Dynamic power consumption optimization control method based on bidirectional DC-DC power management chip

Abstract

The invention discloses a dynamic power consumption optimization control method based on a bidirectional DC-DC power management chip, relates to the technical field of power consumption control, and solves the problem that power consumption cannot be fully optimized and controlled due to the fact that the power consumption is easily influenced by temperature. The dynamic power consumption optimization of the bidirectional DC-DC power supply under all working conditions is realized through closed-loop control of'precise modeling-dynamic optimization-adaptive regulation and control-gradient convergence ', and the defects of a traditional control method in the aspects of loss balance, parameter drift, dynamic response and the like are effectively overcome; an efficient and reliable power supply management technical scheme is provided for the fields of new energy automobiles, energy storage systems and the like, and when temperature rise and device parameter drift are caused by long-time operation of the circuit, a duty ratio iteration adjustment mechanism is triggered. Experimental data show that the mechanism can control the loss amplification in a high-temperature environment to be within 5%, which is reduced by 15%-20% compared with a traditional scheme.

Description

Technical Field

[0001] The present invention relates to the technical field of power consumption control, and specifically to a dynamic power consumption optimization control method based on a bidirectional DC-DC power management chip. Background Art

[0002] Bidirectional DC-DC power management chips are widely used in new energy vehicles, energy storage systems, uninterruptible power supplies (UPS), etc., and need to achieve efficient energy conversion under bidirectional power transfer (charging / discharging mode).

[0003] The dynamic losses of bidirectional DC-DC mainly come from switching losses and conduction losses. Traditional fixed duty cycle control methods are difficult to balance the relationship between the two under wide input and output voltage ranges and dynamic load conditions. For example, fixed frequency PWM control at light load will cause a significant increase in the proportion of switching losses, while conduction losses caused by on-resistance at heavy load will become the main source of energy consumption.

[0004] In addition, the device parameters drift with temperature (such as the temperature drift of MOSFET on-resistance and the change of parasitic capacitance), which further exacerbates the complexity of loss control. Traditional steady-state models cannot track such changes in real time, resulting in a decrease in system efficiency or even overheating failure.

[0005] Traditional control methods are difficult to meet the soft-switching conditions in real time (such as the mismatch between dead time and resonance parameters), resulting in soft-switching failure or duty cycle loss, weakening the loss optimization effect; at the same time, existing optimization algorithms (such as PID control) have problems such as response lag and being easily trapped in local optima, and cannot achieve dynamic power consumption optimization at the microsecond level.

[0006] Existing bidirectional DC-DC power management technologies have significant deficiencies in dynamic power consumption optimization, multi-mode adaptive control, parameter drift compensation, etc. There is an urgent need for a new control method that can integrate loss modeling, dynamic optimization, and real-time feedback to improve the energy conversion efficiency and system reliability under all working conditions. Summary of the Invention

[0007] In view of the deficiencies of the prior art, the present invention provides a dynamic power consumption optimization control method based on a bidirectional DC-DC power management chip, which solves the problem that power loss is easily affected by temperature and cannot be fully optimized and controlled.

[0008] To achieve the above objectives, the present invention is realized through the following technical solutions: A dynamic power consumption optimization control method based on a bidirectional DC-DC power management chip, comprising the following steps: Collect the operating parameters and preset parameters associated with the power supply, and confirm the switching losses and conduction losses associated with the power supply, so as to confirm the total loss of this power supply circuit. The specific method is as follows: Record the collected switching frequency as f1, calibrate the preset device parasitic capacitance as R1, and synchronously calibrate the collected voltage and current as V1 and I1 respectively. Use: Kh = f1 × (R1 × V1 2 + 0.16 × V1 × I1 × t) to confirm the switching loss Kh, where t is the ratio of the switching rise time to the fall time; Calibrate the preset device on-resistance as R2, calibrate the determined duty cycle as D, and use: Ch = I1 2 × R2 × D to lock the conduction loss Ch, where the method of obtaining the duty cycle is different in different circuits; According to the confirmed total loss of the power supply circuit, adjust the output voltage and input current associated with the circuit. According to the specific adjustment process, confirm the change process of the total loss data, and lock the optimal duty cycle from it. Among them, the specific method of locking the optimal duty cycle is: Confirm the duty cycle D associated with the total loss of the power supply circuit, and lock the change interval [D - Y1, D + Y1] belonging to this duty cycle D according to the preset value Y1, where Y1 is the preset change step; Change the output voltage and input current associated with the circuit in real time. According to the specific change process, confirm the changed duty cycle ZB i in real time, where i represents different change processes, and the changed ZB i ∈ [D - Y1, D + Y1]; According to the changed output voltage and input current, use the same processing method as in step one to record the total loss associated with the current change process, and select the minimum value from the recorded total losses. Record the duty cycle ZB i associated with the minimum value as the optimal duty cycle and execute; After the optimal duty cycle is set, use the built-in ADC module to collect the input voltage, output voltage, input current, and output current associated with the circuit in real time, and generate the circuit loss associated with the corresponding moment in real time according to the specific process of the real-time collection. According to the change of the circuit loss value, determine the duty cycle control signal and the control direction.

[0009] Among them, the specific method of determining the duty cycle control signal is: According to the collected input voltage and input current, confirm the input power associated with the current power supply circuit. Then, according to the collected output voltage and output current, confirm the output power associated with the current power supply circuit. According to the determined input power and output power, use: actual loss = input power - output power; Record the actual loss confirmed at the current moment as the reference loss, and re-confirm the actual loss associated with the next moment. If the actual losses at subsequent moments continuously increase based on the reference loss and the increase duration exceeds 10 seconds, it indicates that the power circuit loss is abnormal, and a duty cycle control signal is generated. The duty cycle is re-adjusted according to the duty cycle control signal. Otherwise, no duty cycle control signal is generated; Among them, the specific method for determining the duty cycle control direction is as follows: Based on the different actual losses confirmed at different moments, confirm the difference value of the actual losses generated between adjacent moments. The difference value = the actual loss at the later moment - the actual loss at the previous moment, and according to the time relationship, sort and confirm the difference value sequence table for several groups of associated values before and after; Increase the duty cycle value by 0.01 per unit moment, and observe whether the difference values generated in the subsequent five unit moments in the difference value sequence table gradually decrease. If so, record the upward change direction as the duty cycle control direction. If not, record the downward change direction as the duty cycle control direction.

[0010] Preferably, the preset parameters include device parasitic capacitance and device on-resistance, and the operating parameters include switching frequency, voltage, current, and duty cycle.

[0011] Preferably, the method for obtaining the duty cycle includes: If the converter used in the current circuit is a Buck converter (step-down type): D = output voltage ÷ input current; If the converter used is a Boost converter (step-up type): D = 1 - (input current ÷ output voltage); If the converter used is a Buck-Boost converter: D = -[(input voltage ÷ input current) × (1 - D)], where the specific output voltage and input current can be directly determined in the circuit.

[0012] According to the determined duty cycle control signal and control direction, iteratively adjust the duty cycle until the adjusted comprehensive characteristic is lower than the convergence threshold, and then confirm and execute the final duty cycle. The specific method is as follows: Q1. Record the currently associated duty cycle as D old , and based on the input power and output power associated with the current moment, confirm the loss Ps associated with the current duty cycle D old ; Q2. Based on the confirmed control direction and the set small perturbation value of 0.01, confirm the loss Pn associated with the duty cycle after applying the small perturbation control; Q3. According to D oldThe associated loss Ps and the associated loss Pn, lock the gradient change value TD, where TD = (Pn - Ps) ÷ 0.01; Q4. According to the determined gradient change value, lock the latest duty cycle: D new = D old -α × TD to confirm the latest duty cycle D associated currently new , and 0.05 ≤ D new ≤ 0.95, where α is a preset learning rate; Q5. Record the confirmed latest duty cycle as D again old , and repeat Q1 - Q4 until the determined |TD| < Y2, then stop, and execute the duty cycle associated in the stop process, take it as the final duty cycle, and execute to effectively control and optimize the loss associated in the circuit, where Y2 is a preset convergence threshold.

[0013] The present invention provides a dynamic power consumption optimization control method for a bidirectional DC - DC power management chip. Compared with the prior art, it has the following beneficial effects: The present invention traverses multiple duty cycle samples during the dynamic adjustment of the output voltage and input current, and locks the optimal duty cycle based on the minimum loss; this method breaks through the limitations of traditional steady - state formulas and can achieve loss minimization within a wide load range (such as 0.1 - 1 times the rated current); When the circuit temperature rises and device parameters drift (such as the increase of MOSFET on - resistance) due to long - term operation, trigger the duty cycle iterative adjustment mechanism. Experimental data show that this mechanism can control the loss increase in a high - temperature environment within 5%, reducing by 15% - 20% compared with the traditional scheme; Through the closed - loop control of "accurate modeling - dynamic optimization - adaptive regulation - gradient convergence", the dynamic power consumption optimization of the bidirectional DC - DC power supply under all working conditions is realized, effectively solving the defects of traditional control methods in aspects such as loss balance, parameter drift, and dynamic response, and providing an efficient and reliable power management technical solution for fields such as new energy vehicles and energy storage systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 It is a schematic flow chart of the method of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0015] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0016] Please refer to Figure 1 , this application provides a dynamic power consumption optimization control method for a bidirectional DC-DC power management chip. The dynamic losses associated with the bidirectional DC-DC power management chip mainly come from switching losses, conduction losses, control circuit losses, and dynamic response losses. Among them, the control circuit losses and dynamic response losses are very small, and the specific losses that seriously affect are switching losses and conduction losses respectively; Based on the bidirectional DC-DC power management chip, this application dynamically optimizes the duty cycle, combines a multi-mode adaptive control strategy, and dynamically adjusts the control parameters by real-time monitoring the input and output voltages, currents, and load status to achieve efficient energy conversion and power consumption optimization under all working conditions. Its specific control system mainly includes a converter main circuit, a sensor module, a control chip (such as a DSP or FPGA), a drive circuit, and an auxiliary power supply. Among them, the sensor module collects the input voltage, output voltage, load current, and chip temperature in real time; Its optimization control method specifically includes the following steps: Step 1: Confirm the total circuit loss: Collect the operating parameters and preset parameters associated with the power supply, and confirm the switching losses and conduction losses associated with the power supply to confirm the total loss of this power supply circuit. Specifically, the preset parameters include device parasitic capacitance and device on-resistance, and such values are preset by relevant operators in advance according to the operating devices in the circuit. Its operating parameters include switching frequency, voltage, current, and duty cycle, etc. Among them, the duty cycle has different calculation methods when different converters are used in the circuit; Among them, the specific method for confirming the total loss of the power supply circuit is: Record the collected switching frequency as f1, calibrate the preset device parasitic capacitance as R1, and synchronously calibrate the collected voltage and current as V1 and I1 respectively. Use: Kh = f1×(R1×V1 2 + 0.16×V1×I1×t) to confirm the switching loss Kh, where t is the ratio of the switching rise time to the fall time; Calibrate the preset device on-resistance as R2, and calibrate the determined duty cycle as D. Use: Ch = I1 2×R2×D locks the conduction loss Ch. Among them, the duty cycle is obtained in different ways in different circuits, and its specific obtaining method is preset by relevant personnel in advance. If the converter used in the current circuit is a Buck converter (step-down type): D = output voltage ÷ input current. If the converter used is a Boost converter (step-up type): D = 1 - (input current ÷ output voltage). If the converter used is a Buck-Boost converter: D = -[(input voltage ÷ input current) × (1 - D)], where the specific output voltage and input current can be directly determined in the circuit; Adopt: Total loss = Kh + Ch to lock the total loss of this power supply circuit in the current operating state; Step 2: According to the confirmed total loss of the power supply circuit, adjust the output voltage and input current associated with the circuit. According to the specific adjustment process, confirm the change process of the total loss data, and lock the optimal duty cycle from it. Specifically, the corresponding duty cycle is associated with the output voltage and input current existing in the circuit. Therefore, during the adjustment process of the output voltage and input current, the associated voltage value and current value inside will also change accordingly, and then the total loss generated will also change. Therefore, from the corresponding change process, the optimal duty cycle can be locked; Among them, the specific method for locking the optimal duty cycle is: Confirm the duty cycle D associated with the total loss of the power supply circuit, and lock the change interval [D - Y1, D + Y1] belonging to this duty cycle D according to the preset value Y1, where Y1 is the preset change step, generally taking a value of 0.05, and is determined by the operator in advance; Change the output voltage and input current associated with the circuit in real time, and confirm the changed duty cycle ZB in real time according to the specific change process i , where i represents different change processes, and the changed ZB i ∈[D - Y1, D + Y1]; According to the changed output voltage and input current, adopt the same processing method as in Step 1 to record the total loss associated with the current change process, and select the minimum value from the recorded total losses. The duty cycle ZB associated with the minimum value iIt is denoted as the optimal duty cycle and executed. Specifically, in the current circuit, there are corresponding operating parameters and preset parameters. Then, the losses associated with the circuit in the current state and the corresponding duty cycle can be determined. Subsequently, by gradually readjusting the numerical parameters associated with the duty cycle, according to the specific adjustment process, the optimal duty cycle can be locked in the current circuit, so that the current operating state reaches the state of minimum loss. In the subsequent operating process, since the longer the operating time, the greater the temperature impact, by adjusting the duty cycle, the overall effect of loss optimization control can be further achieved; Step 3: After setting the optimal duty cycle, use the built-in ADC module to collect the input voltage, output voltage, input current, and output current associated with the circuit in real time. According to the specific process of real-time collection, generate the circuit loss associated with the corresponding moment in real time, and determine the control signal and control direction of the duty cycle according to the change of the circuit loss value. Specifically, as the circuit operates for a long time, a large amount of high temperature will be generated inside the circuit, and the high temperature will affect the operating state of related devices, which will cause the loss to increase again. Therefore, in order to re-regulate such abnormal loss situations, it is necessary to re-regulate the duty cycle to ensure the optimization process of the duty cycle control, so that the corresponding duty cycle is in the optimal duty cycle state; The method for determining the duty cycle control signal is: Based on the collected input voltage and input current, confirm the input power associated with the current power supply circuit. Then, based on the collected output voltage and output current, confirm the output power associated with the current power supply circuit. According to the determined input power and output power, use: actual loss = input power - output power; Denote the actual loss confirmed at the current moment as the reference loss, and re-confirm the actual loss associated with the next moment. If the actual losses in subsequent moments continue to increase based on the reference loss and the increase duration exceeds 10 seconds, it means that the power supply circuit loss is abnormal, and a duty cycle control signal is generated. According to the duty cycle control signal, the duty cycle is re-adjusted. Otherwise, no duty cycle control signal is generated; The method for determining the duty cycle control direction is: Based on the different actual losses confirmed at different moments, confirm the difference value of the actual losses generated between adjacent moments. The difference value = actual loss at the later moment - actual loss at the previous moment. According to the time relationship, sort and confirm the difference value sequence table associated with the front and back; Increase the duty cycle value by 0.01 per unit time, and observe whether the difference values generated in the subsequent five groups of unit times in the difference value sequence table gradually become smaller. If so, record the upward change direction as the duty cycle control direction. If not, record the downward change direction as the duty cycle control direction; Specifically, in the actual regulation and control process, there are only two regulation directions, either upward or downward. Among these two regulation directions, there is one direction that can reduce the corresponding loss characteristics. Then, based on the change in the loss characteristics brought about by the corresponding regulation process, the regulation direction of the duty cycle is determined and executed to achieve the optimal regulation and control effect.

[0017] Step 4: According to the determined regulation signal and regulation direction of the duty cycle, iteratively adjust the duty cycle until the adjusted comprehensive characteristic is lower than the convergence threshold, and then confirm the final duty cycle and execute it. The specific method for confirming the final duty cycle is as follows: Q1: Record the currently associated duty cycle as D old (that is, the duty cycle finally confirmed in the process of Step 3, and then this type of duty cycle needs to be corrected). Based on the input power and output power associated with the current moment, confirm the current duty cycle D old and the associated loss Ps; Q2: Based on the determined regulation direction and the set small perturbation (0.01), if the direction is upward, continuously increase by 0.01; if the direction is downward, continuously decrease by 0.01, and confirm the loss Pn associated with the duty cycle after applying the small perturbation regulation. Q3: According to D old and the associated loss Ps and the associated loss Pn, lock the gradient change value TD, where TD = (Pn - Ps) ÷ 0.01; Q4: According to the determined gradient change value, lock the latest duty cycle: D new = D old - α × TD to confirm the currently associated latest duty cycle D new , and 0.05 ≤ D new ≤ 0.95. If the calculated D new exceeds 0.95 or is lower than 0.05, then when it exceeds, D new takes 0.95; if D new is lower than 0.05, D new takes 0.05. Here, α is the preset learning rate, which is determined by the operator in advance according to experience, and α generally takes the value of 0.05; Q5: Re - record the confirmed latest duty cycle as D old , and repeat Q1 - Q4 until |TD| < Y2 is determined and stop, and execute the duty cycle associated with the stop process, take it as the final duty cycle and execute it, effectively control and optimize the loss associated with the circuit. Here, Y2 is the preset convergence threshold, which is determined by the operator in advance according to experience, and Y2 generally takes the value of 0.001.

[0018] Specifically, a detailed example of calculating the optimal duty cycle based on the gradient descent method Example background and parameter settings Suppose we apply the gradient descent method to calculate the optimal duty cycle in a bidirectional DC-DC on-vehicle charging system of an electric vehicle; in the charging mode, this system converts the 400V DC power input from the charging pile into 300V to charge the on-vehicle battery, with a rated load current of 20A; the bidirectional DC-DC converter adopts a synchronous Buck topology, and the power device is a silicon carbide MOSFET; II. Detailed explanation of specific implementation steps 1. Initialization When the system starts or the load changes significantly, first set the initial duty cycle D0 = 0.75 and the learning rate α = 0.05; the control chip will generate a PWM signal according to this initial duty cycle to drive the MOSFET.

[0019] 2. Sampling and calculating losses The control chip uses the built-in ADC module to collect the input voltage, output voltage, input current, and output current in real time. At the duty cycle D0 = 0.75, assume the actual data collected is: Input power = 8200W; Output power = 6000W; Current loss = 8200 - 6000 = 2200W; 3. Perturb the duty cycle Based on the current duty cycle D0 = 0.75, apply a duty cycle perturbation amount (0.01) to obtain a new duty cycle of 0.76; the control chip updates the duty cycle of the PWM signal to make the MOSFET work according to the new duty cycle.

[0020] 4. Calculate the gradient At the duty cycle = 0.76, collect the input and output voltage and current through the ADC module again to calculate the new loss: Input power = 8150W; Output power = 6000W; New loss = 2150W; Determine the associated TD = -5000 according to the gradient calculation formula; 5. Update the duty cycle Based on the preset learning rate and the associated TD, confirm the new duty cycle. Since this duty cycle exceeds 0.95, D new takes the value of 0.95.

[0021] 6. Repeat the iteration The control chip uses the updated duty cycle as the D for the next iteration old, repeat the steps; as the iteration progresses, the loss gradually decreases and the duty cycle continuously approaches the optimal value; assume that after multiple iterations, the absolute value of the calculated gradient is lower than 0.001, at this time the algorithm is considered to have converged, and the current duty cycle is the optimal duty cycle.

[0022] III. Example Results and Analysis Through the iterative calculation of the above gradient descent method, the finally obtained optimal duty cycle may be between 0.85 - 0.92 (the specific value depends on the actual loss change); compared with the initial duty cycle of 0.75, the total loss of the system at the optimal duty cycle is significantly reduced, and the energy conversion efficiency is improved; in practical applications, the influence of temperature on MOSFET parameters also needs to be considered, and the loss model is monitored and corrected in real time through a temperature sensor to ensure that the gradient descent method can always accurately calculate the optimal duty cycle.

[0023] The above example fully demonstrates the whole process of calculating the optimal duty cycle by the gradient descent method.

[0024] Some of the data in the above formula are numerically calculated after removing their dimensions, and the content not described in detail in this specification belongs to the prior art well-known to those skilled in the art.

[0025] The above embodiments are only used to illustrate the technical method of the present invention and not to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical method of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A dynamic power consumption optimization control method for a bidirectional DC-DC power management chip, characterized in that It includes the following steps: Collect the operating parameters and preset parameters associated with the power supply, and confirm the switching losses and conduction losses associated with the power supply to confirm the total losses of this power supply circuit; Adjust the output voltage and input current associated with the circuit according to the confirmed total losses of the power supply circuit. According to the specific adjustment process, confirm the change process of the total loss data, and lock the optimal duty cycle from it; After the optimal duty cycle is set, use the built-in ADC module to collect the input voltage, output voltage, input current, and output current associated with the circuit in real time. According to the specific process of real-time collection, generate the circuit losses associated with the corresponding moment in real time, and determine the control signal and control direction of the duty cycle according to the change of the circuit loss value; Iteratively adjust the duty cycle according to the determined control signal and control direction of the duty cycle until the adjusted comprehensive characteristic is lower than the convergence threshold, and confirm and execute the final duty cycle.

2. The dynamic power consumption optimization control method for a bidirectional DC-DC power management chip according to claim 1, wherein The preset parameters include device parasitic capacitance and device on-resistance, and the operating parameters include switching frequency, voltage, current, and duty cycle.

3. A dynamic power consumption optimization control method for a bidirectional DC-DC power management chip according to claim 1, characterized in that The specific method for confirming the total losses of the power supply circuit is: Record the collected switching frequency as f1, calibrate the preset device parasitic capacitance as R1, and simultaneously calibrate the collected voltage and current as V1 and I1 respectively. Use: Kh = f1×(R1×V1 2 + 0.16×V1×I1×t) to confirm the switching loss Kh, where t is the ratio of the switching rise time to the fall time; Calibrate the preset device on-resistance as R2, and calibrate the determined duty cycle as D. Use: Ch = I1 2 ×R2×D to lock the conduction loss Ch. Among different circuits, the methods for obtaining the duty cycle are different.

4. A dynamic power consumption optimization control method for a bidirectional DC-DC power management chip according to claim 3, characterized in that The method for obtaining the duty cycle includes: If the converter used in the current circuit is a Buck converter (step-down type): D = output voltage ÷ input current; If the converter used is a Boost converter (step-up type): D = 1 - (input current ÷ output voltage); If the converter used is a Buck-Boost converter: D = -[(input voltage ÷ input current) × (1 - D)], where the specific output voltage and input current can be directly determined in the circuit.

5. A dynamic power consumption optimization control method for a bidirectional DC-DC power management chip according to claim 1, characterized in that, The specific method for locking the optimal duty cycle is: Confirm the duty cycle D associated with the total losses of the power supply circuit, and lock the change interval [D - Y1, D + Y1] belonging to this duty cycle D according to the preset value Y1, where Y1 is the preset change step; The output voltage and input current associated in the circuit are changed in real time. According to the specific change process, the changed duty cycle ZB is confirmed in real time i , where i represents different change processes, and the changed ZB i ∈[D - Y1, D + Y1]; According to the changed output voltage and input current, record the total loss associated with the current change process in the same processing manner as in Step 1, select the minimum value from the recorded total losses, and use the duty cycle ZB associated with the minimum value i as the optimal duty cycle and execute it.

6. A dynamic power consumption optimization control method for a bidirectional DC-DC power management chip according to claim 1, characterized in that, The specific method for determining the duty cycle control signal is: According to the collected input voltage and input current, confirm the input power associated with the current power supply circuit, and then according to the collected output voltage and output current, confirm the output power associated with the current power supply circuit. According to the determined input power and output power, use: actual loss = input power - output power; Record the actual loss confirmed at the current moment as the reference loss, and re-confirm the actual loss associated with the next moment. If the actual losses in subsequent moments continue to increase based on the reference loss and the increase duration exceeds 10 seconds, it means that the power supply circuit loss is abnormal, and a duty cycle control signal is generated to re-adjust the duty cycle. Otherwise, no duty cycle control signal is generated.

7. A dynamic power consumption optimization control method for a bidirectional DC-DC power management chip according to claim 6, characterized in that The specific method for determining the duty cycle control direction is: Based on the different actual losses confirmed at different times, confirm the actual loss difference value generated between adjacent times. The difference value = the actual loss at the latter time - the actual loss at the previous time, and according to the time relationship, sort and confirm the difference value sequence table for several groups of related difference values before and after; Increase the duty cycle value upward by 0.01 for each unit time, and observe whether the difference values generated by the subsequent five unit times in the difference value sequence table gradually decrease. If so, record the upward change direction as the duty cycle regulation direction. If not, record the downward change direction as the duty cycle regulation direction.

8. A dynamic power consumption optimization control method for a bidirectional DC-DC power management chip according to claim 1, characterized in that The specific method for confirming the final duty cycle is as follows: Q1. Record the duty cycle associated currently as D old , and based on the input power and output power associated at the current moment, confirm the current duty cycle D old and the associated loss Ps; Q2. Based on the confirmed regulation direction and the set small perturbation, take a value of 0.01, and confirm the loss Pn associated with the duty cycle after applying the small perturbation regulation; Q3. Based on D old Lock the gradient change value TD according to the associated loss Ps and the associated loss Pn, where TD = (Pn - Ps) ÷ 0.01; Q4. Lock the latest duty cycle D according to the determined gradient change value new =D old -α × TD to confirm the latest duty cycle D associated currently new , and 0.05 ≤ D new ≤ 0.95, where α is a preset learning rate; Q5. Re - record the latest confirmed duty cycle as D old , and repeat the execution of Q1 - Q4 until the determined |TD| < Y2, then stop and execute the duty cycle associated with the stop process, take it as the final duty cycle, and execute to effectively control and optimize the losses associated with the circuit, where Y2 is a preset convergence threshold.

Images