Wide voltage input industrial environment adaptive charging method and system

Abstract

The application provides a wide-voltage-input industrial environment adaptive charging method and system, and relates to the technical field of charging control. The method comprises the following steps: collecting running parameters including input voltage in real time, analyzing fluctuation characteristics of the input voltage, and dividing the current input working condition level; based on the current input working condition level, generating an input-side constraint parameter set including alternating current current-limiting threshold and power climbing rate upper limit through a preset constraint parameter mapping relationship, and applying the input-side constraint parameter set to a control loop for constraint; determining a current charging stage and corresponding target charging current and voltage; when the working condition level changes, adjusting the target charging current based on available power; when the input voltage rebounds, restoring the target charging current according to a preset slope, and limiting the control loop. The application effectively solves the charging stability, continuity and availability problems under wide-voltage-input fluctuation by constructing an integrated control link.

Description

Technical Field

[0001] This application relates to the field of charging control technology, and more specifically, to an adaptive charging method and system for industrial environments with wide voltage input. Background Technology

[0002] Currently, industrial charging equipment serves as the core of energy supply for industries such as electric forklifts and AGVs, and its operational stability is directly related to the continuous operation efficiency of industrial logistics and production lines.

[0003] Traditional chargers handle input voltage fluctuations in a relatively simple and passive way. For example, they employ undervoltage or overvoltage protection strategies with fixed thresholds, directly triggering protection and shutting down the output when the input voltage deviates from the rated range. This approach is prone to causing repeated interruptions and restarts of the charging process in industrial environments with frequent grid fluctuations. Another common practice is to simply drate the output power by a fixed percentage or shut it down completely after detecting a voltage fluctuation, and then attempt to restart it once the voltage recovers. However, this approach lacks precise identification of fluctuation characteristics and smooth control of the recovery process, often resulting in a sharp drop in available output power during voltage fluctuations, making it impossible to maintain the current charging stage. Furthermore, at the moment of voltage recovery, if the control loop is not handled properly, a huge input inrush current can be introduced, potentially causing the upstream power distribution switch to trip, interfering with other sensitive equipment on the same grid, and even triggering repeated start-stop oscillations of the entire charging system, affecting the execution of battery management strategies and charging efficiency.

[0004] In recent years, attempts have been made to monitor the input voltage in real time and dynamically adjust the output. For example, software algorithms can be used to appropriately reduce the output power when the input voltage is low, in order to maintain the operation of the charger within a wider input range. However, these improvements still have certain limitations. First, existing technologies mostly focus on power adjustment under steady-state conditions, and do not adequately consider dynamic control during periods of drastic input voltage changes. When the input voltage drops rapidly, due to the lack of active limiting of the input power ramp-up rate and precise feedforward control of the bus voltage, the energy of the bus capacitor can easily be rapidly depleted, causing the downstream converter to lose lockout due to undervoltage, the loop to enter saturation, and ultimately requiring the intervention of the protection circuit. Second, during the recovery process of the input voltage from a drop, existing solutions typically restore power with a simple fixed slope. However, if there are still slight fluctuations in the power grid or changes in the load condition, this open-loop recovery is prone to causing current overshoot and system oscillation. More importantly, during these dynamic adjustments, the status indicators of the charging phase are often abnormally reset or forcibly switched, leading to a mismatch between the battery management system and the charger. This makes it impossible to guarantee the consistency of the entire charging process, resulting in significant fluctuations in charging time. Furthermore, existing technologies rarely combine the constraints of harsh industrial conditions such as low-temperature start-up and long-term heavy load with wide voltage input fluctuation control in a joint decision-making process, raising questions about the long-term reliability of the product in actual industrial environments.

[0005] Therefore, how to construct an adaptive charging control method that can accurately identify input fluctuation characteristics, dynamically generate and apply multi-dimensional constraints, coordinately control input and output side targets, and achieve smooth transition and stage continuity during abnormal disturbances and recovery has become a key technical challenge to improve the engineering availability and environmental adaptability of industrial charging equipment under wide voltage input conditions. Summary of the Invention

[0006] To address the shortcomings of existing technologies, embodiments of this application provide an industrial environment adaptive charging method and system with wide voltage input.

[0007] In a first aspect, embodiments of this application provide an industrial environment adaptive charging method with wide voltage input, including:

[0008] Real-time acquisition of operating parameters, including input voltage, analysis of the fluctuation characteristics of the input voltage, and classification of the current input operating condition level;

[0009] Based on the current input operating condition level, an extended state observer is constructed to estimate the disturbance during the input voltage fluctuation process, and the disturbance estimate is calculated. Based on the disturbance estimate, an input-side constraint parameter set, including at least an AC current limiting threshold and a power ramp-up rate upper limit, is generated through a preset constraint parameter mapping relationship. The input-side constraint parameter set is then applied to a preset control loop for constraint to obtain the input power.

[0010] Determine the current charging stage to obtain the corresponding target charging current and target charging voltage;

[0011] The current input operating condition level is compared with the operating condition level at the previous moment. If the operating condition level changes, based on the disturbance estimate, a correction amount of the adjustment coefficient of the current input power is calculated using a preset proportional-integral algorithm. The adjustment coefficient is used to characterize the utilization ratio of the current input power. The adjustment coefficient at the previous moment is added to the correction amount to obtain the current adjustment coefficient. The current input power is multiplied by the current adjustment coefficient to obtain the current available power. The maximum output current is calculated based on the ratio of the current available power to the current output voltage, and the target charging current is adjusted accordingly.

[0012] When the input voltage rises back to the preset fluctuation threshold, the target charging current is restored according to the preset recovery slope, and the control loop is limited during the recovery process.

[0013] Furthermore, the fluctuation characteristic analysis of the input voltage includes:

[0014] Set fluctuation judgment thresholds, which include amplitude deviation thresholds and rate of change thresholds;

[0015] If the amplitude deviation of the input voltage is greater than or equal to the amplitude deviation threshold, and the rate of change of the input voltage is greater than or equal to the rate of change threshold, the input voltage is determined to be in a fluctuating state.

[0016] The magnitude deviation of the input voltage is obtained by calculating the absolute value of the difference between the input voltage and the standard voltage.

[0017] Based on the amplitude deviation and the rate of change, the fluctuation level corresponding to the fluctuation state is classified as the current input operating condition level.

[0018] Furthermore, the step of classifying the fluctuation level corresponding to the fluctuation state as the current input operating condition level includes:

[0019] The input operating condition level includes at least a first operating condition level, a second operating condition level, and a third operating condition level, wherein the input operating condition level increases sequentially according to the degree of fluctuation.

[0020] The first operating condition level corresponds to an input voltage that is in a fluctuating condition where the amplitude deviation is less than or equal to the first amplitude threshold and the rate of change is less than or equal to the first rate threshold.

[0021] The second operating condition level corresponds to an input voltage that is in a fluctuating condition where the amplitude deviation is greater than the first amplitude threshold but less than or equal to the second amplitude threshold and the rate of change is greater than the first rate threshold but less than or equal to the second rate threshold.

[0022] The third operating condition level corresponds to an input voltage that is in a fluctuating condition where the amplitude deviation is greater than the second amplitude threshold and the rate of change is greater than the second rate threshold.

[0023] Among them, the first amplitude threshold is less than the second amplitude threshold, and the first rate threshold is less than the second rate threshold.

[0024] Furthermore, the real-time estimation of disturbances during input voltage fluctuations by constructing an extended state observer includes:

[0025] The input voltage is used as the input to the extended state observer, and the lumped disturbance is extended into the state variable of the extended state observer.

[0026] The lumped disturbances include: input voltage fluctuations, line impedance changes, and load abrupt changes;

[0027] The observer gain matrix of the extended state observer is calculated based on the current input operating condition level, and the lumped disturbance is estimated, and the disturbance estimate is output.

[0028] Determine the AC current limiting threshold and power ramp-up rate baseline values ​​based on the current input operating condition level;

[0029] The absolute value of the disturbance estimate is compared with the preset rated disturbance value. The disturbance degree coefficient is determined based on the comparison result, and the correction factor is determined based on the disturbance degree coefficient.

[0030] It also obtains the current temperature and determines the temperature correction coefficient through fuzzy logic reasoning;

[0031] The AC current limiting threshold and the power ramp-up rate benchmark are multiplied by the correction factor and the temperature correction coefficient, respectively, to calculate the AC current limiting threshold and the upper limit of the power ramp-up rate, which are then used as the set of input-side constraint parameters.

[0032] Further, calculating the observer gain matrix of the extended state observer includes:

[0033] The observer bandwidth of the extended state observer is determined based on the current input condition level, wherein the observer bandwidth increases as the current input condition level increases;

[0034] The observer bandwidth is substituted into the pole configuration calculation, and an observer gain matrix corresponding to the observer bandwidth is generated by a preset gain determination rule.

[0035] Furthermore, the correction amount for calculating the adjustment coefficient of the current input power using a preset proportional-integral algorithm includes:

[0036] When the current input condition level increases relative to the previous moment, the adjustment coefficient is used as a deregulation coefficient, and a negative correction amount is calculated based on the disturbance estimate using an incremental proportional-integral algorithm.

[0037] When the current input condition level decreases relative to the previous moment, the adjustment coefficient is used as an increase coefficient, and a positive correction amount is calculated based on the disturbance estimate using an incremental proportional-integral algorithm.

[0038] The correction amount includes a proportional correction amount and an integral correction amount;

[0039] The target charging current is compared with the maximum output current. If the target charging current is greater than the maximum output current, the maximum output current is taken as the target charging current.

[0040] Specifically, the current charging stage remains unchanged during the adjustment of the target charging current.

[0041] Further, restoring the target charging current according to the preset recovery ramp includes:

[0042] The cumulative calorific value is obtained by integrating the square of the target charging current over time using the conditional integral anti-saturation method, and a first correction coefficient is determined.

[0043] The first correction coefficient is multiplied by the preset recovery step size to obtain the second step size value, and then increased cycle by cycle based on the preset charging cycle to obtain the current value to be processed.

[0044] The ratio of the current available power to the current output voltage is obtained as the upper limit of the current, and the current value to be processed is compared with the upper limit of the current, and the smaller value of the two is taken as the target charging current of the current cycle.

[0045] During the recovery process, the integral term is constrained according to the integral accumulation condition.

[0046] Furthermore, the constraint on the integral term based on the integral accumulation condition during the recovery process includes:

[0047] If the target charging current of the current cycle reaches the upper limit of the current and the current control error direction is the same as the integral accumulation direction, the accumulation of the integral term is stopped.

[0048] The current control error direction is determined by the sign of the error value in the current cycle, and the integral accumulation direction is determined by the sign of the difference between the integral value in the current cycle and the integral value in the previous cycle.

[0049] Furthermore, maintaining the current charging stage unchanged during the target charging current adjustment process includes:

[0050] A preset gradient descent algorithm is used to construct a cost function based on the deviation between the current output voltage and the target charging voltage, and the deviation between the current output current and the target charging current, with the charging stage status identifier as the variable to be optimized.

[0051] The optimal solution of the cost function is obtained by using the restriction of switching between charging stages during the reduction of the target charging current or the gradual increase of the target charging current as an optimization constraint.

[0052] If the optimal solution obtained by solving is consistent with the current locked charging stage state flag, the current charging stage remains unchanged;

[0053] If the optimal solution is inconsistent with the currently locked charging stage status identifier and the preset optimization convergence condition is met, the locking of the charging stage status identifier is released.

[0054] Secondly, embodiments of this application also provide an industrial environment adaptive charging system with wide voltage input, including: an identification module, a constraint module, a target determination module, an adjustment module, and a recovery module;

[0055] Identification module: Real-time acquisition of operating parameters including input voltage, analysis of fluctuation characteristics of the input voltage, and classification of the current input operating condition level;

[0056] Constraint module: Based on the current input operating condition level, it generates an input-side constraint parameter set that includes at least an AC current limiting threshold and a power ramp-up rate upper limit through a preset constraint parameter mapping relationship, and applies the input-side constraint parameter set to a preset control loop to constrain it, thereby obtaining the input power;

[0057] Target determination module: Determines the current charging stage and obtains the corresponding target charging current and target charging voltage;

[0058] Adjustment module: Compares the current input operating condition level with the operating condition level at the previous moment. When the operating condition level changes, based on the disturbance estimate, it calculates the correction amount of the adjustment coefficient of the current input power using a preset proportional-integral algorithm. The adjustment coefficient is used to characterize the utilization ratio of the current input power. The adjustment coefficient at the previous moment is added to the correction amount to obtain the current adjustment coefficient. The current input power is multiplied by the current adjustment coefficient to obtain the current available power. The maximum output current is calculated based on the ratio of the current available power to the current output voltage, and the target charging current is adjusted accordingly.

[0059] Recovery module: When the input voltage rises back to the preset fluctuation judgment threshold, the target charging current is restored according to the preset recovery slope, and the control loop is limited during the recovery process.

[0060] Compared with the prior art, the effective effects achieved by this application are as follows:

[0061] This application achieves refined classification of input operating condition levels by introducing multi-dimensional fluctuation characteristic analysis such as voltage change rate; by constructing an extended state observer, input voltage fluctuations, line impedance changes, and load abrupt changes are treated as lumped disturbances for real-time estimation, and AC current limiting threshold and power ramp-up rate upper limit are dynamically generated accordingly, so that input-side constraints can be adaptively adjusted with changes in input operating condition levels, effectively mitigating the impact of grid instability on the AC side of the charger; when the input operating condition level changes, this application smoothly adjusts the output current target based on the real-time estimated disturbance and available power through an incremental proportional-integral algorithm. This method ensures that the output-side target maintains a high degree of coordination with the input-side constraints during input voltage dips. By introducing a cost function and optimization constraints, it forcibly maintains the current charging phase state, preventing unnecessary resets during the charging phase and effectively solving the problem of charging phase state mismatch during power adjustment in traditional solutions. When the input voltage recovers, a conditional integral anti-saturation method is adopted, combining the accumulated calorific value and the real-time available power limit, and increasing the target charging current cycle by cycle according to a preset recovery step size. By performing direction judgment and accumulation constraints on the integral term, current overshoot and system oscillation during the recovery process are effectively suppressed. In addition, this invention incorporates industrial operating condition constraints such as temperature into the derating decision, enabling the system to remain stable and controllable under extreme conditions of low temperature, heavy load, and wide input fluctuations, significantly improving the engineering availability and environmental adaptability of industrial charging equipment. Attached Figure Description

[0062] Figure 1 A flowchart of an industrial environment adaptive charging method with wide voltage input provided in an embodiment of this application;

[0063] Figure 2A flowchart for generating a set of input-side constraint parameters provided in this application embodiment;

[0064] Figure 3 A flowchart for determining the charging phase lockout for wide voltage input provided in this application embodiment;

[0065] Figure 4 This is a schematic diagram of an industrial environment adaptive charging system with wide voltage input provided in an embodiment of this application. Detailed Implementation

[0066] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0067] See Figure 1 This is a flowchart of an industrial environment adaptive charging method with wide voltage input provided in this embodiment. The method includes steps S101 to S105, wherein:

[0068] S101: Real-time acquisition of operating parameters including input voltage, analysis of fluctuation characteristics of the input voltage, and classification of the current input operating condition level;

[0069] S102: Based on the current input operating condition level, generate an input-side constraint parameter set that includes at least an AC current limiting threshold and a power ramp-up rate upper limit through a preset constraint parameter mapping relationship, and apply the input-side constraint parameter set to a preset control loop for constraint to obtain the input power;

[0070] S103: Determine the current charging stage and obtain the corresponding target charging current and target charging voltage;

[0071] S104: Compare the current input operating condition level with the operating condition level at the previous moment. If the operating condition level changes, calculate the current available power based on the input power and adjust the target charging current.

[0072] S105: When the input voltage rises back to the preset fluctuation judgment threshold, the target charging current is restored according to the preset recovery slope, and the control loop is limited during the recovery process.

[0073] Regarding step S101:

[0074] Wide voltage input means that the charging equipment can adapt to a wide range of input voltage variations. Taking common industrial charging equipment as an example, the conventional input voltage range is usually around 10% above and below the rated voltage. For example, for a device with a rated voltage of 380 volts AC, its conventional operating voltage range is about 342 volts to 418 volts. However, for wide voltage input devices, the adaptability range can be extended to 30% above and below the rated voltage. For example, for the same device with a rated voltage of 380 volts AC, its operating voltage range can reach 266 volts to 494 volts, thus adapting to the more complex power supply environment in industrial sites.

[0075] In practice, the operating parameters of the charging equipment are collected in real time, including: input voltage, input current, input voltage conversion rate, input frequency, output voltage, output current, and temperature parameters of key components; for example, the key components include heat-generating elements in the power conversion control loop, such as power switching transistors, transformers, rectifier diodes, and heat sinks.

[0076] The temperature parameters include, for example, the case temperature of the power switch transistor, the core temperature of the transformer, the tube temperature of the rectifier diode, and the substrate temperature of the heat sink.

[0077] In practice, the input voltage is analyzed for fluctuation characteristics to classify the current input operating condition level.

[0078] As an optional implementation method, the dq transformation detection method based on instantaneous reactive power theory can be used. The phase information of the input voltage is extracted by a phase-locked loop, and the dq transformation is performed on the collected three-phase input voltage to obtain the d-axis component and the q-axis component.

[0079] The d-axis component represents the active component or amplitude-dependent component of the input voltage, reflecting the amplitude information of the input voltage; the q-axis component represents the reactive component or phase-dependent component of the input voltage, reflecting the phase information of the input voltage.

[0080] By monitoring changes in the d-axis component, events such as increases or decreases in the input voltage can be detected.

[0081] For example, when the d-axis component decreases, the input voltage decreases; similarly, when the d-axis component increases, the input voltage increases; when the d-axis component remains relatively stable, the input voltage is normal.

[0082] In a specific implementation, taking a charging device with a rated voltage of 380 volts AC as an example, the d-axis component can be set to 100, that is, for every change of 1 in the d-axis component, the corresponding input voltage changes by 3.8 volts; the d-axis component is fixedly set in the controller.

[0083] Calculate the difference between the current d-axis component and the reference value 100, multiply the difference by 3.8 to obtain the voltage deviation value; add the voltage deviation value to the input voltage of 380 volts to obtain the effective value of the input voltage.

[0084] For example, when the input voltage drops, the d-axis component will be less than 100. Assuming the current d-axis component is 95, the calculated effective value of the input voltage is 361 volts.

[0085] When the input voltage increases, the d-axis component will be greater than 100. Assuming the current d-axis component is 110, the effective value of the input voltage can be calculated to be 418 volts.

[0086] When the input voltage is at the standard value, the d-axis component is 100, and the difference from the reference value of 100 is 0. The effective value of the input voltage is 380 volts.

[0087] Similarly, by calculating the change of the d-axis component per unit time, the rate of change of the input voltage can be obtained; the calculation period of the rate of change can be set as needed, for example, once per second.

[0088] In practice, the difference between the average value of the d-axis component in the current second and the average value of the d-axis component in the previous second is calculated. This difference is multiplied by 3.8 to obtain the voltage change, and then divided by one second to obtain the rate of change, in volts per second.

[0089] For example, if the d-axis component is 100 in the previous second and 95 in the current second, the rate of change can be calculated to be 19 volts per second.

[0090] As an optional implementation, a fluctuation judgment threshold is set, which includes an amplitude deviation threshold and a rate of change threshold; the amplitude deviation threshold is used to determine the degree to which the input voltage deviates from the standard voltage, and the rate of change threshold is used to determine the severity of the voltage change.

[0091] The fluctuation judgment threshold is determined based on the wide voltage input range of the charging device.

[0092] In practical implementation, taking a rated voltage of 380 volts as an example, the amplitude deviation threshold is set to 19 volts as the standard voltage, corresponding to 5% of the rated voltage; the rate of change threshold is set to 19 volts per second, corresponding to 5% change of the rated voltage per second.

[0093] If the amplitude deviation of the input voltage is greater than or equal to the amplitude deviation threshold, and the rate of change of the input voltage is greater than or equal to the rate of change threshold, the input voltage is determined to be in a fluctuating state; wherein, the amplitude deviation of the input voltage is obtained by calculating the absolute value of the difference between the input voltage and the standard voltage.

[0094] Based on the amplitude deviation and the rate of change, the fluctuation level corresponding to the fluctuation state is classified as the current input operating condition level.

[0095] The input operating condition level includes at least a first operating condition level, a second operating condition level, and a third operating condition level, wherein the input operating condition level increases sequentially according to the degree of fluctuation.

[0096] The first operating condition level corresponds to an input voltage that is in a fluctuating condition where the amplitude deviation is less than or equal to the first amplitude threshold and the rate of change is less than or equal to the first rate threshold.

[0097] The second operating condition level corresponds to an input voltage that is in a fluctuating condition where the amplitude deviation is greater than the first amplitude threshold but less than or equal to the second amplitude threshold and the rate of change is greater than the first rate threshold but less than or equal to the second rate threshold.

[0098] The third operating condition level corresponds to an input voltage that is in a fluctuating condition where the amplitude deviation is greater than the second amplitude threshold and the rate of change is greater than the second rate threshold.

[0099] Among them, the first amplitude threshold is less than the second amplitude threshold, and the first rate threshold is less than the second rate threshold.

[0100] In specific implementation, the first amplitude threshold is the aforementioned amplitude deviation threshold, set to 19 volts; the second amplitude threshold can be set to 57 volts, corresponding to 15% of the rated voltage.

[0101] The first rate threshold is the aforementioned rate of change threshold, set at 19 volts per second; the second rate threshold is set at 57 volts per second, corresponding to a 15% change in rated voltage per second.

[0102] For example, if an industrial charging device is connected to a three-phase power grid, the controller calculates the average value and rate of change of the d-axis component every second; after calibration, the d-axis component 100 corresponds to an input voltage of 380 volts.

[0103] During the first one-second cycle, the d-axis component is 100; the effective value of the input voltage is 380 volts, and the amplitude deviation is 0 volts.

[0104] Compared to the previous second, the difference in the d-axis component is 0, the voltage change is 0 volts, and the rate of change is 0 volts per second. At this time, the amplitude deviation is 0 volts less than 19 volts, and the rate of change is 0 volts per second less than 19 volts per second, which does not meet the fluctuation judgment condition, and the input voltage is in a normal state.

[0105] During the second one-second cycle, the d-axis component drops from 100 to 95, the effective value of the input voltage is 361 volts, and the amplitude deviation is 19 volts; the rate of change is equal to 19 volts divided by 1 second, which equals 19 volts per second.

[0106] Compared to the previous second cycle, the amplitude deviation of 19 volts equals the first amplitude threshold, and the rate of change of 19 volts per second equals the first rate threshold. Both conditions simultaneously reach their thresholds, indicating that the input voltage is in a fluctuating state. Since both the amplitude deviation and the rate of change equal the first rate threshold, it falls under the first operating condition level.

[0107] During the third one-second cycle, the d-axis component decreased from 95 to 90. The effective value of the input voltage was 323 volts, the amplitude deviation was 38 volts, and the rate of change was 38 volts per second.

[0108] At this point, the amplitude deviation of 38 volts is greater than the first amplitude threshold of 19 volts and less than the second amplitude threshold of 57 volts, and the rate of change of 38 volts per second is greater than the first rate threshold of 19 volts per second and less than the second rate threshold of 57 volts per second, which is determined to be the second operating condition level.

[0109] Regarding step S102:

[0110] See Figure 2 The flowchart for generating a set of input-side constraint parameters provided in this application embodiment includes:

[0111] S201: Establish the state-space expression of the charging device through mechanism modeling, and establish an extended state model;

[0112] S202: Construct an extended state observer based on the extended state model to estimate the bus voltage, bus voltage change rate and lumped disturbance in real time, and output the disturbance estimate;

[0113] S203: Determine the observer bandwidth based on the current input operating condition level, calculate the observer gain matrix corresponding to the observer bandwidth using the pole placement method, and adaptively adjust the tracking speed of the observer disturbance in the extended state.

[0114] S204: Calculate the disturbance degree coefficient based on the disturbance estimate, determine the correction factor according to the disturbance degree coefficient, and obtain the temperature correction coefficient through fuzzy logic reasoning. Apply the disturbance degree coefficient and the temperature correction coefficient together to the preset constraint parameters to generate the input-side constraint parameter set.

[0115] Regarding S201:

[0116] In a specific implementation, the state-space expression of the controlled object is first established using the mechanism modeling method. Taking the power conversion circuit of the charging device as the object, and based on the circuit topology and Kirchhoff's laws, the voltage across the bus capacitor is selected as the first state variable, the current in the output filter inductor is selected as the second state variable, the input voltage is selected as the measurable input quantity, and the duty cycle of the drive signal of the power switch is selected as the control input quantity.

[0117] According to the circuit topology, the rate of change of the bus voltage is equal to the output filter inductor current minus the load current divided by the bus capacitance value, and the rate of change of the output filter inductor current is equal to the input voltage multiplied by the duty cycle minus the bus voltage divided by the filter inductor value.

[0118] The dynamic behavior of the circuit can be represented by two relations, and the system of differential equations is as follows:

[0119] ;

[0120] ;

[0121] in, Bus voltage versus time rate of change, For the output filter inductor current versus time rate of change, Bus voltage To output the filter inductor current, For load current, This is the bus capacitance value. This is the value of the filter inductance. Input voltage, The duty cycle of the drive signal for the power switch transistor is used as a control input.

[0122] In practical implementation, the above differential equations are rearranged to separate the relationship between the second derivative of the bus voltage and the duty cycle; by differentiating the first equation and substituting it into the second equation, the expression for the second derivative of the bus voltage is obtained as follows:

[0123] ;

[0124] In practical implementation, by incorporating input voltage fluctuations into the lumped disturbance, a first parameter can be defined. The expression for the second derivative of the bus voltage can be modified as follows:

[0125]

[0126] Where f includes all uncertainties such as input voltage fluctuations, line impedance changes, load current changes, and model inaccuracies, as a lumped disturbance.

[0127] In practice, the three extended state variables are: the first state is the bus voltage, the second state is the rate of change of the bus voltage, and the third state is the lumped disturbance.

[0128] In specific implementation, the dynamic relationship of the state variables is that the rate of change of the bus voltage is equal to the second state variable; the rate of change of the bus voltage is equal to the third state variable plus the first parameter multiplied by the control input; wherein the rate of change of the lumped disturbance is unknown but bounded.

[0129] The lumped disturbances include: input voltage fluctuations, line impedance changes, and load abrupt changes.

[0130] As an optional implementation, based on the above third-order structure, an extended state observer is constructed. The structure of the extended state observer in the continuous time domain includes three differential equations:

[0131] The first equation represents the rate of change of the bus voltage estimate, which is equal to the bus voltage rate of change estimate plus the first gain coefficient multiplied by the output estimation error;

[0132] The second equation represents the rate of change of the estimated bus voltage change, which is equal to the lumped disturbance estimate plus the first parameter multiplied by the control input, plus the second gain coefficient multiplied by the output estimation error.

[0133] The third equation represents the rate of change of the lumped disturbance estimate, which is equal to the third gain coefficient multiplied by the output estimation error.

[0134] The first gain coefficient, the second gain coefficient, and the third gain coefficient are determined based on the observer bandwidth of the extended state observer.

[0135] For example, typically the first gain factor can be equal to three times the observer bandwidth; the second gain factor can be equal to the square of three times the observer bandwidth; and the third gain factor can be equal to the cube of the observer bandwidth.

[0136] The output estimation error is defined as the difference between the actual measured value of the bus voltage and the estimated value of the bus voltage.

[0137] Regarding S202:

[0138] In practice, the inputs of the extended state observer include control inputs and voltage measurements; the outputs of the extended state observer include estimates of the bus voltage, the rate of change of the bus voltage, and the lumped disturbance.

[0139] In a practical implementation, the extended state observer operates within a digital controller with a fixed sampling period. The sampling period is determined by a hardware timer in the controller, for example, set to one hundred microseconds.

[0140] Within each sampling period, the extended state observer performs a recursive calculation, discretizing the continuous differential equation to obtain two steps: prediction and correction. The prediction step calculates the predicted state value for the current moment based on the disturbance estimate from the previous moment and the current control input, according to the dynamic relationship of the extended third-order model. The correction step uses the error between the actual measured value and the predicted value of the bus voltage at the current moment to correct the three state prediction values, obtaining the final disturbance estimate for the current moment.

[0141] Regarding S203:

[0142] The observer gain matrix of the extended state observer is calculated based on the current input operating condition level, and the lumped disturbance is estimated, and the disturbance estimate is output.

[0143] The observer bandwidth of the extended state observer is determined based on the current input condition level, wherein the observer bandwidth increases as the current input condition level increases.

[0144] As an optional implementation, the observer gain coefficient is determined by the pole placement method. The rules for calculating the three gain coefficients of the observer bandwidth are as follows: the first gain coefficient is equal to three times the observer bandwidth, the second gain coefficient is equal to three times the square of the observer bandwidth, and the third gain coefficient is equal to the cube of the observer bandwidth.

[0145] Regarding S204:

[0146] As an optional implementation, a segmented mapping algorithm based on the degree of perturbation is adopted, and temperature information is used for correction.

[0147] In practice, the disturbance degree coefficient is defined by comparing the absolute value of the lumped disturbance estimate with the rated disturbance value.

[0148] When the absolute value of the lumped disturbance estimate is less than the rated disturbance value, the disturbance degree coefficient is the absolute value divided by the rated disturbance value; when the absolute value is greater than or equal to the rated disturbance value, the disturbance degree coefficient takes the maximum value of 1.

[0149] The rated disturbance value is predetermined based on the design tolerance of the charging equipment. For example, for a 3 kW device, the rated disturbance value can be set to 15 amps, which represents the upper limit of disturbance that the device can withstand.

[0150] As an optional implementation, a mapping curve is designed for each constraint parameter based on the perturbation degree coefficient.

[0151] The mapping curve is piecewise linear, determined by three key points. The first key point corresponds to the rated parameter value when the perturbation coefficient is 0, the second key point corresponds to the intermediate parameter value when the perturbation coefficient is at the intermediate threshold, and the third key point corresponds to the minimum parameter value when the perturbation coefficient is 1. The intermediate threshold is typically 0.5, but can be adjusted according to specific needs.

[0152] As an optional implementation, temperature correction can be achieved using a fuzzy logic-based method. The highest temperature of the key component is used as input, and a temperature correction coefficient is obtained through fuzzy inference.

[0153] In practice, the temperature is divided into three fuzzy sets: low temperature, medium temperature, and high temperature. Each set is defined by a triangular membership function.

[0154] The low-temperature group covers the range from room temperature to the first temperature threshold, the medium-temperature group covers the range from the first threshold to the second threshold, and the high-temperature group covers the range from the second threshold to the maximum allowable temperature. The parameters of the membership function are set according to the statistical law of the influence of temperature on device performance. The first and second temperature thresholds are set based on the maximum allowable operating temperature of key devices and the derating requirements of industrial equipment.

[0155] In practical implementation, the fuzzy inference rule base includes: if the temperature is low, the temperature correction coefficient is 1; if the temperature is medium, the temperature correction coefficient decreases linearly with the increase of temperature; if the temperature is high, the temperature correction coefficient decreases rapidly to the minimum protection value.

[0156] The output of each rule is also a fuzzy set. When defuzzifying, a weighted average method is used, that is, based on the membership degree of the input temperature in each fuzzy set, the center value of the output of each rule is weighted and averaged to obtain an accurate temperature correction coefficient.

[0157] The final constraint parameters applied to the control loop are the basic constraint parameters multiplied by the temperature correction factor.

[0158] In practice, after analyzing the characteristics of the input voltage fluctuation, the current d-axis component is 90, the effective value of the input voltage is 342 volts, the amplitude deviation is 38 volts, and the rate of change is 38 volts per second.

[0159] According to the above input condition level classification rules, the amplitude deviation of 38 volts is between 19 volts and 57 volts, and the rate of change of 38 volts per second is between 19 volts per second and 57 volts per second. Therefore, the current condition level is determined to be the second condition level.

[0160] Based on the second operating condition level, the basic constraint parameters are as follows:

[0161] AC current limiting threshold benchmark: Take 80% of the rated current of 15 Amperes, i.e., 12 Amperes.

[0162] The power ramp-up rate benchmark is 60% of the rated power ramp-up rate of 1000 watts per second, which is 600 watts per second.

[0163] The target voltage reference value for the bus is 95% of the rated voltage of 400 volts, which is 380 volts.

[0164] The equivalent power command baseline is 80% of the rated power of 3000 watts, which is 2400 watts.

[0165] Meanwhile, based on the second operating condition level, the gain scheduling extended state observer uses a bandwidth of 60 radians per second, and the estimated value of the lumped disturbance at the current moment is -23.7 amperes after recursive calculation.

[0166] The disturbance level correction is initiated, with the preset disturbance value set at 7.9 amperes. The absolute value of the lumped disturbance estimate is 23.7 amperes, which is greater than the preset disturbance value of 7.9 amperes. Therefore, the disturbance level coefficient is set to 1.

[0167] The basic constraint parameters are dynamically scaled according to the current disturbance by using a correction factor.

[0168] For example, the correction factor uses a linearly decreasing function with a reduction slope of 0.5, meaning the correction factor equals 1 minus 0.5 multiplied by the perturbation level coefficient. When the perturbation level coefficient is 1, the correction factor equals 0.5.

[0169] Multiplying the basic constraint parameters by a correction factor of 0.5 yields the disturbance-corrected parameters, where the AC current limiting threshold is 12 amps multiplied by 0.5, which equals 6 amps; the power ramp-up rate is capped at 300 watts per second; the bus target voltage is 190 volts; and the equivalent power command is capped at 1200 watts.

[0170] Entering temperature correction, assuming the highest temperature of the current key component is 90 degrees Celsius, according to the fuzzy logic temperature correction algorithm, when the temperature is 90 degrees Celsius, the membership degree of the low-temperature set is 0, the membership degree of the medium-temperature set is 0.5, and the membership degree of the high-temperature set is 0.5.

[0171] Set the activation levels of the three rules to 0, 0.5, and 0.5 respectively, and output the center values ​​to 1, 0.65, and 0.3 respectively.

[0172] Using the weighted average method, the temperature correction factor is calculated to be 0.475.

[0173] Multiplying the disturbance-corrected parameters by the temperature correction factor of 0.475 yields the final constraint parameters, where the AC current limiting threshold is 6 amps multiplied by 0.475, which equals 2.85 amps; the upper limit of the power ramp rate is 142.5 watts; the target bus voltage is 90.25 volts; and the upper limit of the equivalent power command is 570 watts.

[0174] For the current second operating condition level, combined with the lumped disturbance estimate of -23.7 amperes and the maximum temperature of 90 degrees Celsius, the generated set of input-side constraint parameters includes four specific values.

[0175] In contrast, if the current condition is classified as Level 1, the basic constraint parameters will be set to higher values. For example, under Level 1, the AC current limiting threshold baseline value can be 7.9 amps, the power ramp-up rate baseline value can be 1000 watts per second, the bus target voltage baseline value can be 400 volts, and the equivalent power command baseline value can be 3000 watts.

[0176] As an optional implementation, the set of constraint parameters is applied to the control loop of power generation.

[0177] In specific implementation, the control loop includes a control loop for the front-end PFC converter and a control loop for the rear-end DC / DC converter. The AC current limiting threshold is sent to the inner current loop of the PFC controller as a limiting value, and the upper limit of the power ramp-up rate is sent to the outer power loop of the PFC controller as a limit on the power command change rate, so as to ensure the input current waveform quality and system stability.

[0178] The above four constraint parameters act on different control loops. For example, the AC current limiting threshold of 2.85 amps is sent to the input current limiting loop. By monitoring the input current sampling value in real time, when the current value approaches or reaches the AC current limiting threshold, the limiter automatically adjusts the duty cycle command to prevent the current from continuing to rise.

[0179] The upper limit of the power ramp-up rate, 142.5 watts per second, is sent to the power command generation stage. When calculating the change in power command, the magnitude of the change per unit time is limited to ensure that the rate of change of power command is less than the upper limit of the power ramp-up rate.

[0180] The target bus voltage of 90.25 volts is fed into the voltage regulation stage. This stage compares the actual bus voltage with the target value and generates the setpoint for the inner current loop through a proportional-integral regulator.

[0181] The equivalent power command upper limit of 570 watts is sent to the power limiting stage. The power command value output by the power command generation stage is limited to ensure that it is less than the equivalent power command upper limit.

[0182] The controller ultimately outputs a duty cycle signal to drive the power switching transistor, thus limiting the actual input power to a range determined by these parameters.

[0183] Regarding step S103:

[0184] Determine the current charging stage and obtain the corresponding target charging current and target charging voltage.

[0185] In practice, the division of charging stages in industrial environments is determined based on battery type and charging characteristics. For example, charging stages include a pre-charging stage, a constant current charging stage, a constant voltage charging stage, and a float charging stage. The conversion thresholds and charging parameters for each stage differ for different types of batteries.

[0186] For example, lead-acid batteries typically use multi-stage constant current or pulse charging, while lithium batteries mostly use constant current and constant voltage two-stage charging.

[0187] The transitions between stages are determined by preset voltage and current thresholds, which are set based on the specifications provided by the battery manufacturer and relevant industry standards.

[0188] In practice, the pre-charge recovery voltage is set based on the battery's over-discharge protection threshold. For lead-acid batteries, the over-discharge voltage of a single cell is typically 10.5 volts. Therefore, the pre-charge recovery voltage for a 48-volt battery pack with four cells connected in series is set to 42 volts, leaving a safety margin. When the battery voltage is lower than this value, it indicates that the battery is in an over-discharged state and needs to be restored to charge with a small current to avoid damage to the battery from a large current surge.

[0189] The constant voltage conversion voltage is set based on the battery's full-charge voltage. For lead-acid batteries, the full-charge voltage of a single cell is typically 14.4 volts. Therefore, the constant voltage conversion voltage for a 48-volt battery pack consisting of four cells connected in series is set to 57.6 volts. When the battery voltage reaches this value, it indicates that the battery is about to be fully charged and it is necessary to switch from constant current charging to constant voltage charging to prevent overcharging.

[0190] The constant voltage termination current is set based on the battery's saturation level. It's typically determined by a percentage of the rated charging current, such as 10% to 20%. For a battery pack with a rated charging current of 20 amps, the constant voltage termination current is set to 2 amps (10% of the rated current). When the charging current drops to this value, it indicates that the battery is essentially fully charged, and charging can be terminated or switched to float charging.

[0191] The float charge voltage is set based on the battery's self-discharge compensation requirements. For lead-acid batteries, the float charge voltage of a single cell is typically 13.6 volts, so the float charge voltage of a 48-volt battery pack with four cells connected in series is set to 54.4 volts. This voltage is lower than the full charge voltage, which helps maintain the battery's charge level while preventing overcharging.

[0192] Taking an industrial charging device for charging a 48-volt lead-acid battery pack as an example, the battery pack consists of four 12-volt lead-acid batteries connected in series, with a rated charging current of 20 amps. The charging parameters for the lead-acid batteries are pre-stored in the controller as follows:

[0193] The pre-charge recovery voltage is 42 volts; the constant current charging current, i.e. the rated current, is 20 amps; the pre-charge current is 20% of the rated current, which is 4 amps; the constant voltage conversion voltage is 57.6 volts; the constant voltage charging voltage is 57.6 volts; the constant voltage termination current is set to 10% of the rated current, which is 2 amps; and the float charge voltage is 54.4 volts.

[0194] After charging begins, the controller reads the battery terminal voltage as 40 volts, which is lower than the pre-charge recovery voltage of 42 volts, thus determining that it has entered the pre-charge stage. At this time, the target charging current is set to 4 amps, and the target charging voltage is set to 42 volts as the upper limit. The controller then uses a current of 4 amps to restore the battery's charge.

[0195] When the battery voltage rises to 42 volts, reaching the pre-charge recovery threshold, the controller automatically switches to the constant current charging stage. At this time, the target charging current is set to 20 amps, and the target charging voltage is set to 57.6 volts (as the upper limit). The controller charges at a constant current of 20 amps, and the battery voltage continues to rise.

[0196] When the battery voltage rises to 57.6 volts, it reaches the constant voltage conversion threshold, and the controller switches to the constant voltage charging stage. At this time, the target charging voltage is set to 57.6 volts, and the target charging current is no longer directly given, but is automatically adjusted by the voltage loop according to the voltage error, and the charging current begins to gradually decrease.

[0197] As charging continues, the charging current gradually decreases. When the charging current drops to the preset constant voltage termination current, such as 2 amps, which is 10% of the rated charging current, the controller determines that the battery is basically fully charged and switches to the float charging stage.

[0198] During the float charging phase, the target charging voltage is set to the float charging voltage, corresponding to 13.6 volts for a single lead-acid battery. This voltage is lower than the constant voltage charging voltage, which can compensate for the battery's self-discharge loss without causing overcharging. At this time, the controller operates in constant voltage mode, and the output current automatically decreases to near zero as the battery state changes, maintaining the battery at a fully charged state.

[0199] Regarding step S104:

[0200] The current input operating condition level is compared with the operating condition level at the previous moment. If the operating condition level changes, the current available power is calculated based on the input power, and the target charging current is adjusted accordingly.

[0201] As an optional implementation, when the input operating condition level increases, the correction amount of the adjustment coefficient is calculated based on the disturbance estimate using an incremental proportional-integral algorithm, wherein the adjustment coefficient includes a deregulation coefficient and an escalation coefficient, and the correction amount includes a proportional correction amount and an integral correction amount.

[0202] The incremental proportional-integral algorithm consists of a proportional branch and an integral branch. The proportional branch generates a proportional correction based on the deviation between the current disturbance estimate and the preset benchmark value. The integral branch accumulates the deviation to generate an integral correction to eliminate steady-state error. The correction is the sum of the two.

[0203] In practice, the absolute value of the disturbance estimate is first compared with a preset disturbance reference value to obtain the deviation; the reference value is set according to the typical disturbance level under normal operating conditions of the equipment.

[0204] The proportional coefficient and integral coefficient are key parameters of the incremental proportional-integral algorithm, and an adaptive calculation method based on perturbation normalization can be adopted for the proportional-integral coefficient.

[0205] In practice, the disturbance estimate output by the extended state observer is used as the basis. First, its absolute value is taken, and then it is divided by the preset rated disturbance value to obtain the normalized disturbance intensity ratio. In this embodiment, the rated disturbance value is 10 amperes.

[0206] The initial reference values ​​for the proportional coefficient and integral coefficient are set to 0.9 and 0.06, respectively. These reference values ​​are the optimal parameters obtained by measuring the amplitude-frequency response of the output current loop of the charging equipment under the nominal operating condition where the input voltage is stable and without fluctuations, i.e., the disturbance estimate is 0, and by using the frequency domain pole placement method to pre-tune them according to the engineering criteria of a phase margin of not less than 60 degrees and an amplitude margin of not less than 6 dB.

[0207] When a disturbance occurs, both the proportional coefficient and the integral coefficient are dynamically scaled by the normalized disturbance intensity ratio, which is the sum of the initial reference value divided by one and the normalized disturbance intensity ratio.

[0208] For example, when the current disturbance estimate is -6 amperes, the absolute value is 6 amperes, the normalized disturbance strength ratio is 0.6, the one-plus-normalized disturbance strength ratio is 1.6, the proportionality coefficient is 0.9 ÷ 1.6 = 0.5625, and the integral coefficient is 0.06 ÷ 1.6 = 0.0375.

[0209] If the current disturbance estimate is -15 amperes, the absolute value is 15 amperes, the normalized disturbance intensity ratio is 1.5, the one-plus ratio is 2.5, the proportionality coefficient is 0.36, and the integral coefficient is 0.024.

[0210] Multiplying the proportional coefficient by the deviation amount yields the proportional correction; multiplying the integral coefficient by the deviation amount and adding it to the cumulative integral value of the previous period gives the current integral correction, while simultaneously limiting the cumulative integral value to prevent integral saturation. Adding the proportional correction and the integral correction gives the correction amount for the current period.

[0211] Since it is a reduction, the correction amount is usually negative.

[0212] The reduction factor is updated by making the current reduction factor equal to the previous reduction factor plus the correction amount.

[0213] In practice, the range of variation of the reduction coefficient is limited so that it is not lower than the preset minimum reduction coefficient and not higher than 1.

[0214] For the rationing process when the input operating condition level is reduced, the incremental proportional-integral algorithm is also used, but the correction amount is positive and the rationing coefficient is updated in the opposite direction.

[0215] The current adjustment coefficient is obtained by adding the adjustment coefficient from the previous time step to the correction amount.

[0216] Multiply the current input power by the adjustment coefficient to obtain the current available power, and calculate the maximum output current;

[0217] The target charging current is compared with the maximum output current. If the target charging current is greater than the maximum output current, the maximum output current is used as the target charging current.

[0218] Specifically, the current charging stage remains unchanged during the adjustment of the target charging current.

[0219] As an optional implementation, a flag is set to prevent the stage switching logic from being triggered by current changes. The charging stage state is only allowed to be updated when preset safety conditions or stage switching conditions are met (such as voltage reaching a threshold, current falling below a threshold, etc.).

[0220] In practice, if the current input condition level of the industrial charging equipment is the second condition level, while the previous time was the first condition level, it indicates that the input condition level has increased and the derating is performed.

[0221] The controller invokes the incremental proportional-integral algorithm to calculate the correction amount for the current period as negative, based on the deviation between the disturbance estimate and the baseline value.

[0222] Assuming the depreciation factor stored in the previous time step was 0.9, after adding the correction amount to the previous depreciation factor, the current depreciation factor is 0.7.

[0223] Based on the current input power of 500 watts, multiplying the input power by a derating factor of 0.7 yields a usable power of 350 watts; the currently sampled output voltage is 50 volts, and the calculated maximum output current is 7 amps. The current charging stage is the constant current charging stage, and the stage lock flag is currently zero, indicating that stage switching is allowed.

[0224] Upon entering the current regulation process, the controller sets the stage lock flag to lock the current stage state. The target charging current for this stage was originally 30 amps. The target charging current of 30 amps is compared with the maximum output current of 7 amps. Since the target charging current is greater than the maximum output current, the target charging current is adjusted to 28 amps and output to the current loop.

[0225] In practice, the stage lock flag remains set, the controller will not perform stage switching judgment, and the charging stage will remain a constant current charging stage.

[0226] Regarding step S105:

[0227] When the input voltage rises back to the preset fluctuation threshold, the target charging current is restored according to the preset recovery slope, and the control loop is limited during the recovery process.

[0228] In practice, the controller compares the amplitude deviation of the current input voltage with a preset amplitude deviation threshold within each control cycle, and simultaneously compares the current rate of change with a preset rate of change threshold. When both the amplitude deviation and the rate of change are below the amplitude deviation threshold and the rate of change threshold, the input voltage is determined to have returned to normal, and the fluctuation state is exited. At this point, the controller records the current moment as the recovery start point and begins executing the recovery ramp algorithm.

[0229] The cumulative calorific value is obtained by integrating the square of the target charging current over time using the conditional integral anti-saturation method, and a first correction coefficient is determined.

[0230] As an alternative implementation, a thermal accumulation model is introduced to adjust the recovery rate, and the thermal effect of power devices is simulated by using a first-order inertial element.

[0231] The recursive algorithm for the cumulative calorific value includes: calculating the square of the target charging current in each sampling period, and then updating the cumulative calorific value according to the following recursive relationship: the current cumulative calorific value is equal to the cumulative calorific value of the previous period multiplied by the thermal decay coefficient, plus the square of the current output current multiplied by the sampling period, as the first correction coefficient.

[0232] The thermal decay coefficient is determined based on the thermal time constant of the radiator. The calculation formula is: thermal decay coefficient equals the negative sampling period of the natural constant divided by the power of the thermal time constant. This coefficient ranges from 0 to 1; the larger the thermal time constant, the slower the decay.

[0233] The upper limit of the cumulative calorific value is calculated by dividing the maximum allowable temperature rise of the power device by the thermal resistance and then multiplying by the thermal time constant. When the cumulative calorific value reaches the upper limit, further accumulation stops, and the value remains at the upper limit. The lower limit of the cumulative calorific value is zero.

[0234] The first correction coefficient is multiplied by the preset recovery step size to obtain the second step size value, and then increased cycle by cycle based on the preset charging cycle to obtain the current value to be processed.

[0235] As an optional implementation, the recovery step size adjustment algorithm adopts proportional control, that is, the current recovery step size is equal to the basic recovery step size multiplied by the thermal constraint coefficient.

[0236] The thermal constraint coefficient is inversely proportional to the cumulative calorific value. The thermal constraint coefficient is equal to a power of a sum minus the cumulative calorific value divided by the upper limit. The power is determined based on the degree of nonlinearity of the device's thermal characteristics, and is typically 1 or 2.

[0237] When the cumulative calorific value is zero, the thermal constraint coefficient is 1, and the recovery speed is the fastest; when the cumulative calorific value approaches the upper limit, the thermal constraint coefficient approaches zero, and the recovery speed slows down significantly.

[0238] The recovery ramp is generated in a cycle-by-cycle increment. In each recovery cycle, the target charging current of the previous cycle is added to the recovery step size of the current cycle to obtain the current value to be processed. The target charging current at the start of recovery is the current value when exiting the fluctuation state.

[0239] The ratio of the current available power to the current output voltage is obtained as the upper limit of the current, and the current value to be processed is compared with the upper limit of the current, and the smaller value between the two is taken as the target charging current of the current cycle.

[0240] During the recovery process, the integral term is constrained according to the integral accumulation condition.

[0241] As an optional implementation, if the target charging current of the current cycle reaches the upper limit of the current and the current control error direction is the same as the integral accumulation direction, the accumulation of the integral term is stopped.

[0242] The current control error direction is determined by the sign of the error value in the current cycle, and the integral accumulation direction is determined by the sign of the difference between the integral value in the current cycle and the integral value in the previous cycle.

[0243] In practice, the controller calculates the following three parameters for each control cycle:

[0244] First, the target charging current is reached. The current target charging current is compared with the upper limit of the current. If the absolute value of the difference between the two is less than a preset proximity threshold, the target charging current is considered to have reached the upper limit. The proximity threshold is usually set to one percent of the upper limit of the current.

[0245] Second, the direction of the error. Calculate the deviation between the target charging current and the actual output current. If the deviation is greater than zero, that is, the actual output current is less than the target charging current, then the error direction is positive; if the deviation is less than zero, then the error direction is negative.

[0246] Third, the direction of integration accumulation. Calculate the difference between the integral value of the current period and the integral value of the previous period. If the difference is greater than zero, it means that the integral term is increasing and the direction of integration accumulation is positive; if the difference is less than zero, it means that the integral term is decreasing and the direction of integration accumulation is negative.

[0247] The condition for enabling integral accumulation is as follows: when the target charging current reaches the true value and the error direction is the same as the integral accumulation direction, integral accumulation is prohibited, that is, the integral value remains unchanged; otherwise, the integral is accumulated normally.

[0248] In practice, the current input operating condition level is Level 2, with a lumped disturbance estimate of -23.7 amps. After disturbance and temperature corrections, the final constraint parameters are: AC current limiting threshold of 2.85 amps, power ramp-up rate upper limit of 142.5 watts per second, target bus voltage of 90.25 volts, and equivalent power command upper limit of 570 watts. The current charging stage is constant current charging, and the target charging current has been adjusted to 4 amps.

[0249] At this point, the input voltage begins to rise, and the controller continuously monitors the fluctuation characteristics of the input voltage. After several control cycles, assuming the input voltage amplitude deviation drops to 15 volts and the rate of change drops to 15 volts per second, both below the fluctuation judgment thresholds of 19 volts and 19 volts per second, the controller determines that the input voltage has returned to normal and exits the fluctuation state. The controller records the current moment as the recovery start point and initiates the recovery ramp process.

[0250] The parameters of the thermal accumulation model were set to a sampling period of 100 microseconds and a radiator thermal time constant of 300 seconds, and the calculated thermal decay coefficient was 0.99999967.

[0251] The upper limit of the cumulative calorific value is calculated to be 1000 ampere-seconds based on the parameters of the power switching tube. The formula for calculating the thermal constraint coefficient adopts a linear relationship, that is, the thermal constraint coefficient equals 1 minus the cumulative calorific value divided by the upper limit value.

[0252] The current cumulative calorific value is 350 amperes squared seconds. According to the recursive algorithm, the current cumulative calorific value is equal to the previous cumulative calorific value multiplied by 0.99999967, plus the square of the current current 2.85 (8.1225) multiplied by 0.0001 seconds. The calculated updated cumulative calorific value is still approximately 350.

[0253] The thermal constraint factor equals 1 minus 350 divided by 1000, which equals 0.65. The basic recovery step size is set to 0.00005 amperes per cycle, multiplied by the thermal constraint factor of 0.65, resulting in a current recovery step size of 0.0000325 amperes per cycle.

[0254] The target charging current at the start of recovery is 4 amps. In the first recovery cycle, the target charging current of the previous cycle of 1.5 amps is added to the recovery step size of 0.0000325 amps to obtain the current value to be processed, which is 1.5000325 amps.

[0255] The current input power is 500 watts, multiplied by the current adjustment factor of 0.7, yielding a usable power of 350 watts. The current sampled output voltage is 50 volts, and the calculated upper limit of the current is 350 divided by 50, which equals 7 amps. The approach threshold is set to one-hundredth of 7 amps, i.e., 0.07 amps. It is determined that the target charging current has not reached the upper limit.

[0256] The current value to be processed, 1.5000325 amps, is compared with the upper limit of current, 7 amps, and the smaller value, 1.5000325 amps, is taken as the target charging current for the current cycle.

[0257] As the recovery process progresses, the accumulation is repeated each cycle, assuming the target charging current has risen to 6.9 amps. At this point, the current sampled output voltage is still 50 volts, the available power is still 350 watts, and the current limit is still 7 amps.

[0258] The difference between the target charging current of 6.9 amps and 7 amps is 0.1 amps, which is greater than the threshold of 0.07 amps. Therefore, the target charging current is reached and the flag is false, and the integral is accumulated normally.

[0259] When the target charging current rises to 6.95 amps, the difference between it and 7 amps is 0.05 amps, which is less than 0.07 amps, indicating that the target charging current has been reached. The current actual output current is 6.94 amps, with a deviation of 0.01 amps, and the error direction is positive. The current cycle integral value is 180, and the previous cycle integral value was 178, with the integral accumulation direction being positive. Since the error direction is the same as the integral accumulation direction, both being positive, the integral accumulation enable condition is not met, so integral accumulation is paused, and the integral value remains unchanged at 180.

[0260] When the actual output current rises to 6.99 amps and the deviation is -0.04 amps, the error direction becomes negative, opposite to the direction of integral accumulation. At this point, integral accumulation is resumed.

[0261] See Figure 3 The flowchart for determining the charging phase lockout using a wide voltage input method provided in this application embodiment includes:

[0262] As an optional implementation, a sequential quadratic programming-gradient descent algorithm is adopted to construct a cost function based on the deviation between the current output voltage and the target charging voltage and the deviation between the current output current and the target charging current, with the charging stage state identifier as the variable to be optimized.

[0263] The optimal solution of the cost function is found by using the restriction of switching between charging stages during the reduction of the target charging current or the gradual increase of the target charging current as an optimization constraint.

[0264] In practice, the algorithm's inputs include: current charging stage identifier, target charging voltage, target charging current, actual output voltage sample value, actual output current sample value, and a flag indicating whether it is currently in derating or scaling adjustment.

[0265] The algorithm outputs an optimized charging stage identifier, which is used to determine whether to keep the current stage unchanged or allow stage switching.

[0266] The cost function consists of a weighted sum of two error terms. The first error term is the absolute value of the deviation between the output voltage and the target charging voltage, reflecting the voltage tracking accuracy. The second error term is the absolute value of the deviation between the output current and the target charging current, reflecting the current tracking accuracy. The two error terms are multiplied by preset weighting coefficients and then summed to form the total cost function value.

[0267] The weighting coefficients are set differently depending on the charging stage. For example, the current deviation has a higher weight in the constant current charging stage, while the voltage deviation has a higher weight in the constant voltage charging stage.

[0268] In specific implementation, the variable to be optimized is the charging stage status identifier, which is usually represented by discrete integers. For example, 0 represents the pre-charging stage, 1 represents the constant current charging stage, 2 represents the constant voltage charging stage, and 3 represents the float charging stage.

[0269] As an optional implementation, the algorithm introduces a hard constraint that prohibits stage switching during the derating or scaling adjustment. When the controller detects that it is currently in the derating or scaling adjustment state, a penalty function term is added to the optimization problem. This penalty function term takes a maximum value when the variable to be optimized deviates from the current stage identifier, thereby forcing the optimization result to be locked at the current stage.

[0270] When the adjustment state ends, the penalty function term is removed, allowing phase switching.

[0271] In its implementation, the algorithm employs a sequential quadratic programming framework for iterative solution. At each iteration step, a quadratic Taylor expansion of the cost function is first performed at the current point to construct a quadratic programming subproblem. The objective function of this subproblem is a quadratic form, and the constraints are the linearized original constraints.

[0272] After solving the quadratic programming subproblem to obtain the search direction, the gradient descent method is used to perform a one-dimensional line search along that direction to determine the optimal step size and update the variable to be optimized.

[0273] In practice, the Hessian matrix correction parameter in the quadratic programming is set to a small positive number. The initial step size of gradient descent is set to 0.1, the step size decay factor is set to 0.5, the convergence threshold is set to 0.001, and the maximum number of iterations is set to twenty.

[0274] The penalty function term is set to a value much larger than the normal cost function during the decrement or escalation adjustment period. For example, if the normal cost function value is between 0 and 10, the penalty function term is set to 1000 to ensure that the optimization result is forcibly locked in the current stage.

[0275] The iteration terminates when the change in cost function value between two consecutive iterations is less than a preset convergence threshold, or the change in the variable to be optimized is less than a preset threshold, or the maximum number of iterations is reached. After the iteration ends, the optimized continuous variables are rounded to obtain the final charging stage identifier.

[0276] The optimized charging stage identifier is compared with the currently locked charging stage identifier. If they match, the current charging stage remains unchanged and the lock continues. If they do not match, it is further determined whether the preset optimization convergence conditions are met. The convergence conditions include: the cost function value is lower than the preset convergence threshold, the optimization iteration has converged, and the current state is not in derating or maximizing adjustment. When these conditions are met simultaneously, the stage lock is released, allowing a switch to a new charging stage; otherwise, even if the optimization results are different, the current charging stage remains unchanged.

[0277] In practical implementation, assuming the current input operating condition level is level two, the final constraint parameters after disturbance and temperature correction are as follows: AC current limiting threshold 2.85 amps, power ramp-up rate upper limit 142.5 watts per second, target bus voltage 90.25 volts, and equivalent power command upper limit 570 watts. The current charging stage is constant current charging, the target charging current has been adjusted to 4 amps, and the target charging voltage is a constant voltage conversion voltage of 57.6 volts. The actual output voltage sampling value is 50 volts, and the actual output current sampling value is 1.48 amps. Currently, it is in derating adjustment state (due to disturbance), and the derating flag is true.

[0278] The cost function of the algorithm is constructed, with the current deviation weight set to 5 and the voltage deviation weight set to 1 during the constant current charging stage.

[0279] The current deviation is 0.02 amperes; the voltage deviation is 7.6 volts, and the cost function value is 7.7.

[0280] The variable to be optimized is the charging stage identifier, with a current value of 1 indicating the constant current charging stage; the optimization variable relaxation is a continuous variable with a value range of 0 to 3.

[0281] Since the deregulation flag is true, a penalty function term is introduced into the constraints, with a penalty coefficient set to 1000. When the optimization variable deviates from its current value of 1, the penalty function term is 1000 multiplied by the square of the deviation.

[0282] The sequential quadratic programming iteration begins. In the first iteration, a quadratic programming subproblem is constructed at the current point 1, and the solution yields a search direction of -0.001. The gradient descent method searches along this direction with a step size of 0.1, resulting in a new point 0.9999.

[0283] The cost function value is calculated by adding the penalty function term. The total cost is 7.7 plus 1000 multiplied by the square of 0.0001, which is still approximately 7.7. The second iteration yields a point of 0.9998, with the total cost remaining essentially unchanged. After five iterations, the variable stabilizes at 0.9995, and the cost function value changes by less than 0.001, indicating convergence.

[0284] The optimized continuous variable is 0.9995, which, after rounding, equals 1, consistent with the currently locked constant current charging stage identifier. Therefore, the algorithm determines to keep the current charging stage unchanged, and the stage lock flag remains set.

[0285] Based on the same inventive concept, this application also provides a wide voltage input industrial environment adaptive charging system for use with the wide voltage input industrial environment adaptive charging method.

[0286] See Figure 4 The diagram shown illustrates an industrial environment adaptive charging system with wide voltage input, as provided in an embodiment of this application. The system includes: an identification module 10, a constraint module 20, a target determination module 30, an adjustment module 40, and a recovery module 50, wherein:

[0287] Identification module 10: Real-time acquisition of operating parameters including input voltage, analysis of fluctuation characteristics of the input voltage, and classification of the current input operating condition level;

[0288] Constraint module 20: Based on the current input operating condition level, it generates an input-side constraint parameter set that includes at least an AC current limiting threshold and a power ramp-up rate upper limit through a preset constraint parameter mapping relationship, and applies the input-side constraint parameter set to a preset control loop for constraint to obtain the input power;

[0289] Target determination module 30: Determines the current charging stage and obtains the corresponding target charging current and target charging voltage;

[0290] Adjustment module 40: compares the current input operating condition level with the operating condition level at the previous moment; if the operating condition level changes, it calculates the current available power based on the input power and adjusts the target charging current.

[0291] Recovery module 50: When the input voltage rises back to the preset fluctuation judgment threshold, the target charging current is restored according to the preset recovery slope, and the control loop is limited during the recovery process.

[0292] Those skilled in the art will understand that, in the above-described method of the specific implementation, the order in which each step is written does not imply a strict execution order and does not constitute any limitation on the implementation process. The specific execution order of each step should be determined by its function and possible internal logic.

[0293] In the description of this specification, references to terms such as "exemplary," "for example," and "specifically" indicate that a particular feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

Claims

1. An adaptive charging method for industrial environments with wide voltage input, characterized in that, The method includes: Real-time acquisition of operating parameters, including input voltage, analysis of the fluctuation characteristics of the input voltage, and classification of the current input operating condition level; Based on the current input operating condition level, an extended state observer is constructed to estimate the disturbance in the input voltage fluctuation process in real time, and the disturbance estimate is calculated. Based on the disturbance estimate, an input-side constraint parameter set including at least an AC current limiting threshold and a power ramp-up rate upper limit is generated through a preset constraint parameter mapping relationship. The input-side constraint parameter set is then applied to a preset control loop for constraint to obtain the input power. Determine the current charging stage to obtain the corresponding target charging current and target charging voltage; The current input operating condition level is compared with the operating condition level at the previous moment. If the operating condition level changes, based on the disturbance estimate, a correction amount of the adjustment coefficient of the current input power is calculated using a preset proportional-integral algorithm. The adjustment coefficient is used to characterize the utilization ratio of the current input power. The adjustment coefficient at the previous moment is added to the correction amount to obtain the current adjustment coefficient. The current input power is multiplied by the current adjustment coefficient to obtain the current available power. The maximum output current is calculated based on the ratio of the current available power to the current output voltage, and the target charging current is adjusted accordingly. When the input voltage rises back to the preset fluctuation threshold, the target charging current is restored according to the preset recovery slope, and the control loop is limited during the recovery process.

2. The industrial environment adaptive charging method with wide voltage input according to claim 1, characterized in that, The fluctuation characteristic analysis of the input voltage includes: Set fluctuation judgment thresholds, which include amplitude deviation thresholds and rate of change thresholds; If the amplitude deviation of the input voltage is greater than or equal to the amplitude deviation threshold, and the rate of change of the input voltage is greater than or equal to the rate of change threshold, the input voltage is determined to be in a fluctuating state. The magnitude deviation of the input voltage is obtained by calculating the absolute value of the difference between the input voltage and the standard voltage. Based on the amplitude deviation and the rate of change, the fluctuation level corresponding to the fluctuation state is classified as the current input operating condition level.

3. The industrial environment adaptive charging method with wide voltage input according to claim 2, characterized in that, The fluctuation level corresponding to the fluctuation state is used as the current input operating condition level. include: The input operating condition level includes at least a first operating condition level, a second operating condition level, and a third operating condition level, wherein the input operating condition level increases sequentially according to the degree of fluctuation. The first operating condition level corresponds to an input voltage that is in a fluctuating condition where the amplitude deviation is less than or equal to the first amplitude threshold and the rate of change is less than or equal to the first rate threshold. The second operating condition level corresponds to an input voltage that is in a fluctuating condition where the amplitude deviation is greater than the first amplitude threshold but less than or equal to the second amplitude threshold and the rate of change is greater than the first rate threshold but less than or equal to the second rate threshold. The third operating condition level corresponds to an input voltage that is in a fluctuating condition where the amplitude deviation is greater than the second amplitude threshold and the rate of change is greater than the second rate threshold. Among them, the first amplitude threshold is less than the second amplitude threshold, and the first rate threshold is less than the second rate threshold.

4. The industrial environment adaptive charging method with wide voltage input according to claim 1, characterized in that, The real-time estimation of disturbances during input voltage fluctuations by constructing an extended state observer includes: The input voltage is used as the input to the extended state observer, and the lumped disturbance is extended into the state variable of the extended state observer. The lumped disturbances include: input voltage fluctuations, line impedance changes, and load abrupt changes; The observer gain matrix of the extended state observer is calculated based on the current input operating condition level, and the lumped disturbance is estimated, and the disturbance estimate is output. Determine the AC current limiting threshold and power ramp-up rate baseline values ​​based on the current input operating condition level; The absolute value of the disturbance estimate is compared with the preset rated disturbance value. The disturbance degree coefficient is determined based on the comparison result, and the correction factor is determined based on the disturbance degree coefficient. It also obtains the current temperature and determines the temperature correction coefficient through fuzzy logic reasoning; The AC current limiting threshold and the power ramp-up rate benchmark are multiplied by the correction factor and the temperature correction coefficient, respectively, to calculate the AC current limiting threshold and the upper limit of the power ramp-up rate, which are then used as the set of input-side constraint parameters.

5. The industrial environment adaptive charging method with wide voltage input according to claim 4, characterized in that, The calculation of the observer gain matrix of the extended state observer includes: The observer bandwidth of the extended state observer is determined based on the current input condition level, wherein the observer bandwidth increases as the current input condition level increases; The observer bandwidth is substituted into the pole configuration calculation, and an observer gain matrix corresponding to the observer bandwidth is generated by a preset gain determination rule.

6. The industrial environment adaptive charging method with wide voltage input according to claim 1, characterized in that, The correction amount for calculating the adjustment coefficient of the current input power using a preset proportional-integral algorithm includes: When the current input condition level increases relative to the previous moment, the adjustment coefficient is used as a deregulation coefficient, and a negative correction amount is calculated based on the disturbance estimate using an incremental proportional-integral algorithm. When the current input condition level decreases relative to the previous moment, the adjustment coefficient is used as an increase coefficient, and a positive correction amount is calculated based on the disturbance estimate using an incremental proportional-integral algorithm. The correction amount includes a proportional correction amount and an integral correction amount; The target charging current is compared with the maximum output current. If the target charging current is greater than the maximum output current, the maximum output current is taken as the target charging current. Specifically, the current charging stage remains unchanged during the adjustment of the target charging current.

7. The industrial environment adaptive charging method with wide voltage input according to claim 1, characterized in that, The step of restoring the target charging current according to the preset recovery slope includes: The cumulative calorific value is obtained by integrating the square of the target charging current over time using the conditional integral anti-saturation method, and a first correction coefficient is determined. The first correction coefficient is multiplied by the preset recovery step size to obtain the second step size value, and then increased cycle by cycle based on the preset charging cycle to obtain the current value to be processed. The ratio of the current available power to the current output voltage is obtained as the upper limit of the current, and the current value to be processed is compared with the upper limit of the current, and the smaller value of the two is taken as the target charging current of the current cycle. During the recovery process, the integral term is constrained according to the integral accumulation condition.

8. The industrial environment adaptive charging method with wide voltage input according to claim 7, characterized in that, The constraint on the integral term based on the integral accumulation condition during the recovery process includes: If the target charging current of the current cycle reaches the upper limit of the current and the current control error direction is the same as the integral accumulation direction, the accumulation of the integral term is stopped. The current control error direction is determined by the sign of the error value in the current cycle, and the integral accumulation direction is determined by the sign of the difference between the integral value in the current cycle and the integral value in the previous cycle.

9. The industrial environment adaptive charging method with wide voltage input according to claim 6, characterized in that, Maintaining the current charging stage unchanged during the target charging current adjustment process includes: A preset gradient descent algorithm is used to construct a cost function based on the deviation between the current output voltage and the target charging voltage, and the deviation between the current output current and the target charging current, with the charging stage status identifier as the variable to be optimized. The optimal solution of the cost function is obtained by using the restriction of switching between charging stages during the reduction of the target charging current or the gradual increase of the target charging current as an optimization constraint. If the optimal solution obtained by solving is consistent with the current locked charging stage state flag, the current charging stage remains unchanged; If the optimal solution is inconsistent with the currently locked charging stage status identifier and the preset optimization convergence condition is met, the locking of the charging stage status identifier is released.

10. A wide-voltage input industrial environment adaptive charging system, used to implement the wide-voltage input industrial environment adaptive charging method according to any one of claims 1-9, characterized in that, The system includes: Identification module: Real-time acquisition of operating parameters including input voltage, analysis of fluctuation characteristics of the input voltage, and classification of the current input operating condition level; Constraint Module: Based on the current input operating condition level, an extended state observer is constructed to estimate the disturbance in the input voltage fluctuation process in real time, calculate the disturbance estimate value, and based on the disturbance estimate value, an input-side constraint parameter set including at least an AC current limiting threshold and a power ramp-up rate upper limit is generated through a preset constraint parameter mapping relationship, and the input-side constraint parameter set is applied to a preset control loop for constraint to obtain the input power; Target determination module: Determines the current charging stage and obtains the corresponding target charging current and target charging voltage; Adjustment module: Compares the current input operating condition level with the operating condition level at the previous moment. When the operating condition level changes, based on the disturbance estimate, it calculates the correction amount of the adjustment coefficient of the current input power using a preset proportional-integral algorithm. The adjustment coefficient is used to characterize the utilization ratio of the current input power. The adjustment coefficient at the previous moment is added to the correction amount to obtain the current adjustment coefficient. The current input power is multiplied by the current adjustment coefficient to obtain the current available power. The maximum output current is calculated based on the ratio of the current available power to the current output voltage, and the target charging current is adjusted accordingly. Recovery module: When the input voltage rises back to the preset fluctuation judgment threshold, the target charging current is restored according to the preset recovery slope, and the control loop is limited during the recovery process.

Images