Portable AC-DC voltage converter

Abstract

The invention discloses a portable AC-DC voltage converter which comprises an AC-DC input loop, a charging module, a lithium battery, an inversion module, an isolation boosting transformer, a central processing unit control unit, a touch screen and an output relay. The input loop charges the lithium battery through the charging module, the lithium battery provides a power supply for the isolation boosting transformer through the inversion module, the boosting transformer outputs multi-stage alternating voltage and distributes the multi-stage alternating voltage to a load through the output relay, and the central processor control unit acquires parameters such as voltage and current in real time based on neural network prediction and fuzzy control rules. The SPWM signal modulation duty ratio of the inversion module is dynamically adjusted, and stable voltage output is achieved. Multi-power-supply compatibility is achieved through the alternating current and direct current self-adaptive charging module, the dynamic response capacity and control precision are improved through neural network predictive control, the voltage stability is ensured through fuzzy control rules, the functions of overload protection, short circuit protection and real-time adjustment are achieved, and the application range is wide.

Description

Technical Field

[0001] The invention belongs to the technical field of power electronic engineering, and in particular relates to a portable AC / DC voltage converter. Background Art

[0002] At present, when accepting the energy meter and load control terminal (or plant terminal) at the outdoor customer's place or the newly built substation, the on-site metering device has no power supply and no working power supply, which makes the load control terminal, energy meter and metering automation master station unable to communicate and debug. It is necessary to wait for the line or voltage transformer to be energized and the voltage transformer to provide 57.7V or 100V voltage to power the energy meter or load control terminal before communication debugging can be carried out. As a result, debugging cannot be carried out in advance, and the work that can be completed in a day has to be completed several times. Sometimes, although there is an AC 220V power supply nearby, the working power supply of the load control terminal is 57.7V or 100V voltage, and some energy meters have no auxiliary power supply. It is necessary to provide AC 220V, 57.7V or 100V voltage to carry out the inspection and acceptance of the energy meter and load control terminal, and realize the normal communication debugging of the load control terminal, energy meter and metering automation master station system. In addition, when checking whether the voltage secondary circuit is broken, wired incorrectly, in poor contact, or has phase sequence errors, some load management terminals and electric energy meters are powered by PT secondary voltage, and the voltage transformer is shut down (or fails) for maintenance. If there is no temporary emergency power supply, communication will be interrupted. When testing the transformer of a newly built substation, construction equipment (such as welding machines, etc.) has a great impact on the test power supply, and power outages may occur at the construction site at any time. If there is no temporary emergency power supply, the test will be interrupted. If the newly shipped electric energy meters and load control terminals cannot provide 57.7V or 100V AC voltage in the office or at the work site, work such as electric energy meter and terminal communication debugging cannot be carried out. Summary of the invention

[0003] The technical problem to be solved by the present invention is to provide a portable AC / DC voltage converter which realizes multi-power compatibility through an AC / DC adaptive charging module, has overload protection, short circuit protection and real-time adjustment functions, and is applicable to a wide range of applications.

[0004] The technical solution of the present invention is: a portable AC / DC voltage converter, comprising: an AC / DC input circuit, a charging module, a lithium battery, an inverter module, an isolation boost transformer, a central processing unit control unit, a touch screen, and an output relay, wherein the AC / DC input circuit is connected to the charging module, the charging module is connected to the lithium battery, the lithium battery is connected to the inverter module, the inverter module is respectively connected to the isolation boost transformer and the central processing unit control unit, the isolation boost transformer is connected to the output relay, the central processing unit control unit is connected to the touch screen, and the output relay comprises a plurality of AC output circuits.

[0005] Furthermore, the AC / DC input circuit includes terminals and an air switch.

[0006] Furthermore, the charging module has a built-in AC / DC adaptive charging module, and the AC / DC adaptive charging module is a HEG75050 AC / DC hybrid conversion module.

[0007] Furthermore, the inverter module includes: an H-bridge power circuit, a gate drive circuit and an SPWM signal generating circuit, the SPWM signal generating circuit is connected to the gate drive circuit, the gate drive circuit is connected to the H-bridge power circuit, and the H-bridge power circuit is connected to the isolation step-up transformer.

[0008] Furthermore, the central processing unit control unit includes: a data acquisition module, a neural network modeling module, a prediction control module, an inverter control module and an optimization algorithm module, the input end of the data acquisition module is connected to the AC / DC input circuit, the lithium battery and the inverter module, the output end of the data acquisition module is connected to the neural network modeling module and the prediction control module, the neural network modeling module is connected to the prediction control module, the prediction control module is connected to the optimization algorithm module, the optimization algorithm module is connected to the inverter control module, and the inverter control module is connected to the inverter module.

[0009] Furthermore, the output relay includes three AC output circuits, namely 220V, 100V, and 57.7V AC output circuits. The 220V AC output circuit uses a three-hole socket and terminal output, and the 100V and 57.7V AC output circuits use a single-pole air switch control terminal or test jack output.

[0010] Furthermore, it also includes: a shell and accessories, the shell is a hollow box structure and is made of insulating material, the AC and DC input circuits, charging module, lithium battery, inverter module, isolation step-up transformer, central processing unit control unit, touch screen, and output relay are arranged in the shell, and a carrying handle is provided on the shell, and the accessories are output wire packages and are connected to the shell.

[0011] An optimization control method for a portable AC / DC voltage converter comprises the following steps:

[0012] Step 1: Real-time monitoring and sampling: The sampling signals include: input voltage V in , output voltage V out and output current I out , the sampling signal is transmitted to the central processing unit control unit (6) through the ADC module;

[0013] Step 2, deviation calculation: Calculate the output voltage deviation and deviation change rate. The calculation formulas for the voltage deviation and deviation change rate are:

[0014] e(t)=Vref -V out (t)

[0015] △e(t)=e(t)-e(t-1),

[0016] Among them, V ref Indicates the target voltage value, V out (t) represents the output voltage at the current moment, e(t) represents the current voltage deviation, e(t-1) represents the voltage deviation at the previous moment, and Δe(t) represents the deviation change rate;

[0017] Step 3: Design of fuzzy control rules, including the following steps:

[0018] S1. According to the output voltage deviation and the deviation change rate, a fuzzy control rule is designed to adjust the SPWM signal modulation duty cycle D of the inverter module: when e(t) is large and Δe(t) increases, D is reduced; when e(t) is small and Δe(t) decreases, D is increased;

[0019] S2. Define fuzzy sets:

[0020] e(t): {negative large NB, negative small NS, zero Z, positive small PS, positive large PB},

[0021] Δe(t): {negative large NB, negative small NS, zero Z, positive small PS, positive large PB}, D: {decrease large BD, decrease small SD, unchanged Z, increase small SI, increase large BI};

[0022] Step 4, fuzzy reasoning: If e(t) = PB and Δe(t) = PB, then D = BD, that is, reduce the duty cycle to reduce the voltage output; if e(t) = Z and Δe(t) = Z, then D = Z, that is, keep the duty cycle unchanged;

[0023] Step 5, defuzzification calculation: According to the result of fuzzy reasoning, the adjusted duty cycle is calculated by weighted average method. The adjusted duty cycle D new The calculation formula is:

[0024]

[0025] Among them, μ i represents fuzzy membership, D i represents the specific value in the fuzzy set, D new Indicates the SPWM signal modulation duty cycle of the new inverter module; μ i (x) represents the membership of input x to fuzzy set i, and its value range is between [0,1]. x represents the input variable, c represents the central value of the fuzzy set, and σ represents the standard deviation, i.e., the width of the fuzzy set.

[0026] Step 6: Output control: Adjust the modulation ratio D of the SPWM signal of the inverter module new， The switching state of the H-bridge power switch of the inverter module is controlled to achieve dynamic adjustment of the output voltage.

[0027] An intelligent control method for a portable AC / DC voltage converter comprises the following steps:

[0028] Step 1: Data collection and preprocessing: The collected signals include: input voltage V in , output voltage V out and output current I out The sampling signal is transmitted to the central processing unit (6) through the ADC module, and the sampling period T is set. s ;

[0029] Step 2: State modeling: The system dynamic behavior is described by the following state equation:

[0030] X(k+1)=f NN (X(k), U(k)),

[0031]

[0032] U(k)=[D(k)],

[0033] Among them, X(k+1) represents the system state vector at the next moment, X(k) represents the current system state vector, and V out (k) represents the current output voltage, I out (k) represents the current output current, U(k) represents the control input vector, D(k) represents the SPWM signal modulation duty cycle of the current inverter module, and f NN Represents the mapping function of the neural network model;

[0034] Step 3: Neural network training, including the following steps:

[0035] S1. Input data: Collect sample data from system operation and construct training sets (X(k), U(k) and X(k+1));

[0036] S2. Neural network model structure, including: input layer: accepts X(k) and U(k); hidden layer: several fully connected layers, activation function is ReLU; output layer: predicts the next state X(k+1);

[0037] S3. Objective function: minimize the prediction error L:

[0038]

[0039] Among them, X(k) represents the current system state vector, U(k) represents the control input vector, and fNN Represents the mapping function of the neural network model, X true (k+1) represents the actual system state vector at the next moment, and N represents the number of training samples;

[0040] Step 4, predictive control, includes the following steps:

[0041] S4. Prediction formula: For a given current state X(k) and control input U(k+j):

[0042] X(k+j+1)=f NN (X(k+1),U(k+j)),j=0,1,...,N p -1,

[0043] Where X(k+j+1) represents the state vector of the system at the k+j+1th moment, f NN represents the mapping function of the neural network model, X(k+j) represents the state vector of the system at the k+jth moment, U(k+j) represents the control input vector at the k+jth moment, k represents the discrete time index of the current moment, j represents the index variable in the prediction time domain, and N p Indicates the prediction time domain length, that is, the number of time steps predicted by the system, N p The larger it is, the farther into the future the system predicts;

[0044] S5. Optimization objective function: Define the optimization objective J to balance voltage stability and control input energy consumption:

[0045]

[0046] ΔU(k+j)=U(k+j)-U(k+j-1),

[0047] Among them, X ref represents the target state, i.e., the target output voltage, ΔU(k+j) represents the control input change at the k+jth moment, X(k+j) represents the state vector of the system at the k+jth moment, U(k+j) represents the control input vector at the k+jth moment, N p represents the length of the prediction time domain, k represents the discrete time index of the current moment, j represents the index variable of the prediction time domain, and λ represents the energy consumption weight coefficient;

[0048] S6. Constraints, including: output voltage limit and duty cycle limit, the expression of the output voltage limit is:

[0049]

[0050] Among them, V out (k+j) represents the predicted output voltage at the k+jth moment, Indicates the minimum allowable value of the output voltage, represents the maximum allowable value of the output voltage, k represents the discrete time index of the current moment, and j represents the index variable of the prediction time domain;

[0051] The duty cycle limit expression is:

[0052] D min ≤D(k+j)≤D max ,

[0053] Where D(k+j) represents the SPWM signal modulation duty cycle of the inverter module at the k+jth moment, D min Indicates the minimum value of the duty cycle, D max represents the maximum value of the duty cycle, k represents the discrete time index of the current moment, and j represents the index variable of the prediction time domain;

[0054] S7. Optimization solution: Use gradient descent or nonlinear programming algorithm to solve the optimal control input sequence {U(k), U(k+1),…, U(k+N p -1)};

[0055] Step 5, real-time control, including the following steps:

[0056] S8, collect system state X(k), and obtain the control input U(k-1) at the previous moment;

[0057] S9. Predicting future states: Using neural networks f NN Predict the future p The state at a moment;

[0058] S10, Optimization control: Calculate the optimal control input U(k) at the current moment according to the objective function J and the constraints;

[0059] S11, input control signal: transmit the optimal duty cycle D(k) to the SPWM signal generation module of the inverter module, and adjust the output voltage of the inverter module in real time;

[0060] S12, repeat cycle: each cycle T s Update the status and repeat steps S8 to S11.

[0061] Beneficial effects of the present invention:

[0062] 1. Strong portability and adaptability: The outer shell is designed as a hollow insulating box with a carrying handle. It is light and easy to carry, suitable for on-site operation, supports AC and DC power input, and the adaptive charging module can automatically identify and adapt to a variety of power types to adapt to complex environments;

[0063] 2. Multi-voltage level output: Provides multiple AC voltage levels such as 220V, 100V, 57.7V, etc. to meet the needs of different equipment, equipped with multiple output circuits, including three-hole sockets and terminals, flexible to use;

[0064] 3. High-efficiency power conversion: Built-in HEG75050 AC / DC hybrid conversion module, high conversion efficiency, low loss, using SPWM signal modulation inverter technology, good output voltage waveform quality, low harmonic distortion;

[0065] 4. Intelligent control and dynamic adjustment: The central processing unit control unit integrates data acquisition, neural network modeling, predictive control and optimization algorithm modules. By real-time sampling of input voltage, output voltage and current, it dynamically adjusts the SPWM modulation duty cycle of the inverter module to achieve precise control and rapid response;

[0066] 5. Strong voltage stability: The fuzzy control rule design adjusts the duty cycle to reduce the error and deviation change rate, ensuring the output voltage is stable, and can maintain the target voltage range even when the load changes dynamically; 6. Complete safety protection function: Equipped with air switch and short-circuit protection circuit, it can effectively protect the equipment and load, support overload protection, overvoltage protection and overcurrent protection, and improve the safety of system operation;

[0067] 7. High intelligence level: Adopt neural network prediction model to model the dynamic behavior of the system, combine with predictive control to optimize output, achieve balance between voltage stability and energy consumption, and improve operation efficiency;

[0068] 8. Real-time monitoring and user interaction: Equipped with a touch screen, it can display the input and output voltage, current and equipment status in real time. Users can adjust the output parameters through the touch screen, which is convenient for on-site operation and management;

[0069] 9. Wide range of application scenarios: Applicable to power system testing, industrial equipment power supply, new energy power conversion, emergency power supply and other fields, providing reliable power support, especially suitable for complex on-site environments and mobile operations; 10. Improve energy utilization efficiency: Combined with neural networks and optimization algorithms, the system can intelligently allocate power resources, reduce unnecessary energy consumption, and the efficient coordinated operation of lithium batteries and inverter modules further improves overall efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] Figure 1 It is a structural schematic diagram of a portable AC / DC voltage converter of the present invention.

[0071] In the figure: 1-AC / DC input circuit, 2-charging module, 3-lithium battery, 4-inverter module, 5-isolation step-up transformer, 6-CPU control unit, 7-touch screen, 8-output relay. DETAILED DESCRIPTION

[0072] like Figure 1 As shown, a portable AC / DC voltage converter includes: an AC / DC input circuit 1, a charging module 2, a lithium battery 3, an inverter module 4, an isolation boost transformer 5, a central processing unit control unit 6, a touch screen 7, and an output relay 8. The AC / DC input circuit 1 is connected to the charging module 2, the charging module 2 is connected to the lithium battery 3, the lithium battery 3 is connected to the inverter module 4, the inverter module 4 is respectively connected to the isolation boost transformer 5 and the central processing unit control unit 6, the isolation boost transformer 5 is connected to the output relay 8, the central processing unit control unit 6 is connected to the touch screen 7, and the output relay 8 includes a plurality of AC output circuits.

[0073] Preferably, the AC / DC input circuit 1 includes terminals and an air switch.

[0074] Preferably, the charging module 2 has a built-in AC / DC adaptive charging module, and the AC / DC adaptive charging module is a HEG75050 AC / DC hybrid conversion module.

[0075] Preferably, the inverter module 4 includes: an H-bridge power circuit, a gate drive circuit and an SPWM signal generating circuit, the SPWM signal generating circuit is connected to the gate drive circuit, the gate drive circuit is connected to the H-bridge power circuit, and the H-bridge power circuit is connected to the isolation step-up transformer 5.

[0076] Preferably, the central processing unit control unit 6 includes: a data acquisition module, a neural network modeling module, a prediction control module, an inverter control module and an optimization algorithm module, the input end of the data acquisition module is connected to the AC / DC input circuit 1, the lithium battery 3 and the inverter module 4, the output end of the data acquisition module is connected to the neural network modeling module and the prediction control module, the neural network modeling module is connected to the prediction control module, the prediction control module is connected to the optimization algorithm module, the optimization algorithm module is connected to the inverter control module, and the inverter control module is connected to the inverter module 4.

[0077] Preferably, the output relay 8 includes three AC output circuits, namely 220V, 100V, and 57.7V AC output circuits. The 220V AC output circuit uses a three-hole socket and terminal output, and the 100V and 57.7V AC output circuits use a single-pole air switch control terminal or test jack output.

[0078] Preferably, it also includes: a shell and accessories, the shell is a hollow box structure and is made of insulating material, the AC / DC input circuit 1, charging module 2, lithium battery 3, inverter module 4, isolation step-up transformer 5, central processing unit control unit 6, touch screen 7, output relay 8 are arranged in the shell, the shell is provided with a carrying handle, and the accessory is an output wire package and is connected to the shell.

[0079] An optimization control method for a portable AC / DC voltage converter comprises the following steps:

[0080] Step 1: Real-time monitoring and sampling: The sampling signals include: input voltage V in , output voltage V out and output current I out , the sampling signal is transmitted to the central processing unit control unit (6) through the ADC module;

[0081] Step 2, deviation calculation: Calculate the output voltage deviation and deviation change rate. The calculation formulas for the voltage deviation and deviation change rate are:

[0082] e(t)=V ref -V out (t)

[0083] △e(t)=e(t)-e(t-1),

[0084] Among them, V ref Indicates the target voltage value, V out (t) represents the output voltage at the current moment, e(t) represents the current voltage deviation, e(t-1) represents the voltage deviation at the previous moment, and Δe(t) represents the deviation change rate;

[0085] Step 3: Design of fuzzy control rules, including the following steps:

[0086] S1. According to the output voltage deviation and the deviation change rate, a fuzzy control rule is designed to adjust the SPWM signal modulation duty cycle D of the inverter module: when e(t) is large and Δe(t) increases, D is reduced; when e(t) is small and Δe(t) decreases, D is increased;

[0087] S2. Define fuzzy sets:

[0088] e(t): {negative large NB, negative small NS, zero Z, positive small PS, positive large PB},

[0089] Δe(t): {negative large NB, negative small NS, zero Z, positive small PS, positive large PB},

[0090] D: {reduce large BD, reduce small SD, keep Z unchanged, increase small SI, increase large BI};

[0091] The fuzzy rule table is designed as follows:

[0092] e(t)\△e(t) NB NS Z PS PB NB BI BI SI Z SD NS BI SI Z SD BD Z SI Z Z Z SD PS SI Z SD BD BD PB Z SD SD BD BD

[0093] Step 4, fuzzy reasoning: If e(t) = PB and Δe(t) = PB, then D = BD, that is, reduce the duty cycle to reduce the voltage output; if e(t) = Z and Δe(t) = Z, then D = Z, that is, keep the duty cycle unchanged;

[0094] Step 5, defuzzification calculation: According to the result of fuzzy reasoning, the adjusted duty cycle is calculated by weighted average method. The adjusted duty cycle D new The calculation formula is:

[0095]

[0096] Among them, μ i represents fuzzy membership, D i represents the specific value in the fuzzy set, D new Indicates the SPWM signal modulation duty cycle of the new inverter module; μ i (x) represents the membership of input x to fuzzy set i, and its value range is between [0,1]. x represents the input variable, c represents the central value of the fuzzy set, and σ represents the standard deviation, i.e., the width of the fuzzy set.

[0097] Step 6: Output control: Adjust the modulation ratio D of the SPWM signal of the inverter module new， The switching state of the H-bridge power switch of the inverter module is controlled to achieve dynamic adjustment of the output voltage.

[0098] An intelligent control method for a portable AC / DC voltage converter comprises the following steps:

[0099] Step 1: Data collection and preprocessing: The collected signals include: input voltage V in , output voltage V out and output current I out The sampling signal is transmitted to the central processing unit (6) through the ADC module, and the sampling period T is set. s ;

[0100] Step 2: State modeling: The system dynamic behavior is described by the following state equation:

[0101] X(k+1)=f NN (X(k), U(k)),

[0102]

[0103] U(k)=[D(k)],

[0104] Among them, X(k+1) represents the system state vector at the next moment, X(k) represents the current system state vector, and V out (k) represents the current output voltage, I out (k) represents the current output current, U(k) represents the control input vector, D(k) represents the SPWM signal modulation duty cycle of the current inverter module, and f NN Represents the mapping function of the neural network model;

[0105] Step 3: Neural network training, including the following steps:

[0106] S1. Input data: Collect sample data from system operation and construct training sets (X(k), U(k) and X(k+1));

[0107] S2. Neural network model structure, including: input layer: accepts X(k) and U(k); hidden layer: several fully connected layers, activation function is ReLU; output layer: predicts the next state X(k+1);

[0108] S3. Objective function: minimize the prediction error L:

[0109]

[0110] Among them, X(k) represents the current system state vector, U(k) represents the control input vector, and f NN Represents the mapping function of the neural network model, X true (k+1) represents the actual system state vector at the next moment, and N represents the number of training samples;

[0111] Step 4, predictive control, includes the following steps:

[0112] S4. Prediction formula: For a given current state X(k) and control input U(k+j):

[0113] X(k+j+1)=f NN (X(k+j),U(k+j)),j=0,1,...,N p -1,

[0114] Where X(k+j+1) represents the state vector of the system at the k+j+1th moment, f NN represents the mapping function of the neural network model, X(k+j) represents the state vector of the system at the k+jth moment, U(k+j) represents the control input vector at the k+jth moment, k represents the discrete time index of the current moment, j represents the index variable in the prediction time domain, and N p Indicates the prediction time domain length, that is, the number of time steps predicted by the system, N p The larger it is, the farther into the future the system predicts;

[0115] S5. Optimization objective function: Define the optimization objective J to balance voltage stability and control input energy consumption:

[0116]

[0117] ΔU(k+j)=U(k+j)-U(k+j-1),

[0118] Among them, X ref represents the target state, i.e., the target output voltage, ΔU(k+j) represents the control input change at the k+jth moment, X(k+j) represents the state vector of the system at the k+jth moment, U(k+j) represents the control input vector at the k+jth moment, N p represents the length of the prediction time domain, k represents the discrete time index of the current moment, j represents the index variable of the prediction time domain, and λ represents the energy consumption weight coefficient;

[0119] S6. Constraints, including: output voltage limit and duty cycle limit, the expression of the output voltage limit is:

[0120]

[0121] Among them, V out (k+j) represents the predicted output voltage at the k+jth moment, Indicates the minimum allowable value of the output voltage, represents the maximum allowable value of the output voltage, k represents the discrete time index of the current moment, and j represents the index variable of the prediction time domain;

[0122] The duty cycle limit expression is:

[0123] D min ≤D(k+j)≤D max ,

[0124] Where D(k+j) represents the SPWM signal modulation duty cycle of the inverter module at the k+jth moment, D min Indicates the minimum value of the duty cycle, D max represents the maximum value of the duty cycle, k represents the discrete time index of the current moment, and j represents the index variable of the prediction time domain;

[0125] S7. Optimization solution: Use gradient descent or nonlinear programming algorithm to solve the optimal control input sequence {U(k), U(k+1),…, U(k+N p -1)};

[0126] Step 5, real-time control, including the following steps:

[0127] S8, collect system state X(k), and obtain the control input U(k-1) at the previous moment;

[0128] S9. Predicting future states: Using neural networks f NN Predict the future p The state at a moment;

[0129] S10, Optimization control: Calculate the optimal control input U(k) at the current moment according to the objective function J and the constraints;

[0130] S11, input control signal: transmit the optimal duty cycle D(k) to the SPWM signal generation module of the inverter module, and adjust the output voltage of the inverter module in real time;

[0131] S12, repeat cycle: each cycle T s Update the status and repeat steps S8 to S11.

[0132] The embodiments of the present invention are described in detail above with reference to the accompanying drawings, but the present invention is not limited to the above embodiments and various changes can be made thereto within the knowledge scope of those skilled in the art.

Claims

1. A portable AC / DC voltage converter, characterized in that: include: An AC / DC input circuit (1), a charging module (2), a lithium battery (3), an inverter module (4), an isolation step-up transformer (5), a central processing unit control unit (6), a touch screen (7), and an output relay (8), wherein the AC / DC input circuit (1) is connected to the charging module (2), the charging module (2) is connected to the lithium battery (3), the lithium battery (3) is connected to the inverter module (4), the inverter module (4) is respectively connected to the isolation step-up transformer (5) and the central processing unit control unit (6), the isolation step-up transformer (5) is connected to the output relay (8), the central processing unit control unit (6) is connected to the touch screen (7), and the output relay (8) comprises a plurality of AC output circuits.

2. A portable AC / DC voltage converter according to claim 1, characterized in that: The AC / DC input circuit (1) comprises terminals and an air switch.

3. A portable AC / DC voltage converter according to claim 1, characterized in that: The charging module (2) has a built-in AC / DC adaptive charging module, and the AC / DC adaptive charging module is a HEG75050 AC / DC hybrid conversion module.

4. A portable AC / DC voltage converter according to claim 1, characterized in that: The inverter module (4) comprises: an H-bridge power circuit, a gate drive circuit and an SPWM signal generating circuit, wherein the SPWM signal generating circuit is connected to the gate drive circuit, the gate drive circuit is connected to the H-bridge power circuit, and the H-bridge power circuit is connected to an isolation step-up transformer (5).

5. A portable AC / DC voltage converter according to claim 1, characterized in that: The central processing unit control unit (6) comprises: a data acquisition module, a neural network modeling module, a prediction control module, an inverter control module and an optimization algorithm module, wherein the input end of the data acquisition module is connected to the AC / DC input circuit (1), the lithium battery (3) and the inverter module (4), the output end of the data acquisition module is connected to the neural network modeling module and the prediction control module, the neural network modeling module is connected to the prediction control module, the prediction control module is connected to the optimization algorithm module, the optimization algorithm module is connected to the inverter control module, and the inverter control module is connected to the inverter module (4).

6. A portable AC / DC voltage converter according to claim 1, characterized in that: The output relay (8) includes three AC output circuits, namely 220V, 100V, and 57.7V AC output circuits. The 220V AC output circuit uses a three-hole socket and terminal output, and the 100V and 57.7V AC output circuits use a single-pole air switch control terminal or a test jack output.

7. A portable AC / DC voltage converter according to claim 1, characterized in that: Also includes: A shell and accessories, wherein the shell is a hollow box structure and is made of insulating material, the AC / DC input circuit (1), the charging module (2), the lithium battery (3), the inverter module (4), the isolation step-up transformer (5), the central processing unit control unit (6), the touch screen (7), and the output relay (8) are arranged in the shell, the shell is provided with a carrying handle, and the accessories are output wire packages and are connected to the shell.

8. An optimization control method for a portable AC / DC voltage converter according to claim 1, characterized in that: The following steps are involved: Step 1: Real-time monitoring and sampling: The sampling signals include: input voltage V in , output voltage V out and output current I out , the sampling signal is transmitted to the central processing unit control unit (6) through the ADC module; Step 2, deviation calculation: Calculate the output voltage deviation and deviation change rate. The calculation formulas for the voltage deviation and deviation change rate are: e(t)=V ref -V out (t) Δe(t)=e(t)-e(t-1), Among them, V ref Indicates the target voltage value, V out (t) represents the output voltage at the current moment, e(t) represents the current voltage deviation, e(t-1) represents the voltage deviation at the previous moment, and Δe(t) represents the deviation change rate; Step 3: Design of fuzzy control rules, including the following steps: S1. According to the output voltage deviation and the deviation change rate, a fuzzy control rule is designed to adjust the SPWM signal modulation duty cycle D of the inverter module: when e(t) is large and Δe(t) increases, D is reduced; when e(t) is small and Δe(t) decreases, D is increased; S2. Define fuzzy sets: e(t): {negative large NB, negative small NS, zero Z, positive small PS, positive large PB}, Δe(t): {negative large NB, negative small NS, zero Z, positive small PS, positive large PB}, D: {reduce large BD, reduce small SD, keep Z unchanged, increase small SI, increase large BI}; Step 4, fuzzy reasoning: If e(t) = PB and Δe(t) = PB, then D = BD, that is, reduce the duty cycle to reduce the voltage output; if e(t) = Z and Δe(t) = Z, then D = Z, that is, keep the duty cycle unchanged; Step 5, defuzzification calculation: According to the result of fuzzy reasoning, the adjusted duty cycle is calculated by weighted average method. The adjusted duty cycle D new The calculation formula is: Among them, μ i represents fuzzy membership, D i represents the specific value in the fuzzy set, D new Indicates the SPWM signal modulation duty cycle of the new inverter module; μ i (x) represents the membership of input x to fuzzy set i, and its value range is between [0,1]. x represents the input variable, c represents the central value of the fuzzy set, and σ represents the standard deviation, i.e., the width of the fuzzy set. Step 6: Output control: Adjust the modulation ratio D of the SPWM signal of the inverter module new , control the switching state of the H-bridge power switch of the inverter module, thereby realizing dynamic adjustment of the output voltage.

9. An intelligent control method for a portable AC / DC voltage converter according to claim 1, characterized in that: The following steps are involved: Step 1: Data collection and preprocessing: The collected signals include: input voltage V in , output voltage V out and output current I out The sampling signal is transmitted to the central processing unit (6) through the ADC module, and the sampling period T is set. s ; Step 2: State modeling: The system dynamic behavior is described by the following state equation: X(k+1)=f NN (X(k),U(k)), U(k)=[[D(k)], Among them, X(k+1) represents the system state vector at the next moment, X(k) represents the current system state vector, and V out (k) represents the current output voltage, I out (k) represents the current output current, U(k) represents the control input vector, D(k) represents the SPWM signal modulation duty cycle of the current inverter module, and f NN Represents the mapping function of the neural network model; Step 3: Neural network training, including the following steps: S1. Input data: Collect sample data from system operation and construct training sets (X(k), U(k) and X(k+1)); S2. Neural network model structure, including: input layer: accepts X(k) and U(k); hidden layer: several fully connected layers, activation function is ReLU; output layer: predicts the next state X(k+1); S3. Objective function: minimize the prediction error L: Among them, X(k) represents the current system state vector, U(k) represents the control input vector, fNN represents the mapping function of the neural network model, and X true (k+1) represents the actual system state vector at the next moment, and N represents the number of training samples; Step 4, predictive control, includes the following steps: S4. Prediction formula: For a given current state X(k) and control input U(k+j): X(k+j+1)=f NN (X(k+j),U(k+j)), j=0,1,...,N p -1, Where X(k+j+1) represents the state vector of the system at the k+j+1th moment, fNN represents the mapping function of the neural network model, X(k+j) represents the state vector of the system at the k+jth moment, U(k+j) represents the control input vector at the k+jth moment, k represents the discrete time index at the current moment, j represents the index variable in the prediction time domain, and N p Indicates the prediction time domain length, that is, the number of time steps predicted by the system, N p The larger it is, the farther into the future the system predicts; S5. Optimization objective function: Define the optimization objective J to balance voltage stability and control input energy consumption: △U(k+j)=U(k+j)-U(k+j-1), Among them, X ref represents the target state, i.e., the target output voltage, ΔU(k+j) represents the control input change at the k+jth moment, X(k+j) represents the state vector of the system at the k+jth moment, U(k+j) represents the control input vector at the k+jth moment, N p represents the length of the prediction time domain, k represents the discrete time index of the current moment, j represents the index variable of the prediction time domain, and λ represents the energy consumption weight coefficient; S6. Constraints, including: output voltage limit and duty cycle limit, the expression of the output voltage limit is: Among them, V out (k+j) represents the predicted output voltage at the k+jth moment, Indicates the minimum allowable value of the output voltage, represents the maximum allowable value of the output voltage, k represents the discrete time index of the current moment, and j represents the index variable of the prediction time domain; The duty cycle limit expression is: D min ≤D(k+j)≤D max , Where D(k+j) represents the SPWM signal modulation duty cycle of the inverter module at the k+jth moment, D min Indicates the minimum value of the duty cycle, D max represents the maximum value of the duty cycle, k represents the discrete time index of the current moment, and j represents the index variable of the prediction time domain; S7. Optimization solution: Use gradient descent or nonlinear programming algorithm to solve the optimal control input sequence {U(k), U(k+1),…, U(k+N p -1)}; Step 5, real-time control, including the following steps: S8, collect system state X(k), and obtain the control input U(k-1) at the previous moment; S9. Predicting future states: Using neural networks f NN Predict the future p The state at a moment; S10, Optimization control: Calculate the optimal control input U(k) at the current moment according to the objective function J and the constraints; S11, input control signal: transmit the optimal duty cycle D(k) to the SPWM signal generation module of the inverter module, and adjust the output voltage of the inverter module in real time; S12, repeat cycle: each cycle T s Update the status and repeat steps S8 to S11.

Images