out the high-frequency voltage noise signal in the voltage change rate through dynamic dead-time filtering logic;

[0032] Step S3: Extract the low-frequency fluctuation characteristics of the grid-side power supply voltage data, and adjust the first judgment boundary value of the trigger power timing adjustment according to the step direction of the low-frequency fluctuation characteristics.

[0033] Step S4: When the voltage change rate after dynamic dead-zone filtering logic processing exceeds the first judgment boundary value after translation adjustment, the torque boost command is suspended, the power rest period is started, and the available power boundary of the auxiliary energy storage unit is read.

[0034] Step S5: If, during the power rest period, the current adjustable power of the auxiliary energy storage unit within the available power boundary reaches the maximum power threshold and the voltage change rate does not fall back to the convergence safety boundary, then the upper limit of the command current of the traction drive subsystem is cut off according to the ratio of the transient overvoltage difference value of the intermediate DC bus voltage data.

[0035] Step S6: When the voltage change rate falls back to within the convergence safety boundary, the suspension state of the torque boost command is released, and the rising slope of the command current of the traction drive subsystem is discretized and constrained according to the current voltage change rate.

[0036] Preferably, step S2, which filters out high-frequency voltage noise signals in the voltage change rate through dynamic dead-zone filtering logic, includes: step S21, acquiring intermediate DC bus voltage data within multiple consecutive control cycles; step S22, calculating the voltage difference sequence between adjacent control cycles; step S23, extracting the amplitude of high-frequency components from the voltage difference sequence and setting the amplitude of high-frequency components as the dynamic dead-zone threshold of the dynamic dead-zone filtering logic; step S24, setting the voltage change rate to 0 and outputting it when the absolute value of the voltage change rate in the current control cycle is less than the dynamic dead-zone threshold; and step S25, outputting the actual voltage change rate when the absolute value of the voltage change rate in the current control cycle is greater than or equal to the dynamic dead-zone threshold.

[0037] Preferably, the first determination boundary value for shifting and adjusting the trigger power timing adjustment in step S3 includes: step S31, low-pass filtering the grid-side power supply voltage data to obtain a grid voltage effective value reference sequence characterizing low-frequency fluctuations; step S32, comparing the grid voltage effective value reference sequence at the current moment with the grid voltage effective value reference sequence at the previous historical moment to generate a voltage step direction identifier; step S33, when the voltage step direction identifier is a positive jump and the jump amplitude exceeds a given grid voltage fluctuation threshold, shifting the first determination boundary value upward by a fixed step increment; step S34, when the voltage step direction identifier is a reverse jump and the jump amplitude exceeds a given grid voltage fluctuation threshold, shifting the first determination boundary value downward by a fixed step increment.

[0038] Preferably, the system further includes a degradation control step S7 for communication anomalies. Step S7 includes: Step S71, acquiring the total network communication delay data between the central control unit, the traction drive subsystem, and the auxiliary power supply system in real time; Step S72, determining that the system has entered a communication congestion state when the total network communication delay data exceeds the rigid communication threshold; Step S73, in the communication congestion state, the central control unit stops adjusting the translation of the first determination boundary value and cutting off the upper limit of the command current of the traction drive subsystem; Step S74, the traction drive subsystem and the auxiliary power supply system switch to independent limiting control state and independently limit the local power output command according to their respective fixed transformer drop power coefficients.

[0039] Preferably, step S4, reading the available power boundary of the auxiliary energy storage unit, includes: step S41, obtaining the current state of charge data and health characteristic variables of the auxiliary energy storage unit; step S42, calculating the maximum allowable charge / discharge rate of the auxiliary energy storage unit at the current ambient temperature based on the current state of charge data and health characteristic variables; and step S43, multiplying the maximum allowable charge / discharge rate by the rated capacity of the auxiliary energy storage unit to obtain the real-time available power boundary.

[0040] Preferably, step S5, which cuts off the upper limit of the command current of the traction drive subsystem according to the transient overvoltage difference value of the intermediate DC bus voltage data, includes: step S51, calculating the difference between the current intermediate DC bus voltage data and the standard rated bus voltage data to obtain the transient overvoltage difference value; step S52, inputting the transient overvoltage difference value into the limiting ratio mapping table to find the corresponding overvoltage cutoff coefficient; step S53, multiplying the overvoltage cutoff coefficient by the rated command current upper limit of the traction drive subsystem to generate the cutoff command current upper limit; and step S54, limiting the actual output command current of the traction drive subsystem within the cutoff command current upper limit.

[0041] Preferably, step S6 involves discretizing the rise slope of the command current in the traction drive subsystem based on the current voltage change rate, including: step S61, obtaining a pre-calibrated reference slope adjustment factor, and calculating the torque boost slope gating coefficient based on the reference slope adjustment factor, the current voltage change rate, and the first judgment boundary value after translation adjustment; step S62, using the torque boost slope gating coefficient to discretize the rise slope of the command current; wherein the torque boost slope gating coefficient is calculated according to the following formula: ,in, This is the torque boost slope gating coefficient; The baseline slope adjustment factor; This represents the current rate of voltage change. This is the first decision boundary value after translation adjustment.

[0042] Preferably, step S62 uses the torque boosting slope gating coefficient to discretize the command current rising slope, including: step S621, obtaining the nominal current rising slope of the traction drive subsystem; step S622, multiplying the nominal current rising slope by the torque boosting slope gating coefficient to generate the target rising slope; step S623, increasing the given current value of the traction drive subsystem according to the target rising slope until the given current value reaches the current required for the target torque.

[0043] Preferably, the criteria for determining that the voltage change rate falls back to the convergence safety boundary include: the absolute value of the voltage change rate is lower than the safety recovery threshold for three consecutive discrete control cycles; and the intermediate DC bus voltage data shows a monotonic trend of convergence to the rated voltage range for three consecutive discrete control cycles.

[0044] Preferably, the method further includes a pulse modulation drive step S8, which includes: step S81, where the central control unit converts the rising slope of the command current after discretization constraint into a drive level signal; step S82, where the drive level signal is transmitted back to the traction converter control board inside the traction drive subsystem through the train communication network; and step S83, where the traction converter control board adjusts the pulse width modulation duty cycle of the inverter switching transistor according to the drive level signal.

[0045] Example 1: When the EMU is operating under multi-tasking conditions, the central control unit periodically acquires intermediate DC bus voltage data and grid-side power supply voltage data via the train bus. At this time, the traction drive subsystem and auxiliary power supply system operate in parallel. After a sudden increase in traction power, a short-term imbalance occurs in the power exchange on both sides of the intermediate DC bus. There is a total network communication delay of 10ms to 15ms between the central control unit and the local actuators of the two subsystems. To verify the effectiveness of this joint control method in handling timing conflicts across subsystems, three sets of joint control tracks with different first judgment boundary values ​​and maximum power thresholds are set up in a standard test bench environment. In the first set, the first judgment boundary value is... Set to 30V / s and set the maximum power threshold of the auxiliary energy storage unit. Set to 50kW; in the second sample group, Set to 50V / s, and Set to 100kW; in the third sample group, Set to 80V / s, and It is set to 150kW.

[0046] The central control unit collects intermediate DC bus voltage data over multiple consecutive control cycles. By performing discrete difference analysis, the voltage change rate for the current control cycle can be obtained. The central control unit calculates the voltage difference sequence between adjacent control cycles. For high-frequency voltage noise signals introduced by sensor and converter switching actions, the central control unit extracts the amplitude of the high-frequency component from this voltage difference sequence and sets this high-frequency component amplitude as the dynamic dead-time threshold of the dynamic dead-time filtering logic. The voltage change rate in the current control cycle... When the absolute value is less than the dynamic dead zone threshold, the voltage change rate is set to 0 and then output. When its absolute value is greater than or equal to the dynamic dead zone threshold, the actual voltage change rate is output. In this test environment, the amplitude of the high-frequency component corresponds to 5% of the effective amplitude of the voltage change signal. The dynamic dead zone filtering logic filters out small-amplitude high-frequency voltage noise signals to avoid noise directly entering subsequent trigger judgments.

[0047] The central control unit simultaneously performs low-pass filtering on the grid-side power supply voltage data to obtain a reference sequence of grid voltage RMS values ​​characterizing low-frequency fluctuations. It then compares the current grid voltage RMS reference sequence with the previous historical reference sequence to generate a voltage step direction indicator. When the voltage step direction indicator shows a positive jump and the jump amplitude exceeds a given grid voltage fluctuation threshold, the first judgment boundary value is set. Shift upwards by a fixed step increment; when the voltage step direction is identified as a reverse jump and the jump amplitude exceeds the given grid voltage fluctuation threshold, Shift downwards by the same fixed step increment, and adjust accordingly. The voltage change rate after dynamic dead-time filtering logic processing serves as the boundary for determining the trigger power timing adjustment within this control cycle. Beyond translation adjustment At that time, the central control unit determines that a transient power conflict has occurred between the traction drive subsystem and the auxiliary power supply system. It then suspends the torque boost command of the traction drive subsystem, starts the power rest period, and reads the available power boundary of the auxiliary energy storage unit.

[0048] During the power rest period, the central control unit sends dispatch commands to the auxiliary power supply system to change the charging and discharging state of the auxiliary energy storage unit and continuously monitors the intermediate DC bus voltage data and its rate of change. To read the available power boundary of the auxiliary energy storage unit, a three-dimensional mesh mapping lookup table is pre-installed in the central control unit's memory. The inputs to this lookup table include the current state of charge data of the auxiliary energy storage unit, health characteristic variables, and the Celsius value collected by the current ambient temperature sensor. When the ambient temperature is in the normal temperature range of 25°C, the current state of charge data is in the energy consumption range of 40% to 60%, and the health characteristic variable is greater than 90%, the maximum allowable charge and discharge ratio output by the lookup table is 5 times. When the ambient temperature drops to 0°C, or the state of charge data exceeds the aforementioned range, the nonlinear weight attenuation coefficient inside the lookup table monotonically reduces the maximum allowable charge and discharge ratio to 2 times. After reading this ratio output, the central control unit multiplies it by the rated capacity of the auxiliary energy storage unit to obtain the real-time available power boundary.

[0049] Before the power rest period ends, the auxiliary energy storage unit's current adjustable power within the available power boundary reaches the maximum power threshold. And the rate of change of voltage If the voltage has not yet fallen back to the convergence safety boundary, the central control unit calculates the difference between the current intermediate DC bus voltage data and the standard rated bus voltage data to obtain the transient overvoltage difference value. This transient overvoltage difference value is then input into the limiting ratio mapping table to find the corresponding overvoltage cutoff coefficient. Subsequently, the central control unit multiplies the overvoltage cutoff coefficient by the rated command current upper limit of the traction drive subsystem to generate the cutoff command current upper limit, and limits the actual output command current of the traction drive subsystem within this cutoff command current upper limit.

[0050] When the rate of change of voltage When the voltage drops back to within the convergence safety boundary, the central control unit releases the suspension of the torque boost command and, in the initial stage of resuming torque boost, adjusts the current voltage change rate accordingly. Discretized constraints are applied to the rise rate of the command current in the traction drive subsystem. Specifically, the central control unit obtains a pre-calibrated reference slope adjustment factor. and based on Current voltage change rate and the first judgment boundary value after translation adjustment Calculate the torque boost slope gating coefficient The torque boost slope gating coefficient Satisfy the following formula: To prevent control logic distortion caused by negative or zero values ​​in the calculation results due to the absolute value of the intermediate DC bus voltage change rate exceeding the first judgment boundary value after translation adjustment during actual operation, the control system is equipped with a mandatory hard clamping physical boundary protection mechanism. When the absolute value of the voltage change rate after dynamic dead-zone filtering is greater than or equal to the first judgment boundary value after translation adjustment, the microprocessor will automatically fuse and interrupt the calculation, and forcibly assign and lock the torque boost slope gating coefficient to the calibrated positive low-limit threshold clamping value of 0.15. This ensures that the target rise slope always remains positive, eliminating the secondary fault defect of reverse inversion of the command current. This is the torque boost slope gate coefficient. As the baseline slope adjustment factor, This refers to the rate of change of the intermediate DC bus voltage, which is obtained in real time by the central control unit before the end of the current control cycle. The first judgment boundary value after translation adjustment, subscript Indicates DC, subscript Indicates the maximum value, subscript Indicating the reference value, the central control unit utilizes torque to boost the slope gating coefficient. Discretize the rise rate of the command current to enable the traction drive subsystem to increase the output command current according to the constrained slope when restoring torque boost.

[0051] In the first group, the intermediate DC bus voltage experiences a transient change from the rated 600V, and the filtered voltage change rate is... Upon reaching 32V / s, the central control unit initiates a power rest period with a total network communication delay of 12.5ms. The auxiliary energy storage unit adjusts its power absorption status within the maximum power threshold of 50kW. The voltage change rate falls back to within the convergence safety boundary within 18.2ms. Before the termination of this control cycle, The voltage drops to 2.4V / s, and the result is calculated using the formula above. The value is 0.92. Based on this, the traction drive subsystem reduces the rise rate of the command current during the recovery phase, thereby suppressing the secondary voltage surge during torque recovery.

[0052] Regarding the internal data structure of the limiting ratio mapping table, the system performs segmented, monotonically decreasing proportional truncation based on the magnitude of the transient overvoltage difference. This limiting ratio mapping table is represented as a continuously allocated discrete one-dimensional array structure in the static random access memory of the central control unit. Its starting address is stored in the configuration pointer register of the control chip. By using the real-time transient overvoltage difference digital signal as the address offset, it is directly mapped at the hardware level to the corresponding overvoltage cutoff coefficient register storage area to quantitatively limit the inverter's current upper limit. Specifically, it executes the following segmented... The control logic works as follows: when the transient overvoltage difference is less than or equal to 10V, the overvoltage cutoff coefficient output by the mapping table is 1, indicating that the traction drive subsystem maintains the upper limit of the rated command current; when the transient overvoltage difference is greater than 10V and less than or equal to 50V, the mapping table internally stores a linearly decreasing array, causing the overvoltage cutoff coefficient to decrease uniformly from 1 to 0.5 as the voltage difference increases; when the transient overvoltage difference exceeds 50V, the mapping table outputs a lower limit clamping value of 0.2 for the overvoltage cutoff coefficient, and the central control unit performs a multiplication operation according to this segmented correspondence to generate the cutoff upper limit of the command current.

[0053] In the second sample group, the rate of change of filtered voltage Reaching 55V / s, the total network communication latency increased to 14.1ms. The central control unit activated the power rest period, and the auxiliary energy storage unit responded within the maximum power threshold of 100kW, bringing the energy conversion adjustment time within 16.5ms. Compared to the peak bus voltage of 820V under conventional open-loop control, the peak bus voltage in this embodiment decreased to 695V, a reduction of over 35%. In the third sample group, the filtered voltage change rate... When the voltage reaches 85V / s, the auxiliary energy storage unit reaches its maximum power threshold of 150kW. Since this voltage change rate has exceeded the first judgment boundary value of 80V / s after the shift adjustment, the central control unit stops using the conventional linear proportional calculation result and activates the low-limit threshold clamping value, adjusting the torque boost slope gating coefficient. The value is set to 0.15, and the rise rate of the command current of the traction drive subsystem is limited accordingly. In contrast, in the missing control group with the power rest period and torque rise rate gating coefficient removed, the traction torque rise command is directly restored after the fixed time wait expires. The auxiliary energy storage unit generates an 865V secondary voltage overshoot before it has completed its response, and triggers the protection cut-off of the local converter.

[0054] During the degradation control process due to communication anomalies, the central control unit acquires the total network communication latency data between itself and the traction drive subsystem and auxiliary power supply system in real time. When the total network communication latency data exceeds the rigid communication threshold, the system enters a communication congestion state, and the central control unit stops processing the first judgment boundary value. The translation adjustment stops the cutoff of the upper limit of the command current of the traction drive subsystem; the traction drive subsystem and the auxiliary power supply system switch to independent amplitude limiting control state, and independently limit the local power output command according to their respective fixed voltage drop power coefficients. The local fixed voltage drop power coefficient of the traction drive subsystem is set to 0.05 per kilowatt, which means that its local power output command is actively reduced by 5% for every 1kW change in power caused by the increase of the bus transient voltage; the local fixed voltage drop power coefficient of the auxiliary power supply system is set to 0.02 per kilowatt. Through the above processing, the EMU forms a closed-loop coordination process between the traction drive subsystem and the auxiliary power supply system with intermediate DC bus voltage data, grid-side power supply voltage data, available power boundary of auxiliary energy storage unit and command current upper limit as the core. Without changing the electrical main circuit and mechanical structure, the transient fluctuation amplitude of DC bus voltage and the joint control response delay are kept within the preset safety range.

[0055] Example 2: In a hardware-in-the-loop simulation test bench simulating the joint operation of the traction transmission subsystem and auxiliary power supply system of a high-speed train, the grid-side voltage monitoring unit collects grid-side power supply voltage data at a sampling rate of 50kHz, and the DC bus voltage sampling module collects intermediate DC bus voltage data at a resolution of 0.1%. To simulate the strong electromagnetic interference environment generated by the high-voltage transient discharge on the roof and the switching of the converter, the signal input terminal of the network converter controller receives raw data superimposed with Gaussian white noise with a signal-to-noise ratio of 20dB and power frequency interference harmonics at a frequency of 50Hz. The data storage components and logic control chip inside the simulation test bench are used to realize information conversion and interaction; and to determine the state adjustment cycle. At that time, the main constraints include the bandwidth of the transient voltage and the timing interrupt constant of the main control unit. When the bandwidth of the cross-subsystem coupling energy impulse waveform excited by load change widens, in order to prevent aliasing under the condition of satisfying the signal sampling theorem, the state adjustment period is adjusted. Converging towards its lower window limit, under typical operating conditions, the state adjustment period... It is set to 0.2ms and used as the discrete control time step for subsequent differential calculations and sequential logic state transitions.

[0056] This embodiment sets up multiple differentiated configurations to verify the response of the joint control method in suppressing the secondary overshoot of the intermediate DC bus voltage and coordinating the energy flow across subsystems. The test group fully deploys dynamic dead-zone filtering logic and shifts and adjusts the first decision boundary value. Timing suspension mechanism and torque boost slope gating coefficient The control loop is partially missing. Control group one removes the dynamic dead-zone filtering logic and directly inputs the original voltage change rate containing Gaussian white noise into the control loop. Control group two retains the suspension action of the torque boost command but removes the slope gating constraint in the recovery phase, adopting a fixed-time recovery method. Control group one beyond the range sets the first judgment boundary value... The value is set to 10V / s, which is below the reasonable range; the first judgment boundary value is set for the out-of-range control group two. Set to 120V / s, exceeding the converter's withstand limit, each sample group operated under the same network transmission delay and operating condition intensity gradient, and the first judgment boundary value was recorded sequentially. The maximum original voltage change rate, the maximum voltage change rate after dynamic dead-time filtering logic processing, the duration of the power rest period, and the torque boost slope gating coefficient. And the highest peak value of the bus voltage.

[0057] In the three test groups that fully deploy this solution, the first test group will determine the first boundary value. The initial voltage change rate was set to 30V / s, with a maximum value of 34.2V / s. After the dynamic dead-time filtering logic removed high-frequency voltage noise signals, the maximum voltage change rate at which the trigger judgment was entered dropped to 31.5V / s. This value still exceeded the first judgment boundary value of 30V / s. The central control unit then suspends the torque boost command and initiates a power rest period, which lasts for 12.4 ms. The torque boost slope gating coefficient calculated during the recovery phase... With a value of 0.915, the rising slope of the command current of the traction drive subsystem is constrained by discretization. The highest peak value of the bus voltage is 692.3V. The second sample group will use the first judgment boundary value. The voltage change rate is set to 50V / s. The maximum original voltage change rate is 58.4V / s, and the maximum filtered voltage change rate is 54.1V / s. Based on this, the central control unit initiates a 14.2ms power rest period and obtains a torque boost slope gating coefficient of 0.724 during the recovery phase. The highest peak value of the bus voltage was 695.1V. The third sample group will be the first judgment boundary value. The voltage change rate is set to 80V / s, with a maximum original voltage change rate of 89.1V / s and a maximum filtered voltage change rate of 84.6V / s. After a 15.1ms power rest period, the torque boost slope gating coefficient during the recovery phase... The value is 0.152, and the highest peak value of the bus voltage is 698.4V. Under the trigger boundaries of 30V / s, 50V / s, and 80V / s, the dynamic dead-time filtering logic filters out the high-frequency components in the original voltage change rate, and then uses the first judgment boundary value after shift adjustment. During the trigger power rest period, the final slope gating coefficient is increased by a torque of less than 1. The ramp rate of the command current during the recovery phase is limited to keep the peak bus voltage stable between 692.3V and 698.4V.

[0058] The partial deletion control group used a first decision boundary value of 50V / s. However, the dynamic dead-time filtering logic was not enabled. The maximum original voltage change rate of this sample group was 58.6V / s. Because no filtering was performed, the maximum voltage change rate of the input control loop was still 58.6V / s. The power rest period of the central control unit was 11.1ms, and the torque boost slope gating coefficient during the recovery phase was... The value was 0.612, and the peak bus voltage rose to 742.6V. Compared with the second sample group, where the bus voltage peak value was 695.1V after filtering from 58.4V / s to 54.1V / s, this control group directly used the original voltage change rate containing noise for control, causing the triggering and recovery process to be affected by noise, and the peak bus voltage peak value increased accordingly. The second partially missing control group also used a first judgment boundary value of 50V / s. The maximum original voltage change rate was 58.3 V / s, and the maximum filtered voltage change rate was 53.9 V / s. Although this sample retained the suspension action of the torque boost command and formed a 14.5 ms power rest period, it did not utilize the torque boost slope gating coefficient to discretize the command current rise slope during the recovery phase. The record was 1.000, and the highest peak bus voltage reached 835.4V. This data is consistent with that in the second sample group. The result of 0.724 and the highest peak bus voltage of 695.1V correspond to the fact that if the rise rate of the command current of the traction drive subsystem is not limited during the recovery phase, the torque recovery after the power rest period will still cause significant secondary overvoltage.

[0059] When the first determination boundary value When the value is below the preset range, the control system is in the corresponding power allocation state, and the sixth sample group will determine the first judgment boundary value. The initial voltage change rate was set to 10V / s, with a maximum original voltage change rate of 15.2V / s and a maximum filtered voltage change rate of 12.1V / s. Since the filtered voltage change rate was still higher than the 10V / s threshold, the system triggered a power rest period, the duration of which was extended to 28.4ms. The torque boost slope gating coefficient during the recovery phase was also adjusted. The voltage dropped to 0.051, with the highest peak bus voltage reaching 712.5V. The seventh sample group also set the first judgment boundary value. With the setting at 10V / s, the maximum original voltage change rate increased to 32.4V / s, and the maximum voltage change rate after filtering was 29.3V / s. The system further extended the power rest period to 35.2ms and adjusted the torque boost slope gating coefficient. When the voltage is reduced to 0.012, the highest peak bus voltage is 725.3V. Of the two sample groups mentioned above, the lower first judgment boundary value... This causes the system to trigger even slight power fluctuations, significantly lengthening the power rest period. The rise rate of the command current during the recovery phase of the traction drive subsystem is excessively restricted, thereby increasing the command transmission delay and the computational load on the main control unit.

[0060] The out-of-range control group two was used to observe the first judgment boundary value. When the control state is too high, the eighth sample group will determine the first decision boundary value. The voltage change rate was set to 120V / s. The maximum original voltage change rate was 58.5V / s, and the maximum filtered voltage change rate was 54.2V / s. Because the filtered voltage change rate did not reach 120V / s, the system did not activate the power rest period, which was recorded as 0.0ms. The torque boost slope gating coefficient during the recovery phase was... The value was 1.000, and the peak bus voltage rose to 824.2V. The ninth sample group also used a first judgment boundary value of 120V / s. The maximum original voltage change rate was 88.9 V / s, and the maximum filtered voltage change rate was 84.3 V / s. This value is still below 120 V / s, and the system did not trigger a power rest period, which lasted 0.0 ms. The torque boost slope gating coefficient... With a value of 1.000, the highest peak value of the bus voltage further increased to 863.7V. Compared with the highest peak values ​​of 692.3V, 695.1V and 698.4V of the bus voltage in the three test groups, the trigger boundary of 120V / s caused the control circuit to fail to suspend the torque increase command under the power coupling impact of 58.5V / s and 88.9V / s, and the intermediate DC bus voltage subsequently entered the overvoltage blockade risk zone.

[0061] As can be seen from the discrete channel test data above, the dynamic dead-time filtering logic first filters the original voltage change rate, reducing the original voltage change rates of 34.2V / s, 58.4V / s, and 89.1V / s to 31.5V / s, 54.1V / s, and 84.6V / s, respectively; subsequently, the central control unit uses the first judgment boundary value after translation adjustment. Determine whether to activate the power rest period, and use the torque boost slope gating coefficient during the recovery phase. By limiting the rise rate of the command current, the test group achieved power rest periods of 12.4 ms, 14.2 ms, and 15.1 ms under different impact intensities. Corresponding to 0.915, 0.724, and 0.152, the highest peak bus voltage remained below 698.4V; in the partially missing control group one, after removing the dynamic dead-time filtering logic, the highest peak bus voltage increased to 742.6V; in the partially missing control group two, after removing the slope gating constraint during the recovery phase, the highest peak bus voltage increased to 835.4V; when the first judgment boundary value... When the voltage drops to 10V / s, the system experiences over-triggered operation, with the power rest period extending to 28.4ms and 35.2ms; when the first judgment boundary value... When the voltage rose to 120V / s, the system failed to activate the power rest period, and the peak bus voltage reached 824.2V and 863.7V. These data support the first judgment boundary value. Dynamic dead-time filtering logic, power rest period, and torque boost slope gating coefficient. The coordination between them enables the traction drive subsystem and the auxiliary power supply system to complete transient power allocation without changing the hardware structure.

[0062] Example 3: The microprocessor in the current central control unit reads the discrete voltage and current signals stored in the digital input buffer register through the peripheral component interconnect bus. The initial bit depth of the digital input buffer register is set to 16 bits, corresponding to an analog-to-digital conversion resolution of 12 bits in the data acquisition environment, and the minimum sampling frequency is maintained at 50kHz. The system bus of the main controller has a state adjustment cycle of 0.2ms. Refreshing the input memory address: Under this condition, transient coupling disturbances between vehicle subsystems and low-frequency fundamental distortion constitute the original input source. The microprocessor executes dynamic dead-time filtering logic and the first decision boundary value based on this. Translation adjustment.

[0063] In the dynamic dead-time filtering logic, the microprocessor reads the intermediate DC bus voltage data of the current sampling period and obtains the voltage change rate of the current control period based on the voltage difference sequence between adjacent control periods. To extract the high-frequency component amplitude, the microprocessor performs a Fast Fourier Transform (FFT) on the voltage difference sequence within a continuous control cycle, converting the time-domain difference sequence to the frequency domain. Subsequently, a 500Hz high-pass cutoff filter is introduced into the frequency domain to remove low-frequency fundamental components below this frequency, separating the high-frequency voltage noise signal. The digital signal processor then performs an Inverse Fast Fourier Transform (IFFT) on the processed frequency-domain signal to reconstruct the high-frequency time-domain pulsating waveform and calculates the root mean square (RMS) value of this waveform within a sliding window of five consecutive control cycles. This RMS value is set as the dynamic dead zone threshold, and the microprocessor sets the voltage change rate for the current control cycle. The absolute value is compared with the dynamic dead zone threshold. When the absolute value is less than the dynamic dead zone threshold, the data bus sends a zero clear command to the digital signal processor to set the voltage change rate to 0 and output it. When the absolute value is greater than or equal to the dynamic dead zone threshold, the actual voltage change rate is output. This process prevents high-frequency disturbances with amplitudes lower than 5% of the effective amplitude of the voltage change signal from entering the subsequent trigger judgment.

[0064] To extract the low-frequency fluctuation characteristics of the grid-side power supply voltage data, the microprocessor calls a discrete low-pass digital filter with a cutoff frequency set to 5Hz. It then performs sliding-window weighted information superposition on the grid-side power supply voltage data from 50 consecutive sampling periods using a shift register to filter out high-frequency chopping ripple components, obtaining a baseline sequence of grid voltage RMS values. During sliding-window weighting, the weighting value of a single sampling point for the grid-side power supply voltage data from the 1st to 10th sampling periods closest to the current time is uniformly configured as 0.04; the weighting value of a single sampling point for the data from the 11th to 30th sampling periods closest to the current time is uniformly configured as 0.02; and the weighting value of a single sampling point for the data from the 31st to 50th sampling periods furthest from the current time is uniformly configured as 0.01. The sum of these weights is 1. Based on this, the microprocessor obtains the baseline sequence of grid voltage RMS values ​​for the current time and compares it with the baseline sequence of grid voltage RMS values ​​from the previous historical time, generating a voltage step direction indicator and the corresponding jump amplitude value.

[0065] The microprocessor invokes the boundary shift adjustment loop based on the voltage step direction indicator. When the voltage step direction indicator is a positive jump and the jump amplitude exceeds the given grid voltage fluctuation threshold, the boundary shift adjustment loop will first determine the boundary value. Shift upwards by a fixed step increment; when the voltage step direction is identified as a reverse jump and the jump amplitude exceeds a given grid voltage fluctuation threshold, the first judgment boundary value is set. The voltage is shifted downwards by the same fixed step increment. This fixed step increment is pre-written into the boundary configuration register of the control state machine and converted into step instructions for the boundary register by the discrete step quantization mapping interface. When a voltage drop occurs on the power grid side, the current effective voltage value reference sequence forms a negative 20V step relative to the previous historical time. When the amplitude of this jump exceeds the given voltage fluctuation threshold, the microprocessor sets the voltage step direction indicator to a reverse jump, triggering the first judgment boundary value. The first judgment boundary value is shifted downward by a fixed step increment. An initial fixed voltage change rate reference value is preset in the microprocessor's non-volatile memory. The physical dimension of this reference value is volts per second, and its initial reference value is explicitly configured as 50V / s. This serves as the absolute discrete comparison source for triggering power timing adjustments. During operation, the boundary shift adjustment loop performs discrete stepwise superposition on the base value based on the low-frequency fluctuation characteristics of the grid-side voltage. Under this test condition, the reference boundary is 50V / s. After downward shifting, the first judgment boundary value used in the current control cycle is... It is 40V / s.

[0066] First decision boundary value After completing the translation adjustment, the microprocessor will process the voltage change rate after dynamic dead-time filtering logic. Compared to this boundary, the rate of voltage change exceeds the first judgment boundary value after translation adjustment. At this time, the central control unit suspends the torque boost command of the traction drive subsystem, initiates a power rest period, and continuously polls the intermediate DC bus voltage data and its rate of change during this period. To perform a status determination when resuming torque boost control, the microprocessor sets a check control interval containing 10 consecutive sampling cycles; within this window, if the rate of change at each discrete sampling point... The absolute values ​​are all less than the convergence safety boundary value, and the intermediate DC bus voltage data shows a monotonic trend of convergence towards the rated voltage range. The state machine counter accumulates to the full value and outputs a logic high level, determining that the voltage change rate has fallen back to within the convergence safety boundary. In the timed interrupt refresh at the bottom layer of this control system, a single complete discrete control cycle is composed of three consecutive signal sampling points. Therefore, when all 10 consecutive sampling points meet the above conditions, the coverage time length has exceeded three consecutive complete discrete control cycles. The variance statistical envelope under stable operating conditions determines the convergence safety boundary value as 5V / s, which is also used as the safety recovery threshold.

[0067] When the state machine confirms the rate of voltage change After falling back to within the convergence safety boundary, the central control unit terminates the power rest period, writes the torque recovery enable word to the register port of the traction converter, and releases the pending state of the torque boost command, resuming the initial stage. The central control unit then adjusts the power output based on the current voltage change rate. The first judgment boundary value after translation adjustment and the calibrated baseline slope adjustment factor Calculate the torque boost slope gating coefficient The coefficient is used to discretize the rise slope of the command current of the traction drive subsystem; the traction converter then restores the pulse width modulation pulse output according to the restricted rise slope of the command current, so that the vehicle joint control system returns to the steady-state operation of normal traction.

[0068] Example 4: This example combines Figures 1 to 2 This section describes a joint control method for power allocation in high-speed trains, such as... Figure 1 As shown, in step S1, the central control unit acquires intermediate DC bus voltage data and grid-side power supply voltage data, which characterize the transient power state between the traction drive subsystem and the auxiliary power supply system. In step S2, the intermediate DC bus voltage data is used to calculate the voltage change rate, and high-frequency voltage noise signals in the voltage change rate are filtered out by dynamic dead-zone filtering logic. In step S3, the low-frequency fluctuation characteristics of the grid-side power supply voltage data are extracted, and the first judgment boundary value of the trigger power timing adjustment is shifted according to the step direction of the low-frequency fluctuation characteristics. In step S4, when the voltage change rate after dynamic dead-zone filtering logic processing exceeds the first judgment boundary value after shifting adjustment... When the value is reached, the torque boost command is suspended, the power rest period is started, and the available power boundary of the auxiliary energy storage unit is read. In step S5, if the current adjustable power of the auxiliary energy storage unit within the available power boundary reaches the maximum power threshold during the power rest period, and the voltage change rate does not fall back to the convergence safety boundary, the upper limit of the command current of the traction drive subsystem is cut off according to the transient overvoltage difference value ratio of the intermediate DC bus voltage data. In step S6, when the voltage change rate falls back to the convergence safety boundary, the suspension state of the torque boost command is released, and the rising slope of the command current of the traction drive subsystem is discretized and constrained according to the current voltage change rate.

[0069] like Figure 2As shown, the data acquisition and filtering logic includes acquiring bus and grid-side voltage data and dynamically filtering high-frequency noise using a dead zone; trigger judgment and boundary adjustment includes extracting grid-side low-frequency fluctuation characteristics and adjusting the first judgment boundary value; timing suspension and rest control includes suspending torque boosting commands and enabling power rest periods; power assessment and limit cutoff includes adjusting adjustable power to the maximum power threshold and cutting off command current upper limit; state recovery and discrete constraints include voltage change rate falling back to a safety boundary and discrete constraint current rising slope. All the above data and logic flows work together to improve transient voltage stability.

[0070] Example 5: Before the on-site deployment of multi-subsystem network interconnection in the train bus controller, the logic arbitration chip in the central control unit first enters the baseline calibration state. The locomotive remains stationary, the main converter transformer is in an unloaded state, and the logic arbitration chip continuously reads the intermediate DC bus voltage data for 1000 sampling cycles. The static voltage fluctuation residual variance value under a fixed strong electromagnetic background interference is extracted by a shift register. Three times the envelope width of this variance value is used as the initial calibration term for the convergence safety boundary value, which is used to determine whether the transient impact has ended.

[0071] During the same calibration process, the logic arbitration chip unidirectionally sends timing probe pulse signals with control words to the local communication bus interface between the traction drive subsystem and the auxiliary power supply system. A hardware timer is started during transmission. After the local actuators of the two subsystems return response handshake data packets, the central control unit reads the timer count difference to obtain the total network communication delay data within the current network topology channel. This data is written to the parameter configuration area of ​​the main controller's non-volatile memory as the preset duration for subsequent corrections. The initial time constant.

[0072] After the logic arbitration chip completes the extraction of state parameters, it writes the convergence safety boundary value, corrected by the initial calibration term, into the threshold comparison register of the control state machine. The main controller then uses the total network communication delay data stored in the parameter configuration area to determine the preset duration. The initial value of the internal step counter is reconstructed so that the joint control sequence still executes with a uniform discrete response step size under different communication delay boundaries.

[0073] To preserve latency acquisition capabilities when bus bandwidth is exhausted or the network is congested, the central control unit sets up a time-division reusable network probing channel based on high-priority hard interrupts at the hardware level. This channel uses the highest physical interrupt level. In the first 0.01ms after the start of each state adjustment cycle, the main controller temporarily suspends the bus transmission requests for regular service data through hardware, using the bus bandwidth within this time window to send micro-latency probe data packets. After the probe data packet completes the handshake with the receiving end and returns latency data, the main controller releases the interrupt and resumes regular service bus transmission. Through this channel, even when the service network is congested, the central control unit can still obtain the total network communication latency data used to determine the communication status.

[0074] The central control unit compares the real-time total network communication delay data with the rigid communication threshold. The rigid communication threshold is pre-stored in the parameter configuration area, and its value is calibrated according to the maximum round-trip transmission time allowed by the train communication network in the joint power allocation control. When the total network communication delay data exceeds the rigid communication threshold, the control state machine determines that the system has entered a communication congestion state and sets the communication congestion flag.

[0075] After the system enters a communication congestion state, the central control unit stops processing the first judgment boundary value. The translation adjustment no longer changes the trigger boundary based on the low-frequency fluctuations on the grid side; at the same time, it stops the centralized cutoff of the command current upper limit of the traction drive subsystem to avoid continuing to issue cross-subsystem joint limiting commands in the event of time delay data distortion or control message lag. Subsequently, the traction drive subsystem and the auxiliary power supply system switch to independent limiting control state, and each limits the local power output command according to the fixed transformer drop power coefficient stored locally.

[0076] In the system parameter configuration area, the voltage drop power factor of the traction drive subsystem is set to 0.05 per kilowatt, which means that when the power changes by 1 kW due to the transient voltage rise of the bus, its local power output command is actively reduced by 5%; the voltage drop power factor of the auxiliary power supply system is set to 0.02 per kilowatt. Both factors are determined by the safety electrical withstand capability of the corresponding subsystem under independent amplitude limiting control, and are directly called by the local control board under communication congestion. The main controller then sends a calibration completion signal to the peripheral port, clears the historical residual content in the data buffer, and switches the power enable ports of different functional vehicle subsystems of the whole vehicle to the initial ready state for receiving transient dispatch commands.

[0077] Example 6: Before performing offline parameter calibration on the combined power distribution system, the current central control unit enters the system identification mode and controls the auxiliary power supply system to operate at a preset power step value. The intermediate DC bus is loaded, and its voltage response waveform is synchronously recorded via a high-speed data sampling link. The central control unit then... The equivalent output impedance of the main circuit of the current train power supply system is identified by the least squares method. ,in, This represents the voltage drop difference at the intermediate DC bus. To assist in understanding the actual output current variation of the power supply system, this value is written into the parameter configuration area of ​​the non-volatile memory as a physical reference for the initial state of the system.

[0078] The logic arbitration chip then establishes a network latency compensation model. In an offline debugging environment, the central control unit traverses the simulated total network communication latency data from 5ms to 25ms and observes the DC bus secondary overshoot amplitude generated when the traction torque is restored. The baseline slope adjustment factor in the central control unit does not actually rely on a continuously changing idealized exponential function for calculation. Instead, it is determined by directly calling the discrete-time mapping address table pre-installed in read-only memory. The microprocessor performs multi-level logical interval matching based on the real-time acquired discrete total network communication delay data. Specifically, the engineering matching benchmark is as follows: when the measured total network communication delay data is in the range of 0ms to 5ms, the baseline slope adjustment factor is directly assigned a value of 1.00; when the total network communication delay data is in the range of greater than 5ms and less than or equal to 10ms, the baseline slope adjustment factor is discretely... The baseline slope adjustment factor is configured to be 0.85. When the total network communication latency data is in the range of greater than 10ms and less than or equal to 15ms, the baseline slope adjustment factor is discretely configured to be 0.70. When the total network communication latency data is in the range of greater than 15ms and less than or equal to 25ms, the baseline slope adjustment factor is forcibly limited to a fixed safety limit value of 0.50. Through the above discrete interval addressing action, a deterministic dimensionless control constant is directly output, thereby completely blocking the introduction of unknown variable constants at the underlying hardware execution level with deterministic discrete numbers. The baseline slope adjustment factor is then determined through nonlinear data regression. Total network communication latency data The functional relationship between them is ,in, This is the initial reference slope adjustment factor under zero-delay conditions. The time-delay sensitive decay constant, This represents the total network communication latency data.

[0079] In actual operation, the central control unit operates according to real-time measurements. calculate Combined with the current voltage change rate and the first judgment boundary value after translation adjustment Calculate the torque boost slope gating coefficient , according to It is confirmed that, among them, The rate of change of voltage during the current sampling period. The first judgment boundary value after translation adjustment is used. When the communication delay increases or the voltage change rate approaches the first judgment boundary value, the limit of the rise slope of the command current in the recovery phase of the traction drive subsystem increases accordingly.

[0080] After releasing the torque boost command from its suspended state, the central control unit first reads the nominal current rise rate of the traction drive subsystem and compares this nominal current rise rate with... Multiply to obtain the target rising slope; then gradually increase the given current value of the traction drive subsystem according to the target rising slope until the given current value reaches the current required for the target torque. The above process makes the current recovery process subject to the current voltage change rate and the total network communication delay data, rather than directly recovering to the original nominal slope after the suspension is released.

[0081] After completing the discretization constraint, the central control unit converts the current increment control quantity corresponding to the target rising slope into a drive level signal, and transmits it back to the traction converter control board inside the traction drive subsystem through the train communication network. After receiving the drive level signal, the traction converter control board adjusts the pulse width modulation duty cycle of the inverter switching transistor according to the signal, so that the traction current recovers according to the target rising slope, thereby maintaining the transient electrical stability of the traction recovery process under different on-board communication load conditions.

[0082] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0083] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A joint control method for power distribution of high-speed trains, characterized in that, Includes the following steps: Step S1: The central control unit acquires intermediate DC bus voltage data and grid-side power supply voltage data, which characterize the transient power state between the traction drive subsystem and the auxiliary power supply system. Step S2: Calculate the voltage change rate based on the intermediate DC bus voltage data, and filter out the high-frequency voltage noise signal in the voltage change rate through dynamic dead-time filtering logic; Step S3: Extract the low-frequency fluctuation characteristics of the grid-side power supply voltage data, and adjust the first judgment boundary value of the trigger power timing adjustment according to the step direction of the low-frequency fluctuation characteristics. Step S4: When the voltage change rate after dynamic dead-zone filtering logic processing exceeds the first judgment boundary value after translation adjustment, the torque boost command is suspended, the power rest period is started, and the available power boundary of the auxiliary energy storage unit is read. Step S5: If, during the power rest period, the current adjustable power of the auxiliary energy storage unit within the available power boundary reaches the maximum power threshold and the voltage change rate does not fall back to the convergence safety boundary, then the upper limit of the command current of the traction drive subsystem is cut off according to the ratio of the transient overvoltage difference value of the intermediate DC bus voltage data. Step S6: When the voltage change rate falls back to within the convergence safety boundary, the suspension state of the torque boost command is released, and the rising slope of the command current of the traction drive subsystem is discretized and constrained according to the current voltage change rate.

2. The joint control method for power distribution of high-speed trains according to claim 1, characterized in that, Step S2 involves filtering out high-frequency voltage noise signals from the voltage change rate using dynamic dead-zone filtering logic, including: Step S21, acquiring intermediate DC bus voltage data over multiple consecutive control cycles; Step S22, calculating the voltage difference sequence between adjacent control cycles; Step S23, extracting the amplitude of high-frequency components from the voltage difference sequence and setting the amplitude of high-frequency components as the dynamic dead-zone threshold of the dynamic dead-zone filtering logic; Step S24, setting the voltage change rate to 0 and outputting it when the absolute value of the voltage change rate in the current control cycle is less than the dynamic dead-zone threshold; Step S25, outputting the actual voltage change rate when the absolute value of the voltage change rate in the current control cycle is greater than or equal to the dynamic dead-zone threshold.

3. The joint control method for power distribution of high-speed trains according to claim 1, characterized in that, The first determination boundary value for shifting and adjusting the trigger power timing adjustment in step S3 includes: step S31, low-pass filtering the grid-side power supply voltage data to obtain a grid voltage effective value reference sequence characterizing low-frequency fluctuations; step S32, comparing the grid voltage effective value reference sequence at the current moment with the grid voltage effective value reference sequence at the previous historical moment to generate a voltage step direction identifier; step S33, when the voltage step direction identifier is a positive jump and the jump amplitude exceeds a given grid voltage fluctuation threshold, shifting the first determination boundary value upward by a fixed step increment; step S34, when the voltage step direction identifier is a reverse jump and the jump amplitude exceeds a given grid voltage fluctuation threshold, shifting the first determination boundary value downward by a fixed step increment.

4. The joint control method for power distribution of high-speed trains according to claim 1, characterized in that, It also includes a degradation control step S7 for communication anomalies. Step S7 includes: Step S71, real-time acquisition of the total network communication delay data between the central control unit, the traction drive subsystem, and the auxiliary power supply system; Step S72, when the total network communication delay data exceeds the rigid communication threshold, determining that the system has entered a communication congestion state; Step S73, in the communication congestion state, the central control unit stops the translation adjustment of the first determination boundary value and the cutoff of the command current upper limit of the traction drive subsystem; Step S74, the traction drive subsystem and the auxiliary power supply system switch to independent limiting control state, and independently limit the local power output command according to their respective fixed transformer drop power coefficients.

5. The joint control method for power distribution of high-speed trains according to claim 1, characterized in that, Step S4 reads the available power boundary of the auxiliary energy storage unit, including: step S41, obtaining the current state of charge data and health characteristic variables of the auxiliary energy storage unit; step S42, calculating the maximum allowable charge and discharge rate of the auxiliary energy storage unit at the current ambient temperature based on the current state of charge data and health characteristic variables; step S43, multiplying the maximum allowable charge and discharge rate by the rated capacity of the auxiliary energy storage unit to obtain the real-time available power boundary.

6. The joint control method for power distribution of high-speed trains according to claim 1, characterized in that, Step S5 involves cutting off the upper limit of the command current of the traction drive subsystem according to the transient overvoltage difference value of the intermediate DC bus voltage data. This includes: Step S51, calculating the difference between the current intermediate DC bus voltage data and the standard rated bus voltage data to obtain the transient overvoltage difference value; Step S52, inputting the transient overvoltage difference value into the limiting ratio mapping table to find the corresponding overvoltage cutoff coefficient; Step S53, multiplying the overvoltage cutoff coefficient by the rated command current upper limit of the traction drive subsystem to generate the cutoff command current upper limit; and Step S54, limiting the actual output command current of the traction drive subsystem within the cutoff command current upper limit.

7. The joint control method for power distribution of high-speed trains according to claim 1, characterized in that, Step S6 involves discretizing the rise slope of the command current in the traction drive subsystem based on the current voltage change rate, including: Step S61, obtaining a pre-calibrated reference slope adjustment factor, and calculating the torque boost slope gating coefficient based on the reference slope adjustment factor, the current voltage change rate, and the first judgment boundary value after translation adjustment; Step S62, using the torque boost slope gating coefficient to discretize the rise slope of the command current; wherein, the torque boost slope gating coefficient is calculated according to the following formula: ,in, This is the torque boost slope gating coefficient; The baseline slope adjustment factor; The current rate of change of voltage; This is the first decision boundary value after translation adjustment.

8. A joint control method for power distribution of high-speed trains according to claim 7, characterized in that, Step S62 uses the torque boosting slope gating coefficient to discretize the command current rising slope, including: step S621, obtaining the nominal current rising slope of the traction drive subsystem; step S622, multiplying the nominal current rising slope by the torque boosting slope gating coefficient to generate the target rising slope; step S623, increasing the given current value of the traction drive subsystem according to the target rising slope until the given current value reaches the current required for the target torque.

9. A joint control method for power distribution of high-speed trains according to claim 1, characterized in that, The criteria for determining whether the voltage change rate falls back to the convergence safety boundary include: the absolute value of the voltage change rate is lower than the safety recovery threshold for three consecutive discrete control cycles; and the intermediate DC bus voltage data shows a monotonic trend of convergence towards the rated voltage range for three consecutive discrete control cycles.

10. A joint control method for power distribution of high-speed trains according to claim 1, characterized in that, It also includes a pulse modulation drive step S8, which includes: step S81, the central control unit converts the rising slope of the command current after discretization constraint into a drive level signal; step S82, the drive level signal is transmitted back to the traction converter control board inside the traction drive subsystem through the train communication network; step S83, the traction converter control board adjusts the pulse width modulation duty cycle of the inverter switching transistor according to the drive level signal.

Images