A double closed loop boost circuit control method and system based on fuzzy neural network

Abstract

The application provides a double closed loop boost circuit control method and system based on a fuzzy neural network, wherein the method comprises the following steps: sampling the output voltage and inductance current of a boost circuit in real time, and calculating voltage error and current error according to a preset reference voltage and reference current; taking the voltage error and its change rate as input, and adopting a fuzzy neural network to perform joint reasoning and weighted calculation, and outputting a parameter correction amount, and correcting the outer loop PID control parameter through the parameter correction amount to obtain a corrected fuzzy neural network; adopting the corrected fuzzy neural network, updating the reference current according to the voltage error, and obtaining a duty cycle adjustment amount according to the current PI control algorithm based on the current error; generating a PWM control signal according to the duty cycle adjustment amount, and generating a target high voltage based on the PWM control signal. The application can realize fast response, stable regulation and safe operation of the boost circuit.

Description

Technical Field

[0001] This invention relates to the field of circuit control technology, and in particular to a dual closed-loop boost circuit control method and system based on fuzzy neural networks. Background Technology

[0002] With the development of electronic technology and power electronic devices, boost circuits have been widely used in high-voltage energy storage, pulsed power supplies, electronic fuse energy management, laser drives, and new energy energy conversion systems due to their advantages such as simple structure, strong boost capability, and flexible control methods. These circuits achieve energy conversion from low input voltage to high input voltage by periodically controlling the on and off states of power switching devices. To ensure the stability of the output voltage under different input voltage and load conditions, boost circuits typically employ a closed-loop feedback control strategy.

[0003] Existing boost circuits mostly employ single-loop PI or PID control algorithms. These control methods are simple in structure, mature in engineering implementation, and can achieve good control performance near rated operating conditions. However, because the control parameters are fixed values, their performance is highly dependent on the initial tuning results. When the input voltage fluctuates, the load changes abruptly, or the system parameters change due to environmental factors and device aging, traditional PID controllers struggle to simultaneously achieve fast dynamic response and steady-state accuracy, easily leading to problems such as output voltage overshoot, response hysteresis, and large steady-state errors. Especially in applications such as high-voltage energy storage and electronic fuses, the system has stringent requirements for boost speed, timely overcurrent protection, and operational safety; traditional single-loop control structures cannot meet the demands of both high dynamic performance and high safety.

[0004] With the development of intelligent control technology in recent years, the advantages of fuzzy control and neural network control in the regulation of nonlinear systems have gradually become apparent. Fuzzy control does not rely on precise mathematical models and can adjust the system state using empirical rules; neural networks have self-learning and adaptive capabilities, and can approximate complex nonlinear mapping relationships. The fuzzy neural network (FNN) control method, which combines the two, can dynamically adjust control parameters during system operation, thereby achieving a better trade-off between steady-state accuracy and dynamic response. However, if only intelligent control methods based on software platforms such as MCUs are used, their computing power and real-time performance are insufficient to support the requirements of high-frequency PWM control and dual closed-loop fast feedback, limiting the engineering application of intelligent control algorithms in high-voltage boost systems.

[0005] Therefore, there is an urgent need for a dual-closed-loop boost circuit control method that can achieve fast response, stable regulation, and safe operation of boost circuits based on fuzzy neural networks. Summary of the Invention

[0006] Therefore, it is necessary to provide a control method and system for a dual closed-loop boost circuit based on a fuzzy neural network to address the aforementioned technical problems.

[0007] A dual-closed-loop boost circuit control method based on fuzzy neural network includes the following steps: sampling the output voltage and inductor current of the boost circuit in real time, and calculating voltage error and current error according to preset reference voltage and reference current; using the voltage error and its rate of change as input, performing joint inference and weighted calculation using fuzzy neural network to output parameter correction amount, and correcting the outer loop PID control parameters through the parameter correction amount to obtain a corrected fuzzy neural network; using the corrected fuzzy neural network to update the reference current according to the voltage error, and obtaining the duty cycle adjustment amount based on the current PI control algorithm according to the current error; generating a PWM control signal according to the duty cycle adjustment amount, and generating a target high voltage based on the PWM control signal.

[0008] In one embodiment, the real-time sampling of the output voltage and inductor current of the boost circuit, and the calculation of voltage and current errors based on preset reference voltage and reference current, includes: acquiring the output voltage of the boost circuit through a voltage sampling circuit and a current sampling circuit, respectively. and inductor current and output voltage With the preset reference voltage By comparison, the voltage error is obtained. : ; inductor current Reference current given by the outer voltage loop By comparison, the current error is calculated. : .

[0009] In one embodiment, the fuzzy neural network performs joint reasoning and weighted calculation using fuzzy membership functions, fuzzy rule bases, and neural network connection weights.

[0010] In one embodiment, the current PI control algorithm employs a proportional-integral algorithm, and its duty cycle is recursively calculated according to the following formula: ; In the formula, u(k) The duty cycle of the output. u(k- 1 ) This represents the duty cycle of the previous cycle. This is the proportionality coefficient. The integral coefficient is... e i This is for current error.

[0011] In one embodiment, after sampling the output voltage and inductor current of the boost circuit in real time, the method further includes: monitoring the inductor current in real time, and outputting a PWM shutdown signal when the inductor current exceeds a preset threshold to forcibly shut down the main power switching device.

[0012] A dual-closed-loop boost circuit control system based on a fuzzy neural network is disclosed to implement the aforementioned dual-closed-loop boost circuit control method based on a fuzzy neural network. The system includes: a sampling feedback module, a voltage outer loop control module, a current inner loop control module, a PWM drive module, a power conversion module, an FPGA main control module, and a host computer monitoring module. The sampling feedback module is connected to the power conversion module and the FPGA main control module to acquire the output voltage and inductor current of the boost circuit and input them to the FPGA main control module. The FPGA main control module calculates the voltage error and current error based on preset reference voltage and reference current. The voltage outer loop control module is connected to the FPGA main control module and adjusts the output reference current according to the voltage error. The current inner loop control module is connected to the voltage outer loop control module and adjusts the output reference current according to the duty cycle. The duty cycle is adjusted by the ratio of the adjustment amount, and the adjusted duty cycle is input to the PWM drive module. The PWM drive module is connected to the current inner loop control module and is used to generate a PWM control signal based on the adjusted duty cycle. The power conversion module is used to boost the output voltage to the target high voltage based on the PWM control signal. The FPGA main control module is used to perform joint inference and weighted calculation using a fuzzy neural network based on the voltage error and its rate of change, and output parameter correction amount. The parameter correction amount is used to correct the outer loop PID control parameters to obtain the corrected fuzzy neural network. The duty cycle adjustment amount is obtained based on the current error using a current-based PI control algorithm. The on-machine monitoring module is connected to the FPGA main control module and is used to send control commands, configure fuzzy neural network parameters, display real-time voltage waveforms, and record data.

[0013] In one embodiment, the power conversion module adopts a flyback boost topology, including a main power switching device MOSFET, a flyback boost transformer, and a voltage doubler rectifier circuit.

[0014] In one embodiment, the PWM drive module uses a gate drive chip to drive the main power switching device MOSFET, thereby realizing high-frequency turn-on and turn-off control of the primary side of the flyback transformer.

[0015] In one embodiment, it further includes a hardware fast-break protection module, which is connected to the sampling feedback module and the PWM drive module, and is used to directly generate a shutdown signal and forcibly shut down the main power switch when the inductor current exceeds a preset threshold.

[0016] In one embodiment, the hardware fast-break protection module includes a high-speed comparator and a digital isolator. The high-speed comparator is used to monitor the current sampling signal in real time and determine whether the sampled current exceeds a set threshold. The digital isolator is used to directly output a PWM shutdown signal to the PWM drive module when the sampled current exceeds the set threshold.

[0017] Compared with existing technologies, the advantages and beneficial effects of this invention are as follows: By sampling the output voltage and inductor current of the boost circuit in real time, and calculating the voltage error and current error based on the preset reference voltage and reference current, high-precision measurement under high-voltage conditions is achieved; using the voltage error and its rate of change as input, a fuzzy neural network is used for joint reasoning and weighted calculation of the output parameter correction amount, which is used to correct the outer loop PID control parameters, resulting in a corrected fuzzy neural network, thereby achieving adaptive adjustment of the boost process; inputting the voltage error into the corrected fuzzy neural network, an updated reference current is obtained, and based on the current PID control algorithm, the duty cycle adjustment amount is obtained according to the current error, realizing dual closed-loop control of voltage and current, which can achieve stable adjustment of the boost process under complex operating conditions; generating a PWM control signal based on the duty cycle adjustment amount, and adjusting the output voltage to the target high voltage based on the PWM control signal, the dynamic performance can be maintained under conditions of input voltage fluctuation, load change and parameter uncertainty, realizing the rapid response, stable adjustment and safe operation of the boost circuit, which is suitable for high-voltage energy storage and electronic fuse energy management applications with high safety and practicality requirements. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating a dual-closed-loop boost circuit control method based on a fuzzy neural network in one embodiment.

[0019] Figure 2 This is a schematic diagram of a dual-closed-loop boost circuit control system based on a fuzzy neural network in one embodiment.

[0020] Figure 3 This is a schematic diagram illustrating the control principle of a dual closed-loop boost circuit control system based on a fuzzy neural network in one embodiment. Detailed Implementation

[0021] Before describing the specific embodiments of the present invention, the overall concept of the present invention will be explained as follows: This invention is mainly based on the development of boost circuit control process. At present, boost circuit control methods have defects such as slow response speed, large overshoot, poor steady-state accuracy and insufficient anti-disturbance capability.

[0022] Therefore, this invention proposes a dual-closed-loop boost circuit control method based on fuzzy neural networks. By sampling the output voltage and inductor current of the boost circuit in real time, and calculating voltage and current errors based on preset reference voltage and current, high-precision measurement under high-voltage conditions is achieved. Using the voltage error and its rate of change as input, a fuzzy neural network is used for joint inference and weighted calculation of the output parameter correction amount, which is then used to correct the outer-loop PID control parameters, resulting in a corrected fuzzy neural network, thereby achieving adaptive adjustment of the boost process. The voltage error is input into the corrected fuzzy neural network to obtain an updated reference current, and based on a current PID control algorithm, the duty cycle adjustment amount is obtained according to the current error, achieving dual-closed-loop control of voltage and current. This enables stable adjustment of the boost process under complex operating conditions. A PWM control signal is generated based on the duty cycle adjustment amount, and the output voltage is adjusted to the target high voltage based on the PWM control signal. This method maintains dynamic performance under conditions of input voltage fluctuations, load changes, and parameter uncertainties, achieving rapid response, stable adjustment, and safe operation of the boost circuit. It is suitable for high-voltage energy storage and electronic fuse energy management applications where safety and practicality are paramount.

[0023] Having introduced the overall concept of the present invention, to make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below through specific embodiments in conjunction with the accompanying drawings. It should be understood that the specific embodiments described herein are merely illustrative of the present invention and are not intended to limit the present invention.

[0024] In one embodiment, such as Figure 1 As shown, a dual-closed-loop boost circuit control method based on a fuzzy neural network is provided, including the following steps: Step S110: Sample the output voltage and inductor current of the boost circuit, and calculate the voltage error and current error based on the preset reference voltage and reference current.

[0025] Specifically, the output voltage and inductor current of the boost circuit are acquired in real time through the sampling circuit, and the difference between the two is calculated based on the preset reference voltage and reference current to obtain the voltage error and current error, which are used for subsequent control of the voltage outer loop and current inner loop.

[0026] The output voltage of the boost circuit is obtained through a voltage sampling circuit and a current sampling circuit, respectively. and inductor current and output voltage With the preset reference voltage By comparison, the voltage error is obtained. : ; inductor current Reference current given by the outer voltage loop By comparison, the current error is calculated. : .

[0027] Specifically, an isolated ADC is used to set up voltage sampling circuits and current sampling circuits to sample the output voltage and inductor current of the boost circuit, respectively, so as to achieve safe isolation and high-precision measurement under high-voltage environment.

[0028] The output voltage is compared with the preset reference voltage to calculate the voltage error. Combined with the voltage error calculated in the previous cycle, the voltage error conversion rate is obtained and used for subsequent voltage outer loop control. The sampled inductor current is compared with the reference current adjusted according to the voltage outer loop in the previous cycle to obtain the current error, which is used for subsequent current inner loop control.

[0029] The reference current is calculated by the PID controller after self-tuning in the previous cycle based on the voltage error signal, and is used as the setpoint for the inner loop control, thereby enabling the outer loop to have online learning and adaptive adjustment capabilities, significantly improving the system's voltage regulation accuracy and dynamic performance.

[0030] In one embodiment, step S110 further includes: monitoring the inductor current in real time, and outputting a PWM shutdown signal when the inductor current exceeds a preset threshold to forcibly shut down the main power switching device.

[0031] Specifically, a hardware fast-break protection mechanism is set in the current sampling circuit to monitor the acquired inductor current in real time. When the high-speed comparator detects that the current exceeds a preset threshold, In this case, the software control path is bypassed, and the shutdown signal is directly output through the digital isolator to force the main power switch to shut down, thereby achieving sub-microsecond overcurrent protection, avoiding the risk of device damage caused by control algorithm delay, and realizing hardware-level fast protection.

[0032] Step S120: Using voltage error and its rate of change as input, a fuzzy neural network is used for joint reasoning and weighted calculation, and the output parameter correction amount is used to correct the outer loop PID control parameters to obtain the corrected fuzzy neural network.

[0033] Specifically, with voltage error and its rate of change As input, a fuzzy neural network is used to perform joint reasoning and weighted calculation through fuzzy membership functions, fuzzy rule bases, and neural network connection weights, outputting the parameter correction amount of the outer loop control module of the voltage. and The PID control parameters of the outer loop are corrected online to obtain an updated fuzzy neural network. Through online optimization of the PID control parameters, adaptive adjustment of the boost process is achieved.

[0034] Among them, the fuzzy neural network performs joint reasoning and weighted calculation through fuzzy membership functions, fuzzy rule bases and neural network connection weights.

[0035] Specifically, the fuzzy neural network includes an input layer, a fuzzy layer, a rule layer, a normalization layer, and an output layer, wherein the input of the input layer is the voltage error. and its error change rate The output of the output layer is the correction amount of the outer loop PID control parameters. and The input membership function adopts the form of triangular membership function. When implemented inside the FPGA, the membership degree is calculated by looking up a table, and the formaldehyde summation is completed by using a fixed weight multiplier to ensure real-time performance and hardware feasibility.

[0036] The voltage outer loop employs a fuzzy neural network structure to achieve self-tuning PID control. The fuzzy neural network takes voltage error and its rate of change as input, fuzzifies the input variables through a membership function layer, and performs inference calculations at the rule layer using preset fuzzy control rules. The output layer achieves nonlinear mapping through a weighted summation method of the neural network, outputting PID parameter correction values ​​for real-time adjustment of the outer loop PID controller parameters.

[0037] After self-tuning, the PID controller calculates the output current reference value based on the voltage error signal and uses it as the setpoint for the inner loop control, thereby enabling the outer loop to have online learning and adaptive adjustment capabilities, significantly improving the system's voltage regulation accuracy and dynamic performance.

[0038] By modifying parameters, the fuzzy neural network acquires self-learning capabilities. Its internal neural weights are modified in real time according to the changing trend of the system output error, enabling the controller to automatically adjust control parameters under conditions such as input voltage fluctuations or load changes, achieving adaptive optimization. Under conditions of input voltage fluctuations, load changes, and device parameter drift, it can continuously optimize control performance, achieve self-learning and online optimization, improve the system's adaptability and anti-disturbance capability, and maintain excellent dynamic performance and steady-state accuracy under different load and input conditions.

[0039] When the input voltage drops or the load changes abruptly, the controller automatically increases the proportional parameter to accelerate the response speed; when the output voltage approaches the target value, it automatically reduces the integral gain to suppress overshoot. Experimental results show that, under a load change rate of 50%, the control method of this invention can reduce voltage overshoot by about 45%, steady-state error by about 60%, and settling time by about 30%, demonstrating excellent dynamic performance.

[0040] In step S130, the modified fuzzy neural network is used to update the reference current based on the voltage error, and the duty cycle adjustment amount is obtained based on the current PI control algorithm.

[0041] Specifically, a dual closed-loop control strategy combining an outer voltage loop and an inner current loop is adopted. The outer loop uses a self-tuning PID control algorithm based on a fuzzy neural network, employing a modified fuzzy neural network. The outer voltage loop control is based on the voltage error. Adjust the output reference current The inner loop uses a current PI control algorithm, and the inner current loop control is based on the current error. The duty cycle adjustment value is obtained, and the duty cycle is quickly adjusted through the PWM control signal to achieve fast response and limiting control of the inductor current; the outer loop focuses on steady-state accuracy and robustness, while the inner loop focuses on dynamic response and safety constraints.

[0042] The outer voltage loop control has a control cycle of 1-5kHz and is used to achieve steady-state regulation of the output voltage. The inner current loop control has a control cycle of 100-200kHz and is used to achieve rapid tracking and current limiting control of the inductor current.

[0043] Both the self-tuning PID control algorithm and the current PI control algorithm of the fuzzy neural network are implemented through the FPGA main control module. Fuzzy inference, neural network weighted operation and control logic are implemented by table lookup and pipeline structure to reduce system control delay and improve real-time performance.

[0044] The PWM control signal is generated based on the control signal output from the inner current loop control and output to the main power switching device MOSFET via the gate driver chip. This controls the periodic on and off of the primary side of the flyback converter, thereby realizing the boost process of inductor energy storage and release.

[0045] The current PI control algorithm uses a proportional-integral algorithm, and its duty cycle is calculated recursively according to the following formula: ; In the formula, u(k) The duty cycle of the output. u(k- 1 ) This represents the duty cycle of the previous cycle. This is the proportionality coefficient. The integral coefficient is...e i This is for current error.

[0046] Specifically, the inner loop employs a high-speed PI controller to calculate the current error in real time. The PI controller, implemented within an FPGA using a fixed-point algorithm, achieves rapid current tracking by adjusting the duty cycle. The proportional element ensures system response speed, while the integral element eliminates steady-state errors. By appropriately selecting the proportional and integral coefficients, it enables rapid dynamic current tracking and smooth limiting, allowing the output current to quickly and stably follow changes in the reference value.

[0047] Meanwhile, the inner loop control cycle is synchronized with the PWM cycle, with a typical control frequency between 100-200kHz, ensuring the real-time performance and response speed of the current control.

[0048] Step S140: Generate a PWM control signal based on the duty cycle adjustment, and generate the target high voltage based on the PWM control signal.

[0049] Specifically, a corresponding PWM control signal is generated based on the duty cycle adjustment obtained from the current inner loop control. The PWM control signal is output to the main power switching device to control the periodic on and off of the primary side of the flyback converter, thereby adjusting the output voltage to the target high voltage and realizing the boost process of inductor energy storage and release.

[0050] In this embodiment, high-precision measurement under high-voltage conditions is achieved by sampling the output voltage and inductor current of the boost circuit in real time and calculating the voltage and current errors based on preset reference voltage and reference current. Using the voltage error and its rate of change as input, a fuzzy neural network is used for joint reasoning and weighted calculation of the output parameter correction amount, which is then used to correct the outer-loop PID control parameters, resulting in a corrected fuzzy neural network that enables adaptive adjustment of the boost process. The voltage error is input into the corrected fuzzy neural network to obtain the updated reference current, and based on the current PID control algorithm, the duty cycle adjustment amount is obtained according to the current error, achieving dual closed-loop control of voltage and current. This enables stable adjustment of the boost process under complex operating conditions. A PWM control signal is generated based on the duty cycle adjustment amount, and the output voltage is adjusted to the target high voltage based on the PWM control signal. This maintains dynamic performance under conditions of input voltage fluctuations, load changes, and parameter uncertainties, achieving rapid response, stable adjustment, and safe operation of the boost circuit. It is suitable for high-voltage energy storage and electronic fuse energy management applications where safety and practicality are paramount.

[0051] like Figure 2-3 As shown, a dual-closed-loop boost circuit control system based on a fuzzy neural network is provided to implement the dual-closed-loop boost circuit control method based on a fuzzy neural network as described above, including: The system includes a sampling feedback module 10, a voltage outer loop control module 20, a current inner loop control module 30, a PWM drive module 40, a power conversion module 50, an FPGA main control module 60, and a host computer monitoring module 70. The sampling feedback module 10 is connected to the power conversion 50 and the FPGA main control module 60. It is used to obtain the output voltage and inductor current of the boost circuit and input them to the FPGA main control module 60. The FPGA main control module 60 calculates the voltage error and current error by combining the preset reference voltage and reference current. The voltage outer loop control module 20 is connected to the FPGA main control module 60 and is used to output a reference current based on the voltage error using a modified fuzzy neural network. The current inner loop control module 30 is connected to the voltage outer loop control module 20, and is used to adjust the duty cycle according to the duty cycle adjustment amount, and input the adjusted duty cycle to the PWM drive module; The PWM drive module 40 is connected to the current inner loop control module 30 and is used to generate a PWM control signal according to the adjusted duty cycle. The power conversion module 50 is used to boost the output voltage to the target high voltage according to the PWM control signal; The FPGA main control module 60 is used to perform joint reasoning and weighted calculation using a fuzzy neural network based on the voltage error and its rate of change, and output parameter correction amount. The parameter correction amount is used to correct the outer loop PID control parameters to obtain the corrected fuzzy neural network. Then, a current-based PI control algorithm is used to obtain the duty cycle adjustment amount based on the current error. The monitoring module 80 is used for sending control commands, configuring fuzzy neural network parameters, displaying real-time voltage waveforms, and recording data.

[0052] Specifically, the modules described above operate collaboratively under unified timing control, and are connected via standard interfaces for easy system integration and expansion. The modules are interconnected through electrical and logic control signals, forming a complete closed-loop control system. The control system's input is a low-voltage DC power supply, and its output is a high-voltage DC signal, primarily used for charging the high-voltage capacitor in electronic fuse energy storage boost systems, but also applicable to other high-voltage power supply applications.

[0053] Furthermore, the aforementioned system completes fuzzy neural network calculation, PWM control generation, and dual closed-loop logic control based on the FPGA platform. By utilizing the parallel processing capabilities and fixed-point arithmetic optimization of the FPGA, the execution latency of the control algorithm is effectively reduced, and the real-time performance and control stability of the system are improved.

[0054] The sampling feedback module 10 is used to acquire the output voltage and inductor current of the boost circuit, and calculate the voltage error and current error based on preset reference voltage and reference current. The voltage and current signals are sampled through the isolation feedback module. The voltage sampling section uses a high-resistance voltage divider network with a voltage division ratio of 1000:1, and achieves electrical isolation between the high-voltage side and the control side through a linear optocoupler. The current sampling obtains the inductor current signal by connecting a 0.5Ω sampling resistor in series with the source of the power MOSFET. This signal is amplified and filtered at high speed and then sent to a high-speed comparator and an ADC module respectively, realizing dual-channel processing for current measurement and overcurrent protection. The isolation feedback module integrates a high-linearity optocoupler, an operational amplifier, and a 12-bit resolution ADC chip. This ADC is connected to the FPGA main control module through the SPI bus, and the maximum sampling rate can reach 200 ksps.

[0055] The voltage outer loop control module 20 employs a fuzzy neural network structure to achieve self-tuning PID control. The fuzzy neural network takes voltage error and its rate of change as input, performs fuzzification processing on the input variables through a membership function layer, and completes inference calculations at the rule layer using preset fuzzy control rules. The output layer achieves nonlinear mapping through a weighted summation method of the neural network, outputting PID parameter correction values ​​for real-time correction of the outer loop PID controller parameters. After self-tuning, the PID controller calculates the output current reference value based on the voltage error signal and uses it as the setpoint for the inner loop control, thereby enabling the outer loop to have online learning and adaptive adjustment capabilities, significantly improving the system's voltage regulation accuracy and dynamic performance.

[0056] The inner current control module 30 employs a high-speed PI controller to achieve rapid current tracking and limiting control. The inner loop controller calculates the PWM control signal based on the current error obtained from real-time sampling and adjusts the duty cycle to achieve a rapid response of the inductor current. The proportional element improves the dynamic response speed of the system, while the integral element eliminates steady-state errors, ensuring that the output current stably follows changes in the reference current. The inner current control cycle is synchronized with the PWM control cycle, with a typical control frequency of 100–200 kHz, thus guaranteeing the real-time performance and stability of the current control.

[0057] The PWM drive module 40 has a PWM control frequency set to 200kHz and a duty cycle adjustment range of 0-90%. The PWM control and drive section generates PWM control signals using the FPGA's internal counting and comparison logic. These signals are amplified by the gate driver chip and then drive the main power switching device MOSFET, enabling the periodic switching of the flyback transformer's primary side. The output high voltage is rectified by a voltage doubler and stored in a high-voltage capacitor for subsequent load release or electronic fuse energy triggering. Precise adjustment of the duty cycle ensures stable energy transfer and high-voltage output from the boost circuit. The entire PWM signal generation, sampling, and control calculation process is implemented in parallel logic within the FPGA. All modules operate synchronously under a unified system clock, resulting in an overall control delay of less than 1μs.

[0058] The power conversion module 50 adopts a flyback boost topology, including a main power switching device MOSFET, a flyback boost transformer, and a voltage doubler rectifier circuit.

[0059] Specifically, the power conversion module 50 is used to boost the output voltage to a set target value. In this embodiment, the power conversion module 10 adopts a flyback boost topology, mainly composed of MOSFET main power switching devices, a flyback boost transformer, and a voltage doubler rectifier circuit. The flyback boost transformer uses an EFD type PC40 ferrite core with a primary-to-secondary winding turns ratio of 3:100 to meet the high-voltage rapid boost requirement. The main power switching devices are MOSFETs suitable for high-frequency, high-voltage operating environments. The secondary rectification uses SP5L type high-voltage rectifier diodes with a withstand voltage rating of 5000V. Through the voltage doubler rectifier structure composed of a high-voltage capacitor and a diode, the output voltage is boosted to the set target value in a short time.

[0060] The FPGA main control module 60 is connected to the sampling feedback module 10, the voltage outer loop control module 20, the current inner loop control module 30, the PWM drive module 40, the power conversion module 50, and the host computer monitoring module 70 via signals. It is used to execute the fuzzy neural network self-tuning PID control algorithm and the current PI control algorithm. The system clock frequency of the FPGA main control module 60 can be set to 50MHz. The FPGA internally adopts a multi-stage pipeline structure, enabling the control algorithm calculation, PWM control, and protection logic to run in parallel under the same timing, thereby significantly reducing system control latency.

[0061] Specifically, the FPGA employs a multi-module parallel design, including a UART communication module, a PWM control signal generation module, a feedback signal processing module, an ADC data acquisition module, and a fuzzy neural network calculation module. All modules operate synchronously under a unified system clock. Within each PWM cycle, the control system first completes ADC sampling, then performs fuzzy neural network inference and PI control calculation, and updates the PWM control signal within the same cycle, thereby achieving adaptive closed-loop control within the cycle.

[0062] In addition, the host computer monitoring module 70 interacts with the FPGA via serial communication, enabling control command transmission, fuzzy neural network parameter configuration, real-time voltage waveform display, and data recording. The host computer interface integrates functions such as mode selection, parameter setting, data plotting, and operating status display, achieving visualized monitoring of the voltage boosting process.

[0063] The system also includes a hardware fast-break protection module 80, which is connected to the sampling feedback module 10 and the PWM drive module 40. It is used to directly generate a shutdown signal and force the main power switch to shut down when the inductor current exceeds a preset threshold.

[0064] The hardware fast-break protection module 80 includes a high-speed comparator and a digital isolator. The high-speed comparator is used to monitor the current sampling signal in real time and determine whether the sampling current exceeds the set threshold. The digital isolator is used to directly output a PWM shutdown signal to the PWM drive module when the sampling current exceeds the set threshold.

[0065] To further enhance system safety, the system also includes a hardware fast-break module 80, which establishes a hardware fast-break channel in the current sampling circuit. This hardware fast-break protection channel utilizes a high-speed comparator LTC6752. When the inductor current exceeds a set threshold, the comparator immediately outputs an overcurrent signal, which is transmitted to the FPGA control logic via a digital isolator. Upon detecting the overcurrent signal, the FPGA bypasses the software control path and directly shuts off the PWM control signal, causing the main power MOSFET to turn off within sub-microseconds, effectively preventing device breakdown and circuit damage caused by abnormal overcurrent. This hardware fast-break protection mechanism operates independently of the control algorithm, exhibiting extremely high response speed and reliability. Experimental results show that the response time of this hardware fast-break mechanism is less than 1μs, significantly improving the system's safety and reliability.

[0066] In summary, this invention achieves rapid, stable, and safe control of the boost circuit by constructing a dual-closed-loop control structure on an FPGA platform, combining a fuzzy neural network voltage outer loop with a PI current-limiting inner loop, and introducing a hardware fast-break protection mechanism. This integration of intelligent control algorithms and hardware logic characteristics enables fast, stable, and safe control of the boost circuit. The outer loop handles overall voltage regulation and adaptive optimization, while the inner loop handles current limiting and protection response. The hardware fast-break mechanism ensures safety under extreme conditions, resulting in an intelligent boost control system with high dynamic response, high robustness, and high reliability. This dual-closed-loop boost circuit control system and method feature fast boost speed, high control accuracy, strong robustness, and high safety, making it widely applicable in electronic fuse power supply, high-voltage charging modules, energy recovery systems, and other power electronic equipment requiring rapid voltage boost and precise control, such as electronic fuse energy storage boost systems, high-voltage capacitor charging power supplies, and other energy conversion applications requiring rapid high-voltage output. It can also be extended to high-voltage pulse power supplies, laser drive power supplies, and new energy conversion devices, demonstrating promising engineering application prospects.

[0067] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0068] It is obvious to those skilled in the art that the modules or steps of the present invention described above can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. Optionally, they can be implemented using computer-executable program code, thereby storing them in a computer storage medium (ROM / RAM, magnetic disk, optical disk) for execution by the computing device. In some cases, the steps shown or described can be performed in a different order than those described herein, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Therefore, the present invention is not limited to any particular hardware and software combination.

[0069] The above description, in conjunction with specific embodiments, provides a further detailed explanation of the present invention. It should not be construed that the specific implementation of the present invention is limited to these descriptions. For those skilled in the art, various simple deductions or substitutions can be made without departing from the concept of the present invention, and all such deductions or substitutions should be considered within the scope of protection of the present invention.

Claims

1. A control method for a dual-closed-loop boost circuit based on a fuzzy neural network, characterized in that, Includes the following steps: The output voltage and inductor current of the boost circuit are sampled in real time, and the voltage error and current error are calculated based on the preset reference voltage and reference current. Using the voltage error and its rate of change as input, a fuzzy neural network is used for joint reasoning and weighted calculation to output parameter correction. The parameter correction is then used to correct the outer loop PID control parameters to obtain the corrected fuzzy neural network. A modified fuzzy neural network is used to update the reference current based on the voltage error, and the duty cycle adjustment amount is obtained based on the current PI control algorithm according to the current error. A PWM control signal is generated based on the duty cycle adjustment, and a target high voltage is generated based on the PWM control signal.

2. The control method for a dual closed-loop boost circuit based on a fuzzy neural network according to claim 1, characterized in that, The real-time sampling of the output voltage and inductor current of the boost circuit, and the calculation of voltage and current errors based on preset reference voltage and reference current, include: The output voltage and inductor current of the boost circuit are obtained through voltage sampling circuit and current sampling circuit, respectively. and output voltage With the preset reference voltage By comparison, the voltage error is obtained. : ； inductor current Reference current given by the outer voltage loop By comparison, the current error is calculated. : 。 3. The dual closed-loop boost circuit control method based on fuzzy neural network according to claim 1, characterized in that, The fuzzy neural network performs joint reasoning and weighted calculations using fuzzy membership functions, fuzzy rule bases, and neural network connection weights.

4. The dual closed-loop boost circuit control method based on a fuzzy neural network according to claim 1, characterized in that, The current PI control algorithm uses a proportional-integral algorithm, and its duty cycle is calculated recursively according to the following formula: ； In the formula, u(k) The duty cycle of the output. u(k- 1 ) This represents the duty cycle of the previous cycle. This is the proportionality coefficient. The integral coefficient is... e i This is the current error.

5. The control method for a dual closed-loop boost circuit based on a fuzzy neural network according to claim 1, characterized in that, After sampling the output voltage and inductor current of the boost circuit in real time, the method further includes: The inductor current is monitored in real time, and a PWM shutdown signal is output when the inductor current exceeds a preset threshold to forcibly shut down the main power switching device.

6. A dual-closed-loop boost circuit control system based on a fuzzy neural network, characterized in that, A method for implementing a dual closed-loop boost circuit control based on a fuzzy neural network as described in any one of claims 1-5 includes: The system includes a sampling feedback module, a voltage outer loop control module, a current inner loop control module, a PWM drive module, a power conversion module, an FPGA main control module, and a host computer monitoring module. The sampling feedback module is connected to the power conversion module and the FPGA main control module, and is used to obtain the output voltage and inductor current of the boost circuit and input them to the FPGA main control module. The FPGA main control module calculates the voltage error and current error in combination with the preset reference voltage and reference current. The voltage outer loop control module is connected to the FPGA main control module and is used to adjust the output reference current according to the voltage error. The current inner loop control module is connected to the voltage outer loop control module and is used to adjust the duty cycle according to the duty cycle adjustment amount, and input the adjusted duty cycle to the PWM drive module; The PWM drive module is connected to the current inner loop control module and is used to generate a PWM control signal according to the adjusted duty cycle. The power conversion module is used to boost the output voltage to the target high voltage according to the PWM control signal; The FPGA main control module is used to perform joint reasoning and weighted calculation using a fuzzy neural network based on the voltage error and its rate of change, and output parameter correction amount. The parameter correction amount is used to correct the outer loop PID control parameters to obtain the corrected fuzzy neural network. The module then uses a current-based PI control algorithm to obtain the duty cycle adjustment amount based on the current error. The monitoring module is connected to the FPGA main control module and is used for sending control commands, configuring fuzzy neural network parameters, displaying real-time voltage waveforms, and recording data.

7. A dual-closed-loop boost circuit control system based on a fuzzy neural network according to claim 6, characterized in that, The power conversion module adopts a flyback boost topology, including a main power switching device MOSFET, a flyback boost transformer, and a voltage doubler rectifier circuit.

8. The dual closed-loop boost circuit control system based on a fuzzy neural network according to claim 6, characterized in that, The PWM drive module uses a gate drive chip to drive the main power switching device MOSFET, thereby realizing high-frequency turn-on and turn-off control of the primary side of the flyback transformer.

9. A dual-closed-loop boost circuit control system based on a fuzzy neural network according to claim 6, characterized in that, Also includes: A hardware fast-break protection module, which is connected to the sampling feedback module and the PWM drive module, is used to directly generate a shutdown signal and forcibly shut down the main power switch when the inductor current exceeds a preset threshold.

10. A dual-closed-loop boost circuit control system based on a fuzzy neural network according to claim 9, characterized in that, The hardware fast-break protection module includes a high-speed comparator and a digital isolator. The high-speed comparator is used to monitor the current sampling signal in real time and determine whether the sampling current exceeds a set threshold. The digital isolator is used to directly output a PWM shutdown signal to the PWM drive module when the sampling current exceeds the set threshold.

Images