Energy-saving power supply regulating system

Abstract

The application discloses an energy-saving power supply adjusting system, which adopts integrated sensors to collect multiple environmental parameters in real time and transmit the parameters to a PC controller; the PC controller is provided with a BP neural network controller; a PFC circuit converts alternating current into direct current input into a DC-AC circuit by PWM closed-loop feedback adjustment based on the power supply voltage amplitude by using a first full-bridge topology circuit; the DC-AC circuit converts the direct current into alternating current by inputting multiple PWM signals into a second full-bridge topology circuit based on the power supply voltage amplitude and the power supply voltage frequency, converts the alternating current into sinusoidal alternating current by using multiple filter inductors, and receives the sinusoidal alternating current at a load end; the multiple switch tubes used in the two full-bridge topology circuits are all SiC Mosfets. The application directly links the actual parameters affecting the power supply output with the power supply voltage amplitude and frequency, thereby adjusting in real time according to the actual situation, achieving the energy-saving effect, and greatly improving the inverter frequency by using SiC Mosfets, which also effectively saves energy.

Description

Technical Field

[0001] This invention relates to the field of power supply regulation, and more specifically to an energy-saving power supply regulation system. Background Technology

[0002] In shopping malls, the main power supply undergoes multiple processes, including voltage reduction, inversion, and rectification, to provide 220-230V AC power to various sub-areas. Power supply-related equipment often uses a unified scheduling method, providing the same power to different sub-areas. If power supply adjustments are needed, each device's power output must be manually modified. This approach presents at least three problems:

[0003] 1) Adjustments are not timely enough and cannot be made in real time;

[0004] 2) Manual operation is cumbersome, inefficient, and prone to errors, resulting in insufficient accuracy;

[0005] 3) During the adjustment process, especially when there are many devices that need to be adjusted, there is a huge loss of power during the voltage reduction, inversion and rectification conversion of the power supply, resulting in serious power waste. Summary of the Invention

[0006] The purpose of this invention is to solve the above three problems by proposing an energy-saving power supply regulation system. This system utilizes sensor registers to receive environmental parameters in real time, enabling real-time adjustment of the power supply. It employs a BP neural network to accurately calculate the target value required by the load and sets up a control interface to intuitively display the data, allowing for timely reporting and repair of errors. Furthermore, it uses a PC controller to perform real-time control of the PFC circuit and DC-AC circuit, and utilizes SiC MOSFETs as switching transistors for conversion, thereby effectively reducing power loss during the conversion process and improving conversion efficiency and performance.

[0007] This was achieved through the following technical solutions:

[0008] An energy-saving power supply regulation system employs a power supply module to supply power to the load. Between the power supply module and the load, there are: an integrated sensor, a PC controller, a PFC circuit, and a DC-AC circuit. The integrated sensor collects multiple environmental parameters in real time and transmits them to the PC controller. The PC controller contains a BP neural network controller, whose BP neural network consists of an input layer, a hidden layer, and an output layer. The input layer receives multiple environmental parameters and the system's set target value as input, and then performs nonlinear transformation processing by the hidden layer. The output layer outputs the real-time required power supply voltage amplitude and frequency for the load. The PFC circuit receives the AC power input from the power supply module and then supplies power to the load in real time. Based on the required supply voltage amplitude, the first full-bridge topology circuit built into the PFC circuit performs PWM closed-loop feedback regulation on the AC power, converting the AC power into DC power input to the DC-AC circuit. After receiving the DC power, the DC-AC circuit uses the real-time required supply voltage amplitude and frequency at the load end as the basis, and uses the second full-bridge topology circuit built into the DC-AC circuit to regulate multiple PWM signals on the DC power input, converting the DC power into AC power. Then, multiple filter inductors built into the DC-AC circuit further convert the converted AC power into sinusoidal AC power, which is received at the load end. Both the first and second full-bridge topologies use multiple switching transistors, and each switching transistor is a SiC MOSFET.

[0009] Using a PC controller as the host computer for control, multiple environmental parameters collected by integrated sensors can be used to adjust the relevant outputs of the PFC circuit and DC-AC circuit in real time. This ensures that the frequency and amplitude of the final sinusoidal AC output match the load in real time, reducing energy loss and achieving energy saving. At the same time, using SiC MOSFETs as switching transistors to convert DC to AC can effectively reduce energy loss during the conversion process, achieving energy saving and improving conversion efficiency.

[0010] Preferably, the input layer uses j input neurons, where j is the number of multiple environmental parameters plus 1, and the output function used by the input layer is... x represents a neuron in the input layer, and the superscript (1) indicates the input layer; the hidden layer receives the output function of the input layer, and the input function used by the hidden layer is... Both k and i are positive integers not greater than j, where k represents a neuron in the hidden layer. The weight coefficients are given by the superscript (2), which indicates the hidden layer. The output function used by the hidden layer is... t is a positive integer greater than k; the output layer receives the output function of the hidden layer, and the input function used by the output layer is... The superscript (3) indicates the output layer, q represents a neuron in the output layer, and the value of q is 1 or 2; the output function of the output layer is ,in, This indicates the real-time supply voltage amplitude required by the load. This represents the real-time supply voltage frequency required by the load. The BP neural network, during calculation, exhibits strong adaptability, fault tolerance, and high computation speed. It can establish a direct relationship between the actual physical quantities affecting the power supply output and the amplitude and frequency of the supply voltage, thereby adjusting the power supply output in real time according to the actual situation on site, achieving energy saving, speed, and accurate calculation.

[0011] Preferably, when performing the corresponding function calculation process in the input layer, hidden layer, and output layer, the steepest descent method is used to adjust the parameters of the corresponding function calculation process. The steepest descent method can accelerate the convergence speed, thereby meeting the requirements of real-time adjustment; it can also minimize the error and improve accuracy.

[0012] Preferably, the PFC circuit includes at least an RC filter circuit, a first full-bridge topology circuit, an LC voltage regulator and filter circuit, and a voltage detection circuit connected in sequence. After the RC filter circuit filters the AC power received by the PFC circuit, multiple switches in the first full-bridge topology circuit input multiple PWM signals to the filtered AC power, converting the filtered AC power into DC power. The LC voltage regulator and filter circuit then further regulates and filters the DC power. The voltage detection circuit then detects the regulated and filtered DC power and adjusts the multiple PWM signals input to the multiple switches in the first full-bridge topology circuit based on the detection results, forming a PWM closed-loop feedback regulation. The voltage detection circuit, when performing power factor correction in the PFC circuit, forms a closed-loop feedback, effectively regulating the signals of multiple switches and ensuring accuracy.

[0013] Preferably, a current detection circuit is also provided between the first full-bridge topology circuit and the LC voltage regulator and filter circuit; when the current detected by the current detection circuit exceeds the maximum current limit of any switching transistor in the PFC circuit, the PFC circuit is turned off. The current detection circuit effectively protects the circuit, preventing excessive current from burning out some components, especially preventing the switching transistors from burning out.

[0014] Preferably, the system is equipped with a control interface, which includes multiple buttons, a temperature control interface, and a power control interface. The buttons control the system's operating status, the temperature control interface displays multiple environmental parameters and allows input of target values, and the power control interface displays multiple actual output values ​​at the load end. This control interface allows for intuitive control and observation, enabling monitoring of the system's operation and ensuring the accuracy and effectiveness of power supply adjustments.

[0015] Preferably, after receiving the DC power output from the PFC circuit at the input terminal of the DC-AC circuit, it is regulated by an RC voltage regulator circuit. Then, a second full-bridge topology circuit composed of multiple parallel switching transistors inputs multiple PWM signals to the regulated DC power, converting it into AC power. After passing through a sinusoidal filter inductor L1, two relays are placed on the transmission path of the converted AC power to control the switching of the DC-AC circuit. The converted AC power then passes through sinusoidal filter inductors L2 and L3, converting it into sinusoidal AC power before transmitting it to the load terminal. By utilizing the second full-bridge topology circuit and sinusoidal filter inductors, the DC power is converted into sinusoidal AC power, facilitating system use.

[0016] Preferably, the multiple environmental parameters include at least indoor temperature, outdoor temperature, outdoor humidity, and the number of people in the mall. Integrating sensors to receive multiple environmental parameters facilitates the subsequent generation of correct circuit outputs based on these parameters, improving the accuracy of the conversion.

[0017] Preferably, the system control method includes the following steps: S1, the system starts supplying power, and after the PC controller initializes the program, it acquires multiple environmental parameters and sets target values; S2, based on the multiple environmental parameters and set target values ​​in step S1, the PC controller's built-in BP neural network controller calculates the required power supply voltage amplitude and frequency for the load; S3, it determines whether the actual parameters corresponding to the required power supply voltage amplitude and frequency for the load have reached the set target values. If yes, control stops; if not, it returns to step S1 where the PC controller acquires multiple environmental parameters and sets target values, and repeats the loop until the actual system parameters reach the set target values, at which point control stops. The goal is to ensure the system's actual parameters reach the set target values, and even if unsuccessful the first time, the loop can be repeated multiple times to guarantee that the load operates in its optimal state.

[0018] The beneficial effects of this invention compared to the prior art are:

[0019] The technical solution of this invention uses a PC controller as the host computer for control. It can collect multiple environmental parameters from integrated sensors and adjust the relevant outputs of the PFC circuit and DC-AC circuit in real time, so that the frequency and amplitude of the final output sinusoidal AC power are matched with the load in real time, thereby effectively reducing energy loss and achieving energy saving. At the same time, SiC MOSFETs are used as inverter switches for conversion, which can effectively reduce the power loss during the conversion process, achieving energy saving and improving conversion efficiency. In addition, a BP neural network is used to calculate the output required by the load, which can quickly and accurately obtain the results required by the load, improving efficiency and accuracy. Attached Figure Description

[0020] Figure 1 This is a structural diagram of an energy-saving power supply regulation system;

[0021] Figure 2 This is a schematic diagram of a PFC circuit in an energy-saving power supply regulation system.

[0022] Figure 3 This is a schematic diagram of the structure of a BP neural network in an energy-saving power supply regulation system.

[0023] Figure 4 This is a schematic diagram of the control interface in an energy-saving power supply regulation system.

[0024] Figure 5 This is a circuit diagram of the voltage regulation input drive terminal of a DC-AC circuit in an energy-saving power supply regulation system.

[0025] Figure 6 This is a circuit diagram of the power supply output drive terminal of a DC-AC circuit in an energy-saving power supply regulation system.

[0026] Figure 7 This is a flowchart of a control method for an energy-saving power supply regulation system. Detailed Implementation

[0027] The technical solutions of the present invention will now be described in detail with reference to the accompanying drawings. Example

[0028] like Figure 1 The diagram shows the structure of an energy-saving power supply regulation system. An integrated sensor, PC controller, PFC circuit, and DC-AC circuit are set between the power supply module and the load end to complete the conversion of AC power to DC power and then to sinusoidal AC power.

[0029] In energy-saving power supply regulation systems, integrated sensors collect multiple environmental parameters in real time, including at least indoor temperature, outdoor temperature, outdoor humidity, and the number of people in the shopping mall. These parameters reflect the actual situation and demand of the current scenario, allowing for subsequent adjustments to the system. The integrated sensor consists of an infrared sensor, a temperature sensor, and a humidity sensor. The infrared sensor detects the current number of people in the shopping mall, the temperature sensor acquires the indoor and outdoor temperatures, and the humidity sensor acquires the outdoor humidity.

[0030] In this embodiment, the PC controller acts as the host computer in the system. It can collect multiple environmental parameters from integrated sensors and adjust the outputs of the PFC and DC-AC circuits in real time. This ensures that the frequency and amplitude of the final sinusoidal AC output match the load in real time, reducing energy loss and achieving energy conservation. The PC controller includes a BP neural network controller. When the system starts working, the BP neural network controller first initializes the program, clearing old data. Then, the BP neural network acquires the multiple environmental parameters transmitted in real time from the integrated sensors and, combined with the system's set target values, calculates the real-time required power supply voltage amplitude and frequency, as well as the corresponding optimal power, for the load.

[0031] like Figure 3 The diagram shows the structure of a BP neural network in an energy-saving power supply regulation system. The BP neural network uses a 5-7-2 structure, consisting of an input layer, hidden layers, and an output layer. The input layer has five neurons: X1, X2, X3, and X4 represent real-time monitoring values ​​of the dynamic environment at the power consumption location, corresponding to environmental parameters collected by integrated sensors; X5 is the set target value, which can be the set temperature. The output layer has two neurons: Y1 and Y2 represent the real-time demand voltage amplitude and frequency at the load end, respectively. The hidden layer has seven neurons, connecting the input and output layers. Through feature extraction, nonlinear mapping, learning, and memory functions, the neural network can handle complex nonlinear problems and has good generalization ability. The hidden layer's role is to transmit information from the input layer to the output layer through a series of nonlinear transformations, thereby achieving abstract representation and feature extraction of the input data. Specifically, each neuron in the hidden layer receives signals from the input layer, processes them through weights and activation functions, generates a new representation, and transmits it to the output layer. Choosing a hidden layer with 7 neurons provides sufficient model complexity to capture the complex relationships between input features and helps the network learn more abstract and high-level features. This helps the network better understand the patterns between input data and allows it to better fit the data during training, improving its generalization ability. The number of neurons in the hidden layer directly affects the network's representational power and its ability to model complex relationships. However, the choice of the number of neurons in the hidden layer also needs to consider the balance between model complexity and training efficiency to avoid overfitting or convergence difficulties during training. Therefore, choosing 7 neurons in the hidden layer can balance model complexity and training efficiency. The BP neural network has strong adaptability, fault tolerance, and fast computation speed. It can establish a direct relationship between the actual physical quantities affecting the power supply output and the amplitude and frequency of the supply voltage, thereby adjusting the power supply output in real time according to the actual situation on site, achieving energy saving, speed, and accurate calculation.

[0032] Specifically, assuming multiple environmental parameters are indoor temperature, outdoor temperature, outdoor humidity, and number of people in the mall, the input layer uses 5 input neurons, and the output function used by the input layer is... , j=1,2,3,4,5; x represents a neuron in the input layer, and the superscript (1) indicates the input layer;

[0033] The hidden layer receives the output function of the input layer, and the input function used by the hidden layer is... k = 1, 2, 3, 4, 5; k represents a neuron in the hidden layer. The weights are set as coefficients, and different weights are assigned to different items. The superscript (2) indicates the hidden layer. The output function used by the hidden layer is... , t=1,2,3,4,5,6,7;

[0034] The output layer receives the output function from the hidden layer, and the input function used by the output layer is... The superscript (3) indicates the output layer, q represents a neuron in the output layer, and the value of q is 1 or 2; the output function of the output layer is ,in, This indicates the real-time supply voltage amplitude required by the load. This indicates the frequency of the power supply voltage required by the load in real time.

[0035] In this embodiment, the sigmoid function is used as the activation function for all neurons, effectively representing the curve changes of signal transmission between different neurons. The sigmoid function introduces nonlinear transformations; without a nonlinear activation function, the linear transformations between different layers of a BP neural network would restrict the entire network to a linear model. Through the nonlinear characteristics of the sigmoid function, the neural network can learn and represent more complex functional relationships, thereby improving the network's expressive power. Furthermore, using the sigmoid function provides smoothness, as it is a smooth and continuously differentiable function, making network training more stable. In addition, the sigmoid function can suppress excessively large gradients because its derivative tends to approach zero in regions where the input is very large or very small. This helps suppress gradient explosion, thereby improving network stability, avoiding significant damage to various parts of the circuit, and ensuring the accuracy of BP neural network calculations.

[0036] In this embodiment, the BP neural network employs the steepest descent method, also known as gradient descent, which primarily serves two purposes: I) Optimizing weights and thresholds: The steepest descent method calculates the gradient of the error function with respect to the weights and thresholds to determine the direction of weight and threshold adjustment, thereby minimizing the error between the output and the set target value. This method can gradually adjust the parameters of the neural network to better fit the training data, thus improving the performance of the neural network; II) Accelerating convergence speed: The steepest descent method adjusts the weights and thresholds along the opposite direction of the error function gradient, continuously reducing the network error. Since this method directly adjusts parameters to reduce the error function, it can converge to a better solution relatively quickly, improving training efficiency.

[0037] like Figure 2 The diagram shows a PFC circuit in an energy-saving power supply regulation system. The PFC circuit includes at least an RC filter circuit, a first full-bridge topology circuit, a current detection circuit, an LC voltage regulator filter circuit, and a voltage detection circuit connected in sequence. After calculating the real-time required supply voltage amplitude and frequency at the load end using a BP neural network, the PFC circuit receives the AC power input from the power supply module, performs power factor correction on the AC power based on the target value, and converts the AC power to DC power. AC represents the AC power input from the power supply module. The AC power is filtered for EMI by the first filter circuit composed of R1', R2', R3', R4', R5', R6', and C4'. R1', R2', and R3' are connected in series as a single unit, and R4', R5', and R6' are also connected in series as a single unit. These two units, along with C4', form a parallel structure. EMI filtering is a method to suppress electromagnetic interference, capable of filtering out or attenuating high-frequency noise in the circuit and interference between different power lines, ensuring the accuracy of the system during operation.

[0038] After EMI filtering, a second filter circuit composed of L1', C2', C3', and C1' filters AC noise. Then, the signal enters the first full-bridge topology circuit composed of Q1', Q2', Q3', and Q4'. A voltage regulator and filter circuit is formed by L2', C5', and C6'. S1 and S2 are two relays whose operating states are synchronized to control whether the output is simultaneously on or off. The working process is as follows: Inductor L1' stores energy; Q1', Q2', Q3', and Q4' generate a PWM signal to regulate the AC power, resulting in DC power. After rectification and filtering by capacitors C5', C6', and L2', the DC power is output. Relays S1 and S2 control the output and stop of the DC power.

[0039] The PFC circuit also includes current detection and voltage detection circuits. P1 and P2 represent two nodes for current detection, and P3 and P4 represent two nodes for voltage detection. During PFC operation, the voltage detection circuit detects the DC current after regulation and filtering. Based on the voltage detection results, it adjusts multiple PWM signals input to multiple switches in the first full-bridge topology circuit, forming a PWM closed-loop feedback regulation to achieve active power factor correction (PFC). Specifically, by adjusting the switching frequencies of the two pairs of transistors Q1, Q4 and Q2, Q3, the output voltage can be adjusted. The switching frequencies of the two pairs of transistors Q1, Q4 and Q2, Q3 can be adjusted by the duty cycle of the PWM signal. Therefore, by adjusting the duty cycle of the PWM signal through the output voltage feedback value, active power factor correction can be achieved. Current detection is also performed during the adjustment process. Current sensors detect the current flowing through P1 and P3 respectively. If the current is too high, the PFC circuit needs to be shut down promptly to prevent overcurrent damage and circuit burnout, especially to avoid burning out the switching transistors.

[0040] In this embodiment, the multiple switching transistors in the first full-bridge topology circuit are all SiC MOSFETs, a special type of switching transistor capable of converting AC to DC power by inputting a PWM signal. Due to the characteristics of SiC material, SiC MOSFETs have low on-resistance and fast switching speed. They also possess excellent high-temperature performance and high blocking voltage, effectively reducing their own losses during energy conversion, thus improving energy conversion efficiency and performance.

[0041] After receiving the DC power rectified by the PFC circuit, the DC-AC circuit uses a second full-bridge topology to convert DC power to AC power based on the real-time supply voltage amplitude and frequency required by the load. The multiple switching transistors in the second full-bridge topology circuit are also SiC MOSFETs.

[0042] like Figure 5The diagram shows a schematic of the voltage regulator input drive terminal of a DC-AC circuit in an energy-saving power supply regulation system. P5 is connected to the L terminal of the PFC circuit output, and P6 is connected to the N terminal of the PFC circuit output. R1, R2, R3, and R4 are connected in series and then in parallel with three parallel capacitors C1, C2, and C3 to form a voltage regulator circuit for regulating the input voltage. Following the voltage regulator circuit, a second full-bridge topology circuit consisting of four switching transistors Q1, Q2, Q3, and Q4 is connected in parallel. Resistors R5, R6, and R7 and diode D1 form the drive circuit for switching transistor Q1; resistors R8, R9, and R10 and diode D3 form the drive circuit for switching transistor Q2; resistors R11, R12, and R13 and diode D2 form the drive circuit for switching transistor Q3; and resistors R14, R15, and R16 and diode D4 form the drive circuit for switching transistor Q4.

[0043] A set of Q1 PWM+ and Q1 PWM- signals are input to switch Q1, a set of Q2 PWM+ and Q2 PWM- signals are input to switch Q2, a set of Q3 PWM+ and Q3 PWM- signals are input to switch Q3, and a set of Q4 PWM+ and Q4 PWM- signals are input to switch Q4. These four sets of PWM signals can adjust the frequency of the corresponding switches, so that the four switches can be synchronized. At the same time, the DC power is converted into AC power.

[0044] like Figure 6 The diagram shown is a schematic of the power output drive terminal of a DC-AC circuit in an energy-saving power supply regulation system. Figure 6 The two nodes P7 and P8 in the text are... Figure 5 Nodes P7 and P8 in the diagram are used here for clarity of the image and explanation; therefore, the DC-AC circuit is divided into... Figure 5 and Figure 6 .

[0045] In the second full-bridge topology, a filter inductor L1 is connected between Q1 and Q2 to pin 2 of the current sensor. The current sensor can be an LTSR 6-NP current sensor, which can sense the magnetic field generated by the current in the circuit to measure the current magnitude, offering high measurement accuracy and sensitivity. A filter inductor L3 is connected between Q3 and Q4 to the output of the current sensor, connecting it to the power supply output driver. Parallel voltage-regulating capacitors C12 and C13 are connected between the 0V ground terminal and the 5V power supply terminal of the current sensor. Resistor R17 is a protective resistor connected between the 5V power supply and the 5V power supply terminal of the current sensor. Resistors R18, R19, R20, and R21, along with capacitor C6, are used to regulate the output of the current sensor, removing transient peak values ​​generated when the switching transistor is turned on and off. After regulating the output of the current sensor, inductors L2 and L3 are connected to both ends for filtering, converting the PWM signal output by the second full-bridge circuit into a sinusoidal AC signal. Inductor L2 is connected to relay KA1, and inductor L3 is connected to relay KA2. Diodes D5, D6, and D7 form the power supply circuit for relay KA1; diodes D8, D9, and D10 form the power supply circuit for relay KA2; resistors R23 and R24 form the Q5 drive circuit, and R24 is grounded to GND3. The Q5 drive circuit controls the on / off state of relays KA1 and KA2, thereby controlling the start and stop of the output. Capacitors C7, C8, C9, C10, and C11, along with resistor R22, are used to filter the outputs of relays KA1 and KA2. Capacitors C7 and C8 are connected in series and grounded to GND4. Relay KA2 can be controlled by the input OPRLY relay signal. Inductor L4, a common-mode inductor, is also used to filter the output signals of KA1 and KA2, eliminating common-mode interference between the two output signals. Capacitors C9 and C10 are also connected in series and grounded to GND5. Finally, the output signal from relay KA1 is connected to P9, and the output signal from relay KA2 is connected to P10. P9 and P10 are used to connect the load.

[0046] The power supply in the DC-AC circuit has the function of automatically adjusting the output load. The use of SiC MOSFETs significantly improves the inverter frequency, reduces energy loss, and improves the operating efficiency of the power supply. In addition, a dual closed-loop control circuit for current and voltage is adopted to stabilize the output voltage and further ensure accuracy and effectiveness.

[0047] In this embodiment, the target values ​​calculated by the BP neural network include the supply voltage frequency and the supply voltage amplitude. In the actual PFC circuit layout, the impedance of the PFC circuit is fixed, so these two values ​​also include corresponding PFC current and voltage data. If the error between the current detection circuit's detected value and the PFC current data exceeds 10%, or the error between the voltage detection circuit's detected value and the PFC voltage data exceeds 10%, the PFC circuit can be temporarily shut down for readjustment and inspection. Timely maintenance and adjustment when errors are not negligible effectively maintain the system, ensure its accuracy, and avoid additional energy waste due to erroneous data.

[0048] In this embodiment, a control method is also provided for the system, specifically as follows: When the system starts supplying power, after the PC controller initializes its program, it acquires multiple environmental parameters and sets target values. Based on these environmental parameters and target values, the PC controller's built-in BP neural network controller calculates the required power supply voltage amplitude and frequency for the load. It then determines whether the actual parameters corresponding to the required power supply voltage amplitude and frequency for the load have reached the set target values. If so, control stops; otherwise, it returns to the step of acquiring multiple environmental parameters and setting target values, and repeats the loop until the actual system parameters reach the set target values, at which point control stops.

[0049] like Figure 4 The diagram shows a control interface in an energy-saving power supply regulation system. The control interface includes multiple buttons, a temperature control interface, and a power control interface. The buttons control the system's operating status, the temperature control interface displays multiple environmental parameters and allows input of set temperatures, and the power control interface displays target values ​​at the load end and input voltage data from the power supply module. This control interface allows for intuitive control and observation, enabling monitoring of the system's operation and ensuring the accuracy and effectiveness of power supply adjustments.

[0050] In addition, the multiple buttons on the control interface can be set as run buttons, stop buttons, and fault buttons. These can be modified according to actual needs and equipment models; for example, a standby button can be added. Furthermore, an indicator light can be added, connected to a buzzer. The buzzer is connected to the output of the DC-AC circuit. When a fault occurs and personnel are on-site, the indicator light flashes to remind them to troubleshoot and repair the problem. When a fault occurs and personnel are not on-site, the indicator light flashes while the buzzer sounds an alarm, prompting personnel to come and troubleshoot the problem as soon as possible.

[0051] It should be noted that energy-saving power supply regulation systems can not only be applied to power supply regulation in shopping malls to improve the energy consumption and efficiency of air conditioning, but also to the charging and power management of vehicles. They can regulate the charging and power consumption process of vehicles in real time, saving energy for the vehicle's power and improving its range, thereby enhancing the user experience.

[0052] Specific application examples

[0053] Within a shopping mall area, an energy-saving power supply regulation system was used to conduct power supply regulation tests on an air conditioner with a rated power of 4500W. Each test lasted 3 hours. The mall's input bus voltage was 220-230V AC, while the air conditioner's actual bus voltage was 228.3V.

[0054] Integrated sensors collect real-time data on indoor and outdoor temperatures, outdoor humidity, and the number of people in the mall. The data is displayed on the control interface of the energy-saving power supply regulation system: indoor temperature 28℃, outdoor temperature 33℃, outdoor humidity 65%, and 93 people in the mall. Based on this data, the desired temperature to be lowered to 25℃ is set on the control interface, and the system begins operation.

[0055] The BP neural network controller in the PC controller initializes the program, clearing old data, and uses a 5-7-2 BP neural network structure to calculate the power supply voltage amplitude and frequency. The neural network input layer has five neurons: X1, X2, X3, and X4 represent indoor temperature, outdoor temperature, outdoor humidity, and the number of people in the mall, respectively; X5 represents the set temperature. The neural network output layer has two neurons: Y1 and Y2 represent the power supply voltage amplitude and frequency, respectively. The hidden layer uses seven neurons as the intermediate processing layer. The output function used by the input layer is... j=1,2,3,4,5; x represents the neuron in the input layer, and the superscript (1) indicates the input layer; the hidden layer receives the output function of the input layer, and the input function used by the hidden layer is k = 1, 2, 3, 4, 5; k represents a neuron in the hidden layer. The weights are set for the five different items, and the superscript (2) indicates the hidden layer; the output function used by the hidden layer is... t=1, 2, 3, 4, 5, 6, 7; the output layer receives the output function of the hidden layer, and the input function used by the output layer is... The superscript (3) indicates the output layer, q represents a neuron in the output layer, and the value of q is 1 or 2; the output function of the output layer is ,in, This indicates the real-time supply voltage amplitude required by the load. This indicates the frequency of the power supply voltage required by the load in real time.

[0056] After the BP neural network calculates the amplitude and frequency of the supply voltage, the PFC circuit receives the AC power corresponding to the actual bus voltage of the air conditioner, performs power factor correction rectification on the AC power, and converts the AC power into DC power. After the DC-AC circuit receives the DC power rectified by the PFC circuit, a SiC MOSFET inverter switch is used to convert the DC power into sinusoidal AC power and output it to the load terminal of the air conditioner. After the air conditioner cools the room temperature to the set temperature of 25°C, it continues to run for 3 hours.

[0057] The number of people in the mall fluctuated frequently within 3 hours, ranging from 86 to 198, but the temperature was consistently maintained at 25°C.

[0058] Ultimately, the air conditioner's output voltage was 210.6V, its output current was 15.9A, and its output power was 3348.5W. Compared to its rated power of 4500W, the energy-saving effect was as high as 25.6%.

[0059] Then, after the first test, wait for a period of time, and repeat the above process to continue the test twice. The results of these three tests are shown in Table 1 below:

[0060] Table 1: Test Results of Operating Efficiency of Energy-Saving Power Supply Regulation System

[0061] bus voltage Output voltage Output current Output power Rated power Energy efficiency 228.3 210.6 15.9 3348.5 4500 25.6% 229.6 215.3 16.8 3701.4 4500 17.7% 224.2 211.7 16.2 3429.5 4500 23.8%

[0062] Three tests were conducted, and the energy-saving effect ranged from a minimum of 17.7% to a maximum of 25.6%, effectively saving electricity.

[0063] In summary, this invention uses a PC controller as the host computer for control. It can collect multiple environmental parameters from integrated sensors, thereby adjusting the outputs of the PFC circuit and the DC-AC circuit in real time. This ensures that the frequency and amplitude of the final sinusoidal AC output match the load in real time, effectively reducing energy loss and achieving energy conservation. Simultaneously, the use of a SiCMosfet inverter switch for conversion effectively reduces energy loss during the conversion process, further enhancing energy efficiency. Furthermore, the use of a BP neural network to calculate the required output at the load end enables rapid and accurate acquisition of the desired results, improving both efficiency and accuracy, demonstrating significant advancements.

[0064] The above embodiments are merely illustrative of the technical concept of the present invention and should not be construed as limiting the scope of protection of the present invention. Any modifications made to the technical solutions based on the technical concept proposed in this invention shall fall within the scope of protection of this invention.

Claims

1. An energy-saving power supply regulation system, comprising a power supply module for supplying power to the load, characterized in that, An integrated sensor, PC controller, PFC circuit, and DC-AC circuit are located between the power supply module and the load. The integrated sensor collects multiple environmental parameters in real time and transmits these parameters to the PC controller; these environmental parameters include at least indoor temperature, outdoor temperature, outdoor humidity, and the number of people in the mall. The PC controller is equipped with a BP neural network controller. The BP neural network in the BP neural network controller consists of an input layer, a hidden layer and an output layer. The input layer obtains multiple environmental parameters and the system's set target value as input, and then performs nonlinear transformation processing by the hidden layer. The output layer outputs the real-time required power supply voltage amplitude and power supply voltage frequency at the load end. After receiving AC power from the power supply module, the PFC circuit uses the real-time supply voltage amplitude required by the load as a basis to perform PWM closed-loop feedback regulation on the AC power using the first full-bridge topology circuit built into the PFC circuit, converting the AC power into DC power for input to the DC-AC circuit. The PFC circuit includes at least an RC filter circuit, a first full-bridge topology circuit, an LC voltage regulator and filter circuit, and a voltage detection circuit connected in sequence. After the RC filter circuit filters the AC power received by the PFC circuit, multiple switches in the first full-bridge topology circuit input multiple PWM signals to the filtered AC power, converting the filtered AC power into DC power. The LC voltage regulator and filter circuit then further regulates and filters the DC power, and the voltage detection circuit detects the DC power after regulation and filtering. Based on the detection results of the voltage detection circuit, the multiple PWM signals input to the multiple switches in the first full-bridge topology circuit are adjusted to form PWM closed-loop feedback regulation. After receiving DC power, the DC-AC circuit uses the real-time required voltage amplitude and frequency of the load to adjust multiple PWM signals of the DC input using the built-in second full-bridge topology circuit, converting the DC power into AC power. Then, the built-in multiple filter inductors in the DC-AC circuit further convert the converted AC power into sinusoidal AC power, which is received by the load. Both the first and second full-bridge topologies employ multiple switching transistors, each of which is a SiC MOSFET.

2. The energy-saving power supply regulation system according to claim 1, characterized in that, The input layer uses j input neurons, where j is the number of environmental parameters plus 1. The output function used by the input layer is... , x represents a neuron in the input layer, and the superscript (1) represents the input layer; The hidden layer receives the output function of the input layer, and the input function used by the hidden layer is... Both k and i are positive integers not greater than j, where k represents a neuron in the hidden layer. The weight coefficients are given by the superscript (2), which indicates the hidden layer. The output function used by the hidden layer is... t is a positive integer greater than k; The output layer receives the output function from the hidden layer, and the input function used by the output layer is... The superscript (3) indicates the output layer, q represents a neuron in the output layer, and the value of q is 1 or 2; the output function of the output layer is ,in, This indicates the real-time supply voltage amplitude required by the load. This indicates the frequency of the power supply voltage required by the load in real time.

3. The energy-saving power supply regulation system according to claim 2, characterized in that, When performing the corresponding function calculation process in the input layer, hidden layer, and output layer, the steepest descent method is used to adjust the parameters of the corresponding function calculation process.

4. The energy-saving power supply regulation system according to claim 3, characterized in that, A current detection circuit is also provided between the first full-bridge topology circuit and the LC voltage regulator and filter circuit; when the current detection value of the current detection circuit exceeds the maximum current limit of any switch in the PFC circuit, the PFC circuit is turned off.

5. The energy-saving power supply regulation system according to claim 1, characterized in that, The system is equipped with a control interface, which includes multiple buttons, a temperature control interface, and a power control interface. The multiple buttons are used to control the operating status of the system, the temperature control interface is used to display multiple environmental parameters and to input set target values, and the power control interface is used to display multiple actual output values ​​at the load end.

6. The energy-saving power supply regulation system according to claim 1, characterized in that, The DC-AC circuit receives DC power from the PFC circuit at its input terminal, regulates it using an RC regulator, and then uses a second full-bridge topology circuit composed of multiple parallel switches to input multiple PWM signals to convert the regulated DC power into AC power. After passing through a sinusoidal filter inductor L1, two relays are placed on the transmission path of the converted AC power to control the switching of the DC-AC circuit. The converted AC power then passes through sinusoidal filter inductors L2 and L3 to be converted into sinusoidal AC power and transmitted to the load terminal.

7. The energy-saving power supply regulation system according to claim 1, characterized in that, The control method of the system includes the following steps: S1. The system starts to supply power. After the PC controller initializes the program, it acquires multiple environmental parameters and sets target values. S2. Based on the multiple environmental parameters and set target values ​​in step S1, the BP neural network controller built into the PC controller is used to calculate the load side to obtain the required power supply voltage amplitude and power supply voltage frequency. S3. Determine whether the actual parameters corresponding to the required power supply voltage amplitude and frequency at the load end have reached the set target values. If yes, stop control; otherwise, return to step S1 where the PC controller obtains multiple environmental parameters and set target values, and repeat the process until the actual parameters of the system reach the set target values ​​and then stop control.

Images