Magnetic field transmitter capable of automatically adjusting control parameters based on AI

Abstract

The invention relates to the technical field of magnetic field generation and control, in particular to a magnetic field transmitter capable of automatically adjusting control parameters based on AI. Comprising the steps that a data acquisition unit acquires multi-dimensional operation parameters; the performance judgment unit judges the current output stability based on the steady-state error; the data analysis unit determines a PID preposition adjusting quantity and a DAC control quantity through deep reinforcement learning; the power supply regulation and control unit generates a PWM signal to control the output voltage of the switching power supply; the compensation judgment unit judges whether feedforward compensation needs to be carried out or not according to the voltage change rate; the feed-forward compensation unit determines a target duty ratio of the H-bridge driving signal; and the deviation analysis unit performs iterative optimization on the PID preposition adjusting quantity based on the deviation between the theoretical current value and the set value. The output precision, the dynamic response speed and the system stability of the magnetic field transmitter are effectively improved.

Description

Technical Field

[0001] This invention relates to the field of magnetic field generation and control technology, and in particular to a magnetic field transmitter based on AI-based automatic adjustment of control parameters. Background Technology

[0002] As key equipment in fields such as electromagnetic exploration, magnetic resonance imaging, and wireless power transmission, the stability of the output current, the accuracy of multi-machine synchronization, and the system reliability of magnetic field transmitters directly affect the application results. With the development of artificial intelligence technology, introducing AI algorithms into the control of magnetic field transmitters has become an important direction for improving system performance.

[0003] Chinese Patent Publication No. CN112484567A discloses a magnetic field control method for a reconnected electromagnetic transmitting device. This method improves the energy transmission efficiency of the device by establishing a magnetic field model and an electromagnetic force model, adding an E-type magnetic core around the rear side of the drive coil, and adjusting the width of the core inside the coil. This increases the peak electromagnetic force and delays the appearance time of the reverse force. However, this method mainly relies on static magnetic field structure optimization, improving the magnetic field distribution through physical means (magnetic core), and does not involve dynamic parameter adjustment of the control system, making it unable to adapt to real-time changes in operating conditions such as load variations and temperature drift.

[0004] Current magnetic field transmitters and their control technologies still face the following prominent problems: 1. Static control parameters and lack of adaptive capability: The PID control parameters of existing magnetic field transmitters are mostly determined by fixed values ​​or manual trial and error. When the load impedance changes, the ambient temperature fluctuates, or the device ages, the fixed parameters cannot guarantee the control performance, resulting in large fluctuations in output current, difficulty in eliminating steady-state error, and slow dynamic response.

[0005] 2. Single control strategy and lack of multi-level collaborative optimization: Existing technologies usually rely only on PID feedback control without combining feedforward compensation mechanism, making it difficult to simultaneously take into account dynamic response speed and steady-state accuracy; when the power supply voltage fluctuates or the load changes suddenly, relying solely on PID regulation results in a slow response, which can easily lead to output current overshoot or oscillation.

[0006] In summary, existing technologies have not yet disclosed an intelligent magnetic field transmitter capable of real-time adaptive adjustment of PID parameters and possessing multi-layer collaborative control with voltage feedforward and current feedback. This invention addresses this technological gap. Summary of the Invention

[0007] To address this issue, the present invention provides a magnetic field transmitter based on AI-based automatic adjustment of control parameters, thereby overcoming the problems of static PID control parameters and lack of adaptive capability in existing magnetic field transmitters.

[0008] To achieve the above objectives, the present invention provides a magnetic field transmitter based on AI-automatic adjustment of control parameters, comprising: The data acquisition unit is used to acquire multi-dimensional operating parameters of the magnetic field transmitter, including the effective value of the output current, the output voltage, the load impedance, the ambient temperature, and the H-bridge switching frequency. The performance determination unit is used to determine the steady-state error value based on the deviation between the effective value of the output current and the set current value, and to determine whether the current output stability of the current magnetic field transmitter meets the standard based on the steady-state error value and the preset error value. The data analysis unit is used to determine the pre-adjustment amount of the PID control parameters and the DAC output control amount based on the multi-dimensional operating parameters using a deep reinforcement learning algorithm. The power supply regulation unit is used to determine the output voltage value of the switching power supply based on the duty cycle preset value and carrier frequency of the PWM modulation signal of the switching power supply. The compensation determination unit is used to determine whether feedforward compensation is needed for the duty cycle of the H-bridge drive signal based on the voltage change rate of the output voltage value within a preset time and the preset change rate. The feedforward compensation unit is used to determine the first target duty cycle of the H-bridge drive signal based on the ratio of the rate of change of voltage change to the preset rate of change. The deviation analysis unit is used to determine whether the pre-adjustment amount of the PID control parameters is qualified based on the absolute value of the deviation between the theoretical output current value and the set current value, and sends the updated pre-adjustment amount to the data analysis unit, so that the effective value of the output current reaches the set current value.

[0009] Furthermore, the steady-state error value is determined based on the deviation between the effective value of the output current and the set current value; The performance determination unit determines that the current output stability of the current magnetic field transmitter is substandard based on the steady-state error value being greater than the preset steady-state error value.

[0010] Furthermore, in response to the current output stability of the current magnetic field transmitter not meeting the standard, the data analysis unit is used to construct the current system state vector based on the steady-state error value, the rate of change of the steady-state error value, the output voltage value, the load impedance value, the ambient temperature value, and the H-bridge switching frequency value. The current system state vector is then input into a deep reinforcement learning model, which outputs the pre-adjustment amount of the PID control parameters.

[0011] Furthermore, the data analysis unit is also used to calculate the updated PID control parameters based on the pre-adjustment amount of the PID control parameters and the basic PID control parameters; Based on the updated PID control parameters, the output control quantity of the PID controller is determined by the PID control algorithm. Based on the output control quantity of the PID controller, the output control quantity of the DAC is determined by digital-to-analog conversion; wherein, the PID control parameters are proportional coefficient, integral coefficient, and derivative coefficient.

[0012] Furthermore, the duty cycle preset value of the PWM modulation signal of the switching power supply is determined based on the DAC output control quantity, the input voltage value of the switching power supply, and the preset power efficiency. The carrier frequency of the PWM modulation signal is determined based on the H-bridge switching frequency value and a preset multiple relationship; The power control unit generates the PWM modulation signal based on the duty cycle preset value and the carrier frequency; Based on the PWM modulation signal, the output voltage value of the switching power supply is determined by the power conversion circuit.

[0013] Furthermore, the compensation determination unit is used to determine, based on the voltage change rate being greater than a preset change rate, that feedforward compensation is needed for the duty cycle of the H-bridge drive signal.

[0014] Furthermore, in response to the need for feedforward compensation of the duty cycle of the H-bridge drive signal, The feedforward compensation unit is used to determine the current compensation coefficient based on the ratio of the voltage change rate to the preset change rate. The duty cycle of the reference H-bridge drive signal is compensated based on the current compensation coefficient to obtain the first target duty cycle.

[0015] Furthermore, the theoretical output current value is determined based on the output voltage value and load impedance value of the switching power supply; The absolute value of the deviation is determined based on the theoretical output current value and the set current value. The deviation analysis unit is used to determine that the pre-adjustment amount of the PID control parameter is unqualified based on the absolute value of the deviation being greater than a preset current deviation threshold.

[0016] Furthermore, in response to the failure of the pre-adjustment amount of the PID control parameter, the deviation analysis unit also includes a correction direction determination module; The correction direction determination module is used to determine the correction direction based on the comparison result between the theoretical output current value and the set current value.

[0017] Furthermore, the deviation analysis unit also includes a correction coefficient calculation module; The correction coefficient calculation module is used to determine the correction coefficient based on the ratio of the absolute value of the deviation to a preset current deviation threshold. The correction coefficient is positively correlated with the absolute value of the deviation; the larger the deviation, the greater the correction magnitude. Based on the correction direction and correction coefficient, the pre-adjustment amounts of the proportional coefficient, integral coefficient, and derivative coefficient are adjusted in the corresponding directions. The updated PID pre-adjustment value is sent to the data analysis unit, which replaces the original pre-adjustment value for subsequent PID parameter updates.

[0018] Compared with the prior art, the beneficial effects of the present invention are that, by using a deep reinforcement learning algorithm, the pre-adjustment of PID control parameters is dynamically optimized based on multi-dimensional operating parameters, overcoming the shortcomings of traditional PID parameters being fixed and unable to adapt to changes in operating conditions; when the load impedance changes, the ambient temperature fluctuates, or the device ages, the system can automatically adjust the control parameters, so that the effective value of the output current is always stable near the set value, and the steady-state error is controlled within 0.1 Amperes, which significantly improves the output accuracy of the magnetic field transmitter.

[0019] Furthermore, this invention adds a feedforward compensation mechanism based on voltage change rate to the PID feedback control. When a rapid fluctuation in output voltage is detected, the system immediately adjusts the duty cycle of the H-bridge drive signal to compensate before the voltage fluctuation is transmitted to the current, effectively suppressing the instantaneous fluctuation of the output current. At the same time, the PID pre-adjustment is corrected based on the deviation between the theoretical current and the set current to ensure the steady-state accuracy of the system in long-term operation, further improving the dynamic response speed and steady-state accuracy of the system.

[0020] Furthermore, this invention uses the theoretical output current value as an intermediate monitoring indicator, and judges whether the PID pre-adjustment is reasonable based on the deviation between the theoretical current and the set current. When the deviation exceeds the preset threshold, the system automatically determines the correction direction and correction coefficient, adjusts the PID pre-adjustment accordingly, and feeds back the updated parameters to the data analysis unit so that the system can actively detect and correct the control deviation.

[0021] Furthermore, this invention sets the PWM carrier frequency of the switching power supply to 10 times the switching frequency of the H-bridge, making the power supply ripple frequency much higher than the H-bridge output base frequency. Combined with the subsequent filter circuit design, the output current ripple coefficient can be reduced to below 1%. At the same time, the PWM duty cycle is dynamically calculated based on the DAC output control quantity, real-time input voltage, and preset power supply efficiency, which compensates for the impact of input voltage fluctuations and efficiency changes on the output voltage, further improving the stability of the output current. Attached Figure Description

[0022] Figure 1 This is a schematic diagram of the system structure of a magnetic field transmitter based on AI-automatic adjustment of control parameters according to an embodiment of the present invention; Figure 2 This is a logic block diagram showing whether the current output stability of the current magnetic field transmitter meets the standard in this embodiment of the invention; Figure 3 A logic block diagram for determining whether the pre-adjustment amount of the PID control parameters is qualified in an embodiment of the present invention; Figure 4 The following is a logic block diagram for determining the PID parameter correction direction and correction coefficient in an embodiment of the present invention. Detailed Implementation

[0023] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0024] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0025] It should be noted that in the description of this invention, the terms "upper", "lower", "left", "right", "inner", "outer", etc., which indicate directions or positional relationships, are based on the directions or positional relationships shown in the accompanying drawings. This is only for the convenience of description and is not intended to indicate or imply that the device or element must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation of this invention.

[0026] Furthermore, it should be noted that, in the description of this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0027] Please see Figure 1 As shown, Figure 1 This is a schematic diagram of the system structure of a magnetic field transmitter based on AI-automatic adjustment of control parameters, according to an embodiment of the present invention.

[0028] This invention relates to an AI-based magnetic field transmitter that automatically adjusts control parameters, comprising: The data acquisition unit is used to acquire multi-dimensional operating parameters of the magnetic field transmitter, including the effective value of the output current, the output voltage, the load impedance, the ambient temperature, and the H-bridge switching frequency. The performance determination unit is connected to the data acquisition unit. It is used to determine the steady-state error value based on the deviation between the effective value of the output current and the set current value, and to determine whether the current output stability of the current magnetic field transmitter meets the standard based on the steady-state error value and the preset error value. The data analysis unit, connected to the performance determination unit, is used to determine the pre-adjustment amount of the PID control parameters and the DAC output control amount based on the multi-dimensional operating parameters and through a deep reinforcement learning algorithm; the power supply regulation unit, connected to the data analysis unit, is used to determine the output voltage value of the switching power supply based on the duty cycle preset value and carrier frequency of the PWM modulation signal of the switching power supply. The compensation determination unit is connected to the power control unit and is used to determine whether feedforward compensation is needed for the duty cycle of the H-bridge drive signal based on the voltage change rate of the output voltage value within a preset time and the preset change rate. A feedforward compensation unit, connected to the compensation determination unit, is used to determine the first target duty cycle of the H-bridge drive signal based on the ratio of the rate of change of voltage change to the preset rate of change. The deviation analysis unit is connected to the feedforward compensation unit and the data analysis unit respectively. It is used to determine whether the pre-adjustment amount of the PID control parameter is qualified based on the absolute value of the deviation between the theoretical output current value and the set current value, and send the updated pre-adjustment amount to the data analysis unit, so that the effective value of the output current reaches the set current value.

[0029] In this embodiment of the invention, the data acquisition unit is connected to the current sensor, voltage sensor, and temperature sensor within the magnetic field transmitter to acquire multi-dimensional operating parameters in real time. These multi-dimensional operating parameters include: the effective value of the output current, the output voltage value, the load impedance value, the ambient temperature value, and the H-bridge switching frequency value. The sampling period of the data acquisition unit is set to 100 μs to meet the requirements of real-time control.

[0030] Please see Figure 2 As shown, Figure 2 This is a logic block diagram for determining whether the current output stability of the current magnetic field transmitter described in this embodiment of the invention meets the standard.

[0031] Specifically, the performance determination unit calculates the steady-state error value based on the effective value of the output current and the user-set current value, and determines whether the current output stability of the current magnetic field transmitter meets the standard based on the comparison between the steady-state error value e and the preset steady-state error value Et. The formula for calculating the steady-state error value is as follows:

[0032] in, This is the effective value of the output current. Here, e represents the set current value and e represents the steady-state error value. Based on e > Et, it is determined that the current output stability of the current magnetic field transmitter is not up to standard; Based on e≤Et, it is determined that the current output stability of the current magnetic field transmitter meets the standard, and the system continues to monitor.

[0033] In this embodiment of the invention, the set current value should not exceed the rated current of the coil to avoid overheating or insulation damage. The rated current of the coil is determined by the cross-sectional area of ​​the conductor, heat dissipation conditions and insulation class. In order to ensure control accuracy, the set current value should not be too low to avoid excessive relative error. In practice, the set current value is 50%-70% of the rated current, preferably 60%.

[0034] In this embodiment of the invention, the preset steady-state error value is preferably 0.1 amperes, corresponding to a relative error range of 0.5%-1% of the set current value. Under this threshold, the stability of the magnetic field strength can be guaranteed, avoiding frequent triggering of unnecessary adjustment actions due to small fluctuations, and ensuring that the system can respond in a timely manner when the load changes, which meets the application requirements of high-precision magnetic field control.

[0035] In this embodiment, the system state vector consists of the following six dimensions: The steady-state error value e, the rate of change of the steady-state error value d, the output voltage value V, the load impedance value Z, the ambient temperature value T, and the H-bridge switching frequency value f.

[0036] The formula for calculating the rate of change of the steady-state error value is as follows:

[0037] in, The sampling period is This represents the steady-state error value at the current sampling time. This is the steady-state error value at the previous sampling time.

[0038] The system state vector is then input into a pre-trained deep reinforcement learning model. This embodiment employs a dual-delay deep deterministic policy gradient algorithm, which includes one Actor network and two Critic networks. The Actor network takes the system state vector as input and outputs the pre-adjustment of the PID control parameters. The Critic networks evaluate the value of the Actor's output action, guiding the network parameter updates.

[0039] Specifically, the training method for the deep reinforcement learning model is as follows: (1) Environmental modeling The control process of the magnetic field transmitter is modeled as a Markov decision process. The state space is the system state vector S.t =[e, d, V, Z, T, f], where the action space is the pre-adjustment A of the PID control parameters. t =[ΔK p ,ΔK i ,ΔK d The reward function is designed to be positively correlated with control performance. This means encouraging the reduction of steady-state error and avoiding excessive adjustment.

[0040] (2) Network construction The following neural network is constructed using the dual-delay deep deterministic policy gradient (TD3) algorithm: Actor Network: Input layer (6-dimensional state) → Hidden layer 1 (256 nodes, ReLU) → Hidden layer 2 (256 nodes, ReLU) → Output layer (3-dimensional action, Tanh activation), with the output range being [−1,1], which is then scaled to the actual action range.

[0041] Two Critic networks: Input layer (state + action, 9-dimensional) → Hidden layer 1 (256 nodes, ReLU) → Hidden layer 2 (256 nodes, ReLU) → Output layer (1-dimensional Q-value).

[0042] Target network: The Actor and Critic network structures are copied separately, and the parameters are synchronized through soft updates.

[0043] (3) Experience collection A magnetic field transmitter model was run in a simulation environment to collect empirical data under different operating conditions. A quintuple (s) was recorded at each time step. t ,a t ,r t ,s t+1 (done), store in the experience replay pool. The replay pool capacity is set to 10. 6 A priority experience playback mechanism is adopted to improve training efficiency.

[0044] (4) Training process The training uses the following hyperparameters: discount factor γ = 0.99, learning rate... The soft update coefficient τ = 0.005, the batch size = 256, the exploration noise N(0, 0.1), the target policy noise N(0, 0.2), and the pruning range [−0.5, 0.5].

[0045] Each training cycle, a batch is randomly sampled from the replay pool, and the network is updated according to the following steps: Calculate the target Q value:

[0046] Update the Critic network: Minimize

[0047] Update the Actor network every 2 steps: Maximize Q1(s,μ(s)) Soft update target network:

[0048] Training continues until the reward function converges, which typically takes 10 seconds. 5 -10 6 Time step.

[0049] (5) After training is completed, the model performance is verified using a test set that was not used in training.

[0050] Evaluation indicators include: The average steady-state error e < 0.1 , Maximum overshoot σmax < 5%, Average settling time <100ms, If the above indicators are met, the model is qualified and can be deployed to the actual system.

[0051] (6) Model Deployment The trained Actor network parameters are exported and deployed to the data analysis unit. In actual operation, the current state vector is collected in each control cycle and input into the Actor network for forward inference to obtain the PID pre-adjustment, thus achieving real-time control.

[0052] Specifically, in response to the current output stability of the current magnetic field transmitter not meeting the standard, the data analysis unit is used to construct the current system state vector based on the steady-state error value, the rate of change of the steady-state error value, the output voltage value, the load impedance value, the ambient temperature value, and the H-bridge switching frequency value. The current system state vector is then input into a deep reinforcement learning model, which outputs the pre-adjustment amount of the PID control parameters.

[0053] Specifically, the data analysis unit is also used to calculate the updated PID control parameters based on the pre-adjustment amount of the PID control parameters and the basic PID control parameters; Based on the updated PID control parameters, the output control quantity of the PID controller is determined by the PID control algorithm. Based on the output control quantity of the PID controller, the output control quantity of the DAC is determined by digital-to-analog conversion; wherein, the PID control parameters are proportional coefficient, integral coefficient, and derivative coefficient.

[0054] PID parameter update and DAC output determination: Pre-adjustment amount ΔK based on the output proportional coefficientp The pre-adjustment amount ΔK of the integral coefficient i The pre-adjustment amount ΔK of the differential coefficient d Update the PID parameters according to the following formula: K p '=K pb +ΔK p ,K i '=K ib +ΔK i ,K d '=K db +ΔK d , Among them, K pb K ib K db These are the basic PID control parameters.

[0055] Then, based on the updated PID control parameters, the output control quantity u(t) of the PID controller is calculated using the digital PID control algorithm:

[0056] Map the PID output control quantity u(t) (normalized to the 0-1 range) to the DAC output control quantity D. DAC This embodiment uses a 12-bit DAC with a reference voltage Vr = 5V, and the mapping formula is as follows: .

[0057] Specifically, the power supply regulation unit is used to determine the duty cycle preset value D of the PWM modulation signal of the switching power supply based on the DAC output control quantity, the input voltage value of the switching power supply, and the preset power efficiency; the formula for the duty cycle preset value is as follows:

[0058] Where K is the power supply voltage gain, that is, the amplification factor from the DAC output voltage to the power supply output voltage, K=9.6; η is the real-time sampled input voltage of the switching power supply, and η is the preset power supply efficiency.

[0059] During the system debugging phase, the input power Pin and output power Pout under different load conditions were measured using a power analyzer, and the actual efficiency was calculated according to the formula η=Pout / Pin. The measured results show that the power supply efficiency is stable between 0.91 and 0.93 within the range of 20%-100% rated load. In this implementation, the intermediate value of 0.92 is taken as the preset efficiency.

[0060] Simultaneously, based on the H-bridge switching frequency f and the preset multiple N=10, the carrier frequency fs of the PWM modulation signal is determined using the following formula: fs=N×f Finally, a PWM modulation signal is generated based on the duty cycle preset value and carrier frequency to drive the power transistor of the switching power supply, thereby obtaining the actual output voltage.

[0061] In this embodiment of the invention, the output voltage ripple of the switching power supply is inversely proportional to the switching frequency. Setting the carrier frequency of the PWM modulation signal to 10 times the switching frequency of the H-bridge makes the power supply ripple frequency much higher than the H-bridge output frequency, which facilitates the design of subsequent filter circuits and effectively reduces the ripple content of the output current.

[0062] Specifically, the feedforward compensation unit is used to determine whether feedforward compensation is needed for the duty cycle of the H-bridge drive signal based on the voltage change rate of the output voltage value within a preset time and a preset change rate. The voltage change rate The calculation formula is as follows:

[0063] in, The output voltage at the current sampling moment. The output voltage at the previous sampling time. The sampling period.

[0064] This invention also sets a preset change rate L1 = 1V / ms, based on the following: the system's rated output voltage is 48V, and the allowable voltage fluctuation range is ±5%, i.e., the maximum allowable fluctuation amplitude is 2.4V. To ensure timely detection and compensation before fluctuations affect the output current, the expected detection response time is no more than 2.4ms. Therefore, the threshold is calculated as 2.4V / 2.4ms = 1V / ms. This means that when the voltage change rate exceeds 1V / ms, the voltage fluctuation is considered to be potentially exceeding the allowable range, requiring triggering feedforward compensation.

[0065] Based on the fact that the voltage change rate is less than or equal to the preset change rate, it is determined that no feedforward compensation is needed for the duty cycle of the H-bridge drive signal; the first target duty cycle is the reference duty cycle. Based on the fact that the voltage change rate is greater than the preset change rate, it is determined that the duty cycle of the H-bridge drive signal needs to be feedforward compensated, and the current compensation coefficient is determined in the following manner: The current compensation coefficient is determined based on the ratio of the voltage change rate to the preset change rate. Based on the fact that the ratio of the rate of change is greater than zero, a current compensation coefficient less than 1 is determined to reduce the duty cycle of the H-bridge drive signal. Based on the fact that the ratio of the rate of change is less than zero, a current compensation coefficient greater than 1 is determined to increase the duty cycle of the H-bridge drive signal. The magnitude of the compensation coefficient is positively correlated with the absolute value of the voltage change rate; the larger the change rate, the greater the compensation amplitude. Specifically, the compensation coefficient X can be calculated as X = 1 ± α × L / L1, where α is a preset proportional coefficient.

[0066] The formula for calculating the first target duty cycle Q1 is: Q1 = Qb × X; where Qb is the duty cycle of the reference H-bridge drive signal.

[0067] In this embodiment, the preset proportional coefficient α is in the range of [0.2-0.5], preferably 0.3. If α is less than 0.2, the compensation is too weak and cannot effectively suppress the current fluctuation caused by voltage drop within the 2.4ms detection window; if α is greater than 0.5, the compensation is too strong and can easily lead to overshoot of H-bridge duty cycle adjustment, causing secondary oscillation.

[0068] Please see Figures 3-4 As shown, Figure 3 A logic block diagram for determining whether the pre-adjustment amount of the PID control parameters is qualified in an embodiment of the present invention; Figure 4 The following is a logic block diagram for determining the PID parameter correction direction and correction coefficient in an embodiment of the present invention.

[0069] Specifically, the deviation analysis unit is used to determine the theoretical output current value I based on the output voltage value Vo and the load impedance value Z of the switching power supply, according to Ohm's law. The calculation formula is as follows: I = Vo / Z; Calculate the absolute value of the deviation between the theoretical output current value It and the set current value Is, ΔI=|It-Is|; and determine whether the pre-adjustment amount of the PID control parameter is qualified based on the comparison result of the absolute value of the deviation ΔI and the preset current deviation threshold ΔIth. Based on the result that the absolute value of the deviation is greater than the preset current deviation threshold, it is determined that the pre-adjustment amount of the PID control parameter is unqualified, and the data analysis unit is triggered to redetermine the pre-adjustment amount of the PID control parameter. Based on the result that the absolute value of the deviation is less than or equal to the preset current deviation threshold, the pre-adjustment amount of the PID control parameter is determined to be qualified, and the current PID parameter is maintained.

[0070] In this embodiment, a preset steady-state error value of 0.1 amperes is used to determine whether the system meets the standard, while a preset current deviation threshold of 0.5A is used to determine whether the PID pre-adjustment needs to be corrected. The two form a 5-fold multiplier relationship to ensure that the various control mechanisms work in coordination and do not interfere with each other.

[0071] When correction is required, the correction direction determination module determines the correction direction based on the relationship between the theoretical output current value and the set current value: Based on the fact that the theoretical output current value is less than the set current value, it is determined that a positive correction is needed, that is, to increase the PID pre-regulation value in order to improve the output voltage of the switching power supply. Based on the fact that the theoretical output current value is greater than the set current value, it is determined that a negative correction is needed, that is, to reduce the PID pre-regulation amount in order to reduce the output voltage of the switching power supply.

[0072] The correction coefficient calculation module is used to determine the correction coefficient based on the ratio of the absolute value of the deviation to the preset current deviation threshold. The larger the absolute value of the deviation, the greater the correction magnitude.

[0073] Specifically, the correction coefficient β can be calculated according to β=ΔI / ΔIth.

[0074] Based on the correction direction and correction coefficient, the pre-adjustment amounts of the proportional coefficient, integral coefficient, and derivative coefficient are adjusted in the corresponding directions to obtain the updated pre-adjustment amount [ΔK]. pn ,ΔK in ,ΔK dn ].

[0075] The updated pre-adjustment value is sent to the data analysis unit, which replaces the original pre-adjustment value for subsequent PID parameter updates. Simultaneously, based on the first target duty cycle and the current deviation detection result, the final target duty cycle of the H-bridge drive signal is determined, driving the H-bridge output to stabilize the effective value of the output current at the set current value.

[0076] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

[0077] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A magnetic field transmitter based on AI-automatic adjustment of control parameters, characterized in that, include: The data acquisition unit is used to acquire multi-dimensional operating parameters of the magnetic field transmitter, including the effective value of the output current, the output voltage, the load impedance, the ambient temperature, and the H-bridge switching frequency. The performance determination unit is used to determine the steady-state error value based on the deviation between the effective value of the output current and the set current value, and to determine whether the current output stability of the current magnetic field transmitter meets the standard based on the steady-state error value and the preset error value. The data analysis unit is used to determine the pre-adjustment amount of the PID control parameters and the DAC output control amount based on the multi-dimensional operating parameters using a deep reinforcement learning algorithm. The power supply regulation unit is used to determine the output voltage value of the switching power supply based on the duty cycle preset value and carrier frequency of the PWM modulation signal of the switching power supply. The compensation determination unit is used to determine whether feedforward compensation is needed for the duty cycle of the H-bridge drive signal based on the voltage change rate of the output voltage value within a preset time and the preset change rate. The feedforward compensation unit is used to determine the first target duty cycle of the H-bridge drive signal based on the ratio of the rate of change of voltage change to the preset rate of change. The deviation analysis unit is used to determine whether the pre-adjustment amount of the PID control parameters is qualified based on the absolute value of the deviation between the theoretical output current value and the set current value, and sends the updated pre-adjustment amount to the data analysis unit, so that the effective value of the output current reaches the set current value.

2. The magnetic field transmitter based on AI-automatic adjustment of control parameters according to claim 1, characterized in that, The steady-state error value is determined based on the deviation between the effective value of the output current and the set current value. The performance determination unit determines that the current output stability of the current magnetic field transmitter is substandard based on the steady-state error value being greater than the preset steady-state error value.

3. The magnetic field transmitter based on AI-automatic adjustment of control parameters according to claim 1, characterized in that, In response to the current output stability of the current magnetic field transmitter not meeting the standard, the data analysis unit is used to construct the current system state vector based on the steady-state error value, the rate of change of the steady-state error value, the output voltage value, the load impedance value, the ambient temperature value, and the H-bridge switching frequency value. The current system state vector is then input into a deep reinforcement learning model, which outputs the pre-adjustment amount of the PID control parameters.

4. The magnetic field transmitter based on AI-automatic adjustment of control parameters according to claim 3, characterized in that, The data analysis unit is also used to calculate the updated PID control parameters based on the pre-adjustment amount of the PID control parameters and the basic PID control parameters. Based on the updated PID control parameters, the output control quantity of the PID controller is determined by the PID control algorithm. Based on the output control quantity of the PID controller, the output control quantity of the DAC is determined by digital-to-analog conversion; wherein, the PID control parameters are proportional coefficient, integral coefficient, and derivative coefficient.

5. The magnetic field transmitter based on AI-automatic adjustment of control parameters according to claim 1, characterized in that, The duty cycle preset value of the PWM modulation signal of the switching power supply is determined based on the DAC output control quantity, the input voltage value of the switching power supply, and the preset power efficiency. The carrier frequency of the PWM modulation signal is determined based on the H-bridge switching frequency value and a preset multiple relationship; The power control unit generates the PWM modulation signal based on the duty cycle preset value and the carrier frequency; Based on the PWM modulation signal, the output voltage value of the switching power supply is determined by the power conversion circuit.

6. The magnetic field transmitter based on AI-automatic adjustment of control parameters according to claim 1, characterized in that, The compensation determination unit is used to determine, based on the voltage change rate being greater than a preset change rate, that feedforward compensation is needed for the duty cycle of the H-bridge drive signal.

7. The magnetic field transmitter based on AI-automatic adjustment of control parameters according to claim 6, characterized in that, In response to the need for feedforward compensation of the duty cycle of the H-bridge drive signal, The feedforward compensation unit is used to determine the current compensation coefficient based on the ratio of the voltage change rate to the preset change rate. The duty cycle of the reference H-bridge drive signal is compensated based on the current compensation coefficient to obtain the first target duty cycle.

8. The magnetic field transmitter based on AI-automatic adjustment of control parameters according to claim 1, characterized in that, The theoretical output current value is determined based on the output voltage value and load impedance value of the switching power supply. The absolute value of the deviation is determined based on the theoretical output current value and the set current value. The deviation analysis unit is used to determine that the pre-adjustment amount of the PID control parameter is unqualified based on the absolute value of the deviation being greater than a preset current deviation threshold.

9. The magnetic field transmitter based on AI-automatic adjustment of control parameters according to claim 8, characterized in that, In response to the failure of the pre-adjustment amount of the PID control parameter, the deviation analysis unit further includes a correction direction determination module; The correction direction determination module is used to determine the correction direction based on the comparison result between the theoretical output current value and the set current value.

10. The magnetic field transmitter based on AI-automatic adjustment of control parameters according to claim 8, characterized in that, The deviation analysis unit also includes a correction coefficient calculation module; The correction coefficient calculation module is used to determine the correction coefficient based on the ratio of the absolute value of the deviation to a preset current deviation threshold. The correction coefficient is positively correlated with the absolute value of the deviation; the larger the deviation, the greater the correction magnitude. Based on the correction direction and correction coefficient, the pre-adjustment amounts of the proportional coefficient, integral coefficient, and derivative coefficient are adjusted in the corresponding directions. The updated PID pre-adjustment value is sent to the data analysis unit, which replaces the original pre-adjustment value for subsequent PID parameter updates.

Images