Control method and device, electronic equipment, storage medium and program product

Abstract

The invention discloses a control method of a multi-phase power supply. The method comprises the following steps: acquiring historical network flow data of the gateway equipment; according to the network flow data and a pre-trained prediction model, generating a state setting instruction of the multi-phase power supply, the prediction model predicting a power utilization state of the gateway device according to the network flow data to obtain a power utilization state prediction result, and generating the state setting instruction of the multi-phase power supply according to the power utilization state prediction result; and controlling the current running state of the multi-phase power supply according to the state setting instruction. Thus, the state setting instruction is generated based on the historical network flow data and the prediction model, prospective adjustment of the operation state of the multi-phase power supply is realized to a certain extent, a passive response mode of feedback control is changed, the problem of response lag is effectively solved, accurate matching of the power supply capacity and the power utilization demand is ensured to a certain extent, and the power utilization efficiency is improved. The fluctuation of the output voltage is avoided, and the stable operation of the gateway equipment is ensured.

Description

Technical Field

[0001] This application relates to the field of electronic circuit technology, and in particular to a control method for a multiphase power supply, a control device for a multiphase power supply, an electronic device, a computer-readable storage medium, and a computer program product. Background Technology

[0002] In related technologies, multiphase power supplies often rely on feedback control mechanisms to adjust their own state. This involves passively adjusting the number of phases and operating mode of the power supply by monitoring fluctuations in parameters such as voltage and current at the power output. However, since fluctuations in parameters such as voltage and current can only be detected after changes in load power consumption, the state adjustment of the multiphase power supply always lags behind changes in load power consumption. Consequently, the electrical energy supplied by the multiphase power supply to the load may not effectively meet the load's power demand, affecting the stable operation of the load. Summary of the Invention

[0003] This application provides a control method for a multiphase power supply, a control device for a multiphase power supply, an electronic device, a computer-readable storage medium, and a computer program product.

[0004] This application provides a control method for a multiphase power supply, wherein the multiphase power supply supplies power to a gateway device, the method comprising: Obtain historical network traffic data from the gateway device; Based on the network traffic data and the pre-trained prediction model, the status setting instruction for the multiphase power supply is generated. The prediction model predicts the power consumption status of the gateway device based on the network traffic data, obtains the power consumption status prediction result, and generates the status setting instruction for the multiphase power supply based on the power consumption status prediction result. The current operating state of the multiphase power supply is controlled according to the state setting instructions.

[0005] In this way, by generating state setting instructions based on historical network traffic data and predictive models, the operating status of multiphase power supplies can be proactively adjusted to a certain extent. This changes the passive response mode of feedback control, effectively solves the problem of response lag, and enables multiphase power supplies to adjust their operating status in advance when gateway devices face rapid load changes such as packet burst processing and route updates. To a certain extent, this ensures a precise match between power supply capacity and power demand, avoids fluctuations in output voltage, and guarantees the stable operation of gateway devices.

[0006] In some implementations, the historical network traffic data includes at least one of packet rate, byte rate, queue depth, and route update events within a preset time period.

[0007] In this way, the multidimensional characteristics of historical network traffic data ensure the comprehensiveness and effectiveness of the input prediction model data to a certain extent, improve the accuracy of power consumption status prediction results, and enable the generated status setting instructions to accurately match the actual power consumption needs of the gateway device. This enhances the stability and reliability of multiphase power supply to a certain extent, effectively solves the problem of lag in traditional feedback control response, and ensures the stable and efficient operation of the gateway device.

[0008] In some implementations, the power consumption state prediction results include expected current prediction and / or current change rate prediction.

[0009] Thus, the synergistic support of the dual-dimensional prediction results makes the generation of state setting instructions more accurate and reasonable, effectively making up for the limitations of traditional single-dimensional prediction, realizing refined control of the operating status of multi-phase power supply, thereby improving the matching degree between power supply capacity and power demand to a certain extent, and effectively avoiding voltage drop caused by insufficient power supply or energy waste caused by oversupply.

[0010] In some implementations, the prediction model includes a decision tree.

[0011] Thus, choosing decision trees as the prediction model effectively balances the requirements of prediction accuracy and real-time performance, and to a certain extent solves the problems of large computational delays in complex models and inaccurate predictions in simple models, thereby improving the effectiveness and reliability of multiphase power supply control.

[0012] In some implementations, controlling the current operating state of the multiphase power supply according to the state setting instruction includes: The status setting instruction is written into the execution chip of the multiphase power supply via a serial communication bus, wherein the execution chip of the multiphase power supply sets the current operating state according to the status setting instruction.

[0013] Thus, the use of a serial communication bus ensures the reliability and stability of command transmission, avoiding command errors caused by signal distortion or loss. The precise analysis and control of the execution chip improves the alignment between the multi-phase power supply's operating state and command requirements. The collaboration between the serial communication bus and the execution chip ensures that the power supply status of the multi-phase power supply accurately matches the predicted power demand, thereby improving the effectiveness of the control scheme and ensuring the stable operation of the gateway device.

[0014] In some implementations, the status setting instruction includes at least one of a phase number setting instruction, an operating mode setting instruction, and a switching frequency setting instruction.

[0015] In this way, the synergistic effect of multiple dimensions such as the number of phases, operating mode, and switching frequency enables comprehensive control of the operating parameters of multi-phase power supplies and in-depth optimization of the power supply status. This improves the adaptability and control accuracy of multi-phase power supplies to a certain extent, effectively avoids the limitations of traditional single control methods, and ensures the stability and reliability of power supply quality.

[0016] In some embodiments, the phase number setting instruction includes a phase increment instruction for increasing the phase number of the multiphase power supply, and controlling the operating state of the multiphase power supply according to the state setting instruction includes: According to the phase-increase command, the first phase unit in the multiphase power supply is controlled to enter the on state.

[0017] In this way, controlling the first phase unit to enter the start state in a timely manner effectively solves the problem of insufficient power supply caused by the untimely access of the new phase unit. To a certain extent, it ensures that the output current capability of the multi-phase power supply can quickly respond to the load growth demand, ensures the stability of the multi-phase power supply output voltage, effectively avoids the impact of voltage fluctuations on the gateway equipment, and improves the reliability and stability of the power supply system.

[0018] In some embodiments, the phase number setting instruction includes a phase increment instruction for increasing the phase number of the multiphase power supply, and the step of controlling the operating state of the multiphase power supply according to the state setting instruction further includes: The first phase unit in the multiphase power supply is controlled to enter the on state, and the multiphase power supply is controlled to perform a preset operation, wherein the preset operation includes at least one of the following operations: Increase the width of the first pulse signal of the first phase unit; Adjust the compensator parameters of the voltage feedback loop in the multiphase power supply; Increase the reference voltage of the voltage feedback loop in the multiphase power supply.

[0019] In this way, by performing the corresponding preset operation while activating the first phase unit, the stable output of the multiphase power supply after phase addition is effectively ensured. This solves the problems of slow current build-up and unstable voltage that exist in the newly added phase unit, enabling the multiphase power supply to quickly enter a stable working state after phase addition. To a certain extent, this enhances the dynamic response capability and stability of the power supply system and improves the operational reliability of the gateway equipment.

[0020] In some embodiments, the phase number setting instruction includes a phase reduction instruction for reducing the number of phases of the multiphase power supply, and controlling the operating state of the multiphase power supply according to the state setting instruction includes: According to the phase reduction command, the duty cycle of the second phase unit in the multiphase power supply is reduced, and the second phase unit is controlled to enter the off state when the inductor current of the second phase unit is less than or equal to a preset value.

[0021] In this way, by gradually reducing the duty cycle of the second phase unit and then turning it off under appropriate current conditions, the problem of sudden changes in inductor current and voltage spikes caused by directly turning off the phase unit can be effectively avoided to a certain extent. This ensures the stability of the multi-phase power supply output voltage and prevents voltage fluctuations from affecting the gateway device.

[0022] This application also provides a control device for a multiphase power supply, wherein the multiphase power supply supplies power to a gateway device, the device comprising: The data acquisition module is used to acquire historical network traffic data of the gateway device; The instruction generation module is used to generate the status setting instruction of the multiphase power supply based on the network traffic data and the pre-trained prediction model. The prediction model predicts the power consumption status of the gateway device based on the network traffic data, obtains the power consumption status prediction result, and generates the status setting instruction of the multiphase power supply based on the power consumption status prediction result. The control module is used to control the current operating state of the multiphase power supply according to the state setting instructions.

[0023] In this way, by generating state setting instructions based on historical network traffic data and predictive models, the operating status of multiphase power supplies can be proactively adjusted to a certain extent. This changes the passive response mode of feedback control, effectively solves the problem of response lag, and enables multiphase power supplies to adjust their operating status in advance when gateway devices face rapid load changes such as packet burst processing and route updates. To a certain extent, this ensures a precise match between power supply capacity and power demand, avoids fluctuations in output voltage, and guarantees the stable operation of gateway devices.

[0024] This application also provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, it implements the methods described in some of the above embodiments.

[0025] This application also provides a computer-readable storage medium, characterized in that the computer-readable storage medium stores a computer program, which, when executed by one or more processors, implements the methods described in some of the above embodiments.

[0026] This application also provides a computer program product, including a computer program / instructions, characterized in that the computer program / instructions, when executed by a processor, implement the methods described in some of the above embodiments.

[0027] The electronic device, computer-readable storage medium, and computer program product provided in this application, when implementing the above method, first acquire historical network traffic data of the gateway device; then, based on the network traffic data and a pre-trained prediction model, generate a state setting instruction for the multi-phase power supply. The prediction model predicts the power consumption state of the gateway device based on the network traffic data, obtains a power consumption state prediction result, and generates the state setting instruction for the multi-phase power supply based on the power consumption state prediction result; finally, it controls the current operating state of the multi-phase power supply according to the state setting instruction. In this way, generating state setting instructions based on historical network traffic data and a prediction model achieves, to a certain extent, a proactive adjustment of the multi-phase power supply's operating state, changing the passive response mode of feedback control and effectively solving the problem of response lag. This allows the multi-phase power supply to adjust its operating state in advance when the gateway device faces rapid load changes such as packet burst processing and route updates, ensuring, to a certain extent, a precise match between power supply capacity and power demand, avoiding output voltage fluctuations, and guaranteeing the stable operation of the gateway device.

[0028] Additional aspects and advantages of embodiments of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of embodiments of this application. Attached Figure Description

[0029] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, wherein: Figure 1 This is one of the flowcharts illustrating a control method for a multiphase power supply according to certain embodiments of this application; Figure 2 This is a schematic diagram illustrating the generation process of state setting instructions in certain embodiments of this application; Figure 3 This is a second schematic flowchart of a multiphase power supply control method according to certain embodiments of this application; Figure 4 This is the third flowchart illustrating a control method for a multiphase power supply according to certain embodiments of this application; Figure 5 This is the fourth flowchart illustrating a control method for a multiphase power supply according to certain embodiments of this application; Figure 6 This is the fifth flowchart illustrating a control method for a multiphase power supply according to certain embodiments of this application; Figure 7 This is a schematic diagram of the system architecture of a control method for a multiphase power supply according to certain embodiments of this application. Detailed Implementation

[0030] The embodiments of this application are described in detail below. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the embodiments of this application, and should not be construed as limiting the embodiments of this application.

[0031] In wired routers, switches, and other network communication devices, the System on Chip (SoC) is the core component ensuring efficient operation, and its power supply stability directly determines the performance of the network device. In related technologies, multi-phase power supplies have become the standard power supply solution for such SoCs. This solution balances high current output capacity with energy efficiency under light load scenarios by dynamically adjusting the number of operating phases, flexibly adapting to the power supply needs of the SoC under different load conditions, and providing strong support for the basic operation of network devices.

[0032] Multiphase power supplies typically adjust their operating phases and mode by monitoring fluctuations in electrical parameters such as voltage and current at the power output in real time. However, fluctuations in these parameters can only be detected after changes in load power consumption. This means that the power supply's adjustments always lag behind actual load power changes, making it difficult to handle rapid load transients (microsecond to nanosecond levels) caused by network events such as burst packet processing and routing protocol updates in the core processor. When such transients occur, the power supply cannot adjust its power supply capacity in time, leading to output voltage fluctuations that affect the stable operation of the core processor and reduce the performance and reliability of the entire network device.

[0033] To compensate for the lag in feedback control, some related technologies build models by analyzing historical power consumption data of the core processor or readings from built-in sensors to adjust the number of power phases in advance. However, the current and power consumption signals on which this approach relies for prediction are themselves the result of load changes, causing the prediction to always lag behind network events that trigger power consumption changes and large-volume bursts, thus failing to achieve true forward-looking control. Furthermore, this approach fails to utilize signals in the network data plane that can directly characterize subsequent load features, such as queue depth, Direct Memory Access (DMA) transfer status, and routing protocol events, resulting in a single information dimension for the prediction model and making it difficult to accurately predict the dynamic power supply needs of the core processor.

[0034] In summary, the existing multiphase power supply systems cannot fully meet the dynamic and rapidly changing power supply requirements of the core processor.

[0035] Based on the above issues, please refer to Figure 1This application provides a method for controlling a multiphase power supply, wherein the multiphase power supply supplies power to a gateway device, and the method includes: 01: Obtain historical network traffic data from the gateway device; 02: Based on network traffic data and a pre-trained prediction model, generate a multi-phase power supply status setting command. The prediction model predicts the power consumption status of the gateway device based on network traffic data, obtains the power consumption status prediction result, and generates the multi-phase power supply status setting command based on the power consumption status prediction result. 03: Control the current operating status of the multiphase power supply according to the status setting instructions.

[0036] This application provides a control device for a multiphase power supply. The control method for the multiphase power supply according to this application can be implemented by the control device. Specifically, the control device includes a data acquisition module, an instruction generation module, and a control module. The data acquisition module acquires historical network traffic data from a gateway device. The instruction generation module generates a state setting instruction for the multiphase power supply based on the network traffic data and a pre-trained prediction model. The prediction model predicts the power consumption status of the gateway device based on the network traffic data, obtains the power consumption status prediction result, and generates the state setting instruction for the multiphase power supply based on the power consumption status prediction result. The control module controls the current operating state of the multiphase power supply according to the state setting instruction.

[0037] This application also provides a server, which includes a memory and a processor. The multiphase power supply control method of this application can be implemented by the server of this application. Specifically, the memory stores a computer program, and the processor is used to acquire historical network traffic data of the gateway device. The processor is used to generate a multiphase power supply status setting instruction based on the network traffic data and a pre-trained prediction model, wherein the prediction model predicts the power consumption status of the gateway device based on the network traffic data, obtains the power consumption status prediction result, and generates the multiphase power supply status setting instruction based on the power consumption status prediction result. The processor is used to control the current operating state of the multiphase power supply according to the status setting instruction.

[0038] Specifically, a multiphase power supply is composed of multiple phase units connected in parallel. By adjusting the number of working phase units and their operating parameters, a power system with different power supply requirements can be achieved, which can power the core processor of the gateway device.

[0039] Gateway devices are network communication devices such as wired routers and network switches. The power requirements of their core processors fluctuate dynamically with changes in network services.

[0040] Historical network traffic data is a collection of network transmission-related data generated by the gateway device within a preset time period, which can reflect the workload change trend of the gateway device.

[0041] The prediction model is an algorithm model that has been pre-trained and has the ability to predict the target result based on the input data. In the embodiments of this application, it can be used to predict the power consumption status of the gateway device based on historical network traffic data.

[0042] The power consumption status prediction result is information output by the prediction model related to the power consumption demand of the gateway device within a preset future time period.

[0043] The status setting command is an instruction to adjust the operating status of a multiphase power supply, which can specify the specific parameters that need to be adjusted.

[0044] The current operating status refers to the specific operating parameters of the multiphase power supply during operation, such as the number of phases, operating mode, and switching frequency.

[0045] During the operation of the multiphase power supply, the power supply system will continuously collect network traffic data related to the gateway device within a preset time period. By capturing various network transmission data during the operation of the gateway device, it provides comprehensive and reliable data support for subsequent power consumption status prediction.

[0046] In addition, the preset duration needs to balance effectiveness and real-time performance. It should be long enough to capture meaningful network traffic patterns and ensure that the collected data accurately reflects the load change characteristics of the gateway device, but it should not be too long to avoid introducing irrelevant noise or causing calculation delays. It is usually in the range of hundreds of microseconds to tens of milliseconds. In practical applications, it can be flexibly adjusted according to the working characteristics of the gateway device and the network environment.

[0047] Then, the instruction generation module inputs the acquired historical network traffic data into the pre-trained prediction model. This prediction model learns the mapping relationship between historical network traffic data and the power consumption status of the gateway device, and can accurately predict the power consumption status of the gateway device based on the input historical network traffic data, thus obtaining the power consumption status prediction result.

[0048] Subsequently, based on the obtained power consumption status prediction results, the instruction generation module generates corresponding multiphase power supply status setting instructions, so that the instruction content matches the predicted power consumption demand, ensuring that the adjusted multiphase power supply can adapt to the upcoming power consumption changes.

[0049] After receiving the status setting command, the multiphase power supply adjusts its own operating status, such as the number of phases and working mode, according to the command requirements, so that the power supply capacity of the multiphase power supply is precisely matched with the power demand of the gateway device.

[0050] In this way, by generating state setting instructions based on historical network traffic data and predictive models, the operating status of multiphase power supplies can be proactively adjusted to a certain extent. This changes the passive response mode of feedback control, effectively solves the problem of response lag, and enables multiphase power supplies to adjust their operating status in advance when gateway devices face rapid load changes such as packet burst processing and route updates. To a certain extent, this ensures a precise match between power supply capacity and power demand, avoids fluctuations in output voltage, and guarantees the stable operation of gateway devices.

[0051] In some implementations, historical network traffic data includes at least one of packet rate, byte rate, queue depth, and route update events within a preset time period.

[0052] Specifically, the preset duration is a specific time interval for collecting historical network traffic data. It needs to be set reasonably to balance the effectiveness and real-time nature of the data, usually in the range of hundreds of microseconds to tens of milliseconds, in order to cover multiple network event cycles and capture complete traffic change patterns.

[0053] Packet rate is the ratio of the number of data packets received or sent by the gateway device within a preset time period to the time period. It can reflect the frequency of data packet transmission. The increase or decrease of packet rate corresponds to the increase or decrease of the workload of the gateway device, and can reflect the changes in power demand.

[0054] The byte rate is the ratio of the total number of data bytes transmitted by the gateway device within a preset duration to the duration. It can reflect the total scale of data transmission. Different byte rates correspond to different processing loads of the gateway device, and its changing trend can effectively reflect the fluctuations in power consumption.

[0055] Queue depth is the length of the queue of data packets to be processed in the gateway device. An increase in queue depth indicates that the gateway device will face greater processing pressure, and the subsequent power demand is likely to increase.

[0056] A route update event refers to an event in which the routing table in a gateway device is updated. When a route update occurs, the gateway device needs to perform a lot of calculations and data processing, which can cause a sharp change in power demand.

[0057] In practical applications, the data acquisition module can continuously collect one or more of the aforementioned traffic data from the gateway device within a preset time period, based on the actual application scenario and prediction accuracy requirements. For example, in scenarios with high real-time requirements, packet rate, queue depth, and route update events can be collected simultaneously. These data can quickly reflect the load change trend of the gateway device. In scenarios sensitive to the total amount of data transmission, byte rate and packet rate can be the primary data collected.

[0058] Understandably, under different network environments and business needs, targeted selection of corresponding key features for data collection and analysis can not only avoid resource consumption caused by redundant data to a certain extent, but also allow data collection to accurately focus on core needs, providing reliable data support for the regulation of multiphase power sources. This enables the regulation of multiphase power sources to flexibly adapt to various application scenarios, improving the adaptability and operational efficiency of the control process.

[0059] Furthermore, when collecting historical network traffic data, it is necessary to set an appropriate data collection time interval to ensure the real-time nature and integrity of the data, and to ensure the accurate capture of events reflecting changes in gateway device load, such as packet bursts, queue accumulation, and route updates. For example, on a 10Gbps link, the transmission time of a 1500-byte data packet is approximately 1.2 microseconds. Setting the preset duration to 50 microseconds is sufficient to cover the transmission process of dozens of data packets. This not only allows for accurate calculation of traffic characteristics such as packet rate and byte rate during this period, but also enables simultaneous monitoring of dynamic changes in queue depth and the occurrence of route update events, providing comprehensive and reliable basic data support for subsequent predictive model training and real-time prediction.

[0060] In addition, the collected multi-dimensional historical network traffic data serves as input to the predictive model. By learning the mapping relationship between multi-dimensional historical network traffic data and the power consumption status of gateway devices, the model can accurately identify precursory signals of load changes and predict the changing trends of power demand in advance, providing solid data support for generating accurate status setting instructions. For example, when both packet rate and queue depth show an upward trend, and a routing update event is detected, the model can predict that the load on the gateway device will increase, and then generate corresponding status setting instructions.

[0061] In this way, the multidimensional characteristics of historical network traffic data ensure the comprehensiveness and effectiveness of the input prediction model data to a certain extent, improve the accuracy of power consumption status prediction results, and enable the generated status setting instructions to be accurately matched with the actual power consumption needs of the gateway device. This enhances the stability and reliability of multiphase power supply to a certain extent, effectively solves the problem of lag in traditional feedback control response, and ensures the stable and efficient operation of the gateway device.

[0062] In some implementations, the power consumption state prediction results include expected current prediction and / or current change rate prediction.

[0063] Specifically, the expected current prediction is the current demand value that the gateway device will generate in the future period of time, which is predicted by the prediction model based on historical network traffic data. This value can reflect the total power demand of the gateway device in the future.

[0064] Current change rate prediction is the rate of change of the future current demand of the gateway device predicted by the prediction model, that is, the amount of change of current per unit time. This parameter can reflect the degree of drastic change in the power demand of the gateway device.

[0065] After receiving historical network traffic data, the prediction model uses a built-in algorithm to calculate the corresponding expected current and current change rate.

[0066] For expected current prediction, the model analyzes the mapping relationship between characteristics such as packet rate, byte rate, and queue depth in historical network traffic data and past load current to estimate the current required by the future load. For example, when historical data shows that queue depth is continuously increasing and packet rate remains high, the model will predict a higher expected current value, indicating that the multiphase power supply needs to have sufficient current output capability.

[0067] For current change rate prediction, the model will focus on capturing abrupt changes in network traffic data, such as the occurrence of routing update events and sudden increases or decreases in packet rate.

[0068] The predictive model combines these characteristics with the correlation between past current change rates to predict the steepness of future current changes. For example, in the event of a routing update event, the model predicts a sharp increase in the current change rate, instructing the multiphase power supply to quickly adjust its operating state to cope with sudden load changes.

[0069] In addition, the prediction model can flexibly output the expected current prediction result or the current change rate prediction result, or both, according to the needs of actual applications.

[0070] For example, when the load changes gradually, appropriate state setting instructions can be generated simply by referring to the expected current prediction results. When the load may experience transient changes, the response strategy of the multi-phase power supply needs to be adjusted in a targeted manner based on the current change rate prediction results to ensure power supply stability.

[0071] Understandably, in different network service scenarios, the prediction results can be output by selecting a single dimension or a combination of dimensions to make the solution adaptable to various application scenarios. This not only avoids the resource consumption caused by redundant calculations to a certain extent, but also accurately matches the needs of the scenario, thereby improving the adaptability and efficiency of the control process.

[0072] Thus, the synergistic support of the dual-dimensional prediction results makes the generation of state setting instructions more accurate and reasonable, effectively making up for the limitations of traditional single-dimensional prediction, realizing refined control of the operating status of multi-phase power supply, thereby improving the matching degree between power supply capacity and power demand to a certain extent, and effectively avoiding voltage drop caused by insufficient power supply or energy waste caused by oversupply.

[0073] In some implementations, the prediction model includes a decision tree.

[0074] Specifically, a decision tree is a machine learning model that makes decisions based on a tree structure. By learning from historical data, it can construct a series of decision rules based on feature variables. It has the characteristics of simple structure, efficient training, fast inference speed, and strong interpretability. It can quickly complete predictions without complex computing resources.

[0075] In practical applications, the instruction generation module must first train the prediction model to pre-build decision rules. Model training can be performed during power supply system initialization or idle periods to avoid consuming resources when the gateway device is running normally.

[0076] The training data for the model consists of historical network traffic data and the actual power consumption data of the corresponding gateway devices. The historical network traffic data includes features such as packet rate, byte rate, queue depth, and routing update events, while the actual power consumption data includes parameters such as load current.

[0077] Decision tree models can analyze historical network traffic data and actual electricity consumption data to uncover the mapping relationship between different network traffic characteristics and electricity consumption status, and construct appropriate decision rules based on this mapping relationship.

[0078] Subsequently, while the gateway device is operating normally, the acquisition module continuously collects current historical network traffic data and inputs it into the trained decision tree model.

[0079] The decision tree model quickly traverses the tree structure based on built-in decision rules and outputs the power consumption status prediction results.

[0080] Understandably, decision trees possess fast and efficient reasoning capabilities, which to a certain extent ensures the real-time output of power consumption state prediction results, providing timely support for the forward-looking control of multiphase power supplies. This enables multiphase power supplies to adjust their operating status in advance, cope with rapid load transients, avoid fluctuations in output voltage, and ensure the stable operation of gateway devices.

[0081] In addition, the decision tree model can accurately capture the correlation between historical network traffic data and power consumption status, with high prediction accuracy. This makes the generated status setting instructions highly matched with actual power consumption demand, thereby improving the effectiveness and reliability of multiphase power supply control to a certain extent.

[0082] Thus, choosing decision trees as the prediction model effectively balances the requirements of prediction accuracy and real-time performance, and to a certain extent solves the problems of large computational delays in complex models and inaccurate predictions in simple models, thereby improving the effectiveness and reliability of multiphase power supply control.

[0083] Please see Figure 2The following explanation uses the workflow of a high-performance wired router as an example to illustrate the generation process of status setting instructions in some of the above implementation methods: When a high-performance wired router handles a large volume of data transmission in a campus network, its application layer or data plane packet scheduler is responsible for distributing the network packets to be processed. The routing protocol processor processes routing update tasks caused by network topology changes in real time, while the Quality of Service Policy Manager (QoS) manages traffic of different priorities in a differentiated manner. The three work together to support the efficient transmission of network data. At the same time, the network driver monitoring module in the operating system or kernel space monitors the operating status of network interfaces in real time, the kernel power management module coordinates the allocation of power resources, and the system load tracker continuously records the overall load change trend of the router, providing basic status data for subsequent analysis.

[0084] During data processing, the real-time queue deep monitoring module accurately captures the dynamic growth of the packet queue to be processed on the router port, the routing update event detection module detects changes in the routing table in real time, and the priority traffic identification module clearly distinguishes between high-priority traffic from real-time services such as video conferencing and ordinary data transmission traffic. The multi-dimensional network status information collected by these three modules, covering queue status, routing changes, and traffic levels, is synchronously transmitted to the multi-dimensional feature extractor, providing comprehensive and accurate data support for subsequent load prediction.

[0085] The multi-dimensional feature extractor filters and refines the collected raw data such as queue depth, route update frequency, and traffic priority to extract features that can characterize the load changes of the router's core processor. These feature data are then input into the intelligent load prediction engine. Driven by an efficient predictive control mechanism, the engine, in conjunction with the load current observer and a decision tree model pre-trained during system idle periods, quickly analyzes the mapping relationship between the input features and historical load data. Finally, it generates a predicted load curve and an estimated load current, predicting the changes in the core processor's power demand within the next 50 microseconds.

[0086] Finally, based on the prediction results output by the intelligent load prediction engine, the model prediction controller generates state setting instructions, including phase number adjustment, operating mode switching, and switching frequency setting. These instructions are written to the execution chip of the multi-phase power supply through the serial communication bus, driving the multi-phase power supply to adjust its operating state in advance to ensure that its power supply capacity is accurately matched with the router's upcoming power demand, effectively cope with the rapid load changes caused by sudden data transmission and routing updates, and ensure the stable and efficient operation of the router.

[0087] Please see Figure 3In some implementations, step 03 includes: 031: The status setting instruction is written to the execution chip of the multiphase power supply through the serial communication bus. The execution chip of the multiphase power supply sets the current operating state according to the status setting instruction.

[0088] In some implementations, the control module is also used to write status setting instructions into the execution chip of the multiphase power supply via a serial communication bus, wherein the execution chip of the multiphase power supply sets the current operating state according to the status setting instructions.

[0089] In some implementations, the processor is also used to write a status setting instruction to the execution chip of the multiphase power supply via a serial communication bus, wherein the execution chip of the multiphase power supply sets the current operating state according to the status setting instruction.

[0090] Specifically, a serial communication bus is a communication line that transmits data serially. It features stable transmission, strong anti-interference capability, and simple hardware implementation. It can meet the instruction transmission requirements between multi-phase power supplies and control modules. Common serial communication buses include the Inter-Integrated Circuit (IIC) bus.

[0091] The execution chip is the internal chip of a multiphase power supply responsible for receiving control commands and controlling the operation of various components. It has the functions of parsing commands and controlling the operation of power supply hardware. It can adjust the operating status of the multiphase power supply, such as the number of phases, operating mode, and switching frequency, according to the received commands. For example, according to the received commands, the number of phases of the multiphase power supply can be set to 2, the operating mode can be set to light-load high-efficiency mode, and the switching frequency can be set to Fsw=500KHz.

[0092] After generating the status setting command, the control module sends the command data to the execution chip of the multi-phase power supply via a pre-deployed serial communication bus. The transmission characteristics of the serial communication bus can effectively resist electromagnetic interference, ensuring the integrity and accuracy of the command data during transmission to a certain extent, and avoiding command errors caused by signal distortion.

[0093] After receiving the status setting instruction, the execution chip first parses the instruction and extracts information such as the phase number adjustment requirements, working mode parameters, and switching frequency settings.

[0094] Subsequently, based on the analysis results, the execution chip adjusts the operating state of the multi-phase power supply through its internal drive circuits and control logic to ensure that the current operating state of the multi-phase power supply is consistent with the instruction requirements. For example, when the instruction requires increasing the number of phases, the execution chip will activate the corresponding phase unit drive signal to control the newly added phase unit to enter the working state. When the instruction requires switching the operating mode, the execution chip will adjust the control parameters of the relevant circuits to adapt the multi-phase power supply to different load scenarios.

[0095] Thus, the use of a serial communication bus ensures the reliability and stability of command transmission, avoiding command errors caused by signal distortion or loss. The precise analysis and control of the execution chip improves the alignment between the multi-phase power supply's operating state and command requirements. The collaboration between the serial communication bus and the execution chip ensures that the power supply status of the multi-phase power supply accurately matches the predicted power demand, thereby improving the effectiveness of the control scheme and ensuring the stable operation of the gateway device.

[0096] In some implementations, the status setting instruction includes at least one of the following: phase number setting instruction, operating mode setting instruction, and switching frequency setting instruction.

[0097] Specifically, the phase number setting command is used to instruct a multiphase power supply to adjust the number of working phase units. The number of phases of a multiphase power supply determines its output current capability. The more phases, the stronger the output current capability. The output current capability of the power supply can be changed by increasing or decreasing the number of phases to match the power demand of different load intensities.

[0098] The operating mode setting command is used to indicate the switching operating mode of a multi-phase power supply. Different operating modes correspond to different efficiency characteristics and response strategies. For example, the light-load high-efficiency mode is suitable for scenarios with low power demand, while the heavy-load stable mode is suitable for scenarios with high power demand and large fluctuations.

[0099] The switching frequency setting command is used to instruct a multiphase power supply to adjust the switching frequency of the switching transistors. The switching frequency can affect the output ripple, conversion efficiency, and dynamic response speed of a multiphase power supply. A suitable switching frequency can improve power efficiency while ensuring power quality.

[0100] After the predictive model outputs the power consumption status prediction results, the instruction generation module generates one or more of the following instructions based on information such as the expected current and current change rate in the prediction results: phase number setting instruction, working mode setting instruction, and switching frequency setting instruction. This ensures that the power supply status of the multi-phase power supply matches the upcoming power demand of the gateway device. In this way, under different network service scenarios, single or partial instructions can be flexibly selected to accurately adapt to the needs based on the actual application scenario. This allows the solution to accurately adapt to various application scenarios, which simplifies the control process and reduces resource consumption to a certain extent, while focusing on core needs to improve control efficiency and enhance the adaptability and practicality of the control process.

[0101] When it is predicted that the gateway device is about to enter a heavy load state, the instruction generation module generates a phase number setting instruction to increase the number of phases, a working mode setting instruction to switch to the heavy load stable mode, and a switching frequency setting instruction to appropriately increase the switching frequency.

[0102] Understandably, increasing the number of phases can improve the power supply's output current capability, heavy-load mode prioritizes power supply stability, and increasing the switching frequency can speed up the response. The synergy of these three factors enables the multi-phase power supply to meet the high current and fast response requirements of heavy-load scenarios.

[0103] Furthermore, when the gateway device is under continuous heavy load, the control module can also control the multi-phase power supply to actively switch the main control phase while maintaining the number of phases, so as to avoid overheating of a single phase and affecting the power supply capacity of the multi-phase power supply.

[0104] When it is predicted that electricity demand will decrease and stabilize, the instruction generation module will generate a phase number setting instruction to reduce the number of phases, a working mode setting instruction to switch to light load high efficiency mode, and a switching frequency setting instruction to reduce the switching frequency.

[0105] Understandably, reducing the number of phases can avoid energy waste, light-load mode can improve power supply efficiency, and reducing the switching frequency can reduce switching losses. The three work together to achieve efficient and energy-saving operation in light-load scenarios.

[0106] When a rapid transient load change is predicted but the overall current demand is moderate, the instruction generation module can generate instructions to adjust the switching frequency. By optimizing the switching frequency, the dynamic response capability of the power supply is improved, thereby ensuring the stability of the output voltage.

[0107] In this way, the synergistic effect of multiple dimensions such as the number of phases, operating mode, and switching frequency enables comprehensive control of the operating parameters of multi-phase power supplies and in-depth optimization of the power supply status. This improves the adaptability and control accuracy of multi-phase power supplies to a certain extent, effectively avoids the limitations of traditional single control methods, and ensures the stability and reliability of power supply quality.

[0108] Please see Figure 4 In some implementations, the phase number setting instruction includes a phase increment instruction for increasing the number of phases in a multi-phase power supply. Step 03 further includes: 032: According to the phase increment command, control the first phase unit in the multiphase power supply to enter the turn-on state.

[0109] In some implementations, the control module is also used to control the first phase unit in the multiphase power supply to enter the on state according to the phase increment command.

[0110] In some implementations, the processor is also configured to control the first phase unit in the multiphase power supply to enter the on state according to the phase increment instruction.

[0111] Specifically, the phase increment instruction is one type of phase number setting instruction, used to instruct a multiphase power supply to increase the number of working phase units in order to improve the output current capability and adapt to the growth of power demand.

[0112] The first phase unit is a newly added phase unit that needs to be activated according to the phase increment command. It can be any phase unit in the multiphase power supply that is in an inactive state. Its hardware structure is the same as other active phase units, with independent switching transistors, inductors, and other components, and it can independently output current. In practical applications, the specific phase unit to be activated can be determined according to the hardware design and control logic of the multiphase power supply to ensure a balanced workload of each phase unit. For example, the first phase unit is usually determined by combining factors such as the cumulative working time and temperature of each phase unit to avoid overheating caused by excessive operation of a single phase unit and to ensure the long-term stable operation of the multiphase power supply.

[0113] The active state is when the switching transistor of the phase unit is turned on and off normally according to the control signal, the inductor current is established normally and participates in the total current output of the multiphase power supply, and the working state provides power to the load.

[0114] In practical applications, when the instruction generation module determines that the number of phases of the multiphase power supply needs to be increased based on the power consumption status prediction results to meet the upcoming power demand, it will generate a corresponding phase increase instruction and transmit it to the execution chip of the multiphase power supply through the serial communication bus.

[0115] After receiving the phase increment command, the execution chip first identifies the first phase unit that needs to be activated.

[0116] After determining the first phase unit, the execution chip sends an enable control signal to the first phase unit to start the drive circuit of the phase unit, so that its switching transistor starts to work according to the preset switching frequency and duty cycle.

[0117] After the first phase unit enters the turn-on state, its inductor current begins to gradually build up, rising from the initial zero value to the average current level with other working phase units, thereby contributing to the total output current of the multiphase power supply and improving the power supply capacity of the multiphase power supply.

[0118] During startup, the execution chip monitors the working status of the first phase unit in real time to ensure that its switching transistors are turned on and off normally, and that the inductor current rises steadily without abnormal fluctuations.

[0119] In this way, controlling the first phase unit to enter the start state in a timely manner effectively solves the problem of insufficient power supply caused by the untimely access of the new phase unit. To a certain extent, it ensures that the output current capability of the multi-phase power supply can quickly respond to the load growth demand, ensures the stability of the multi-phase power supply output voltage, effectively avoids the impact of voltage fluctuations on the gateway equipment, and improves the reliability and stability of the power supply system.

[0120] Please see Figure 5 In some implementations, the phase number setting instruction includes a phase increment instruction for increasing the number of phases in a multi-phase power supply. Step 03 further includes: 033: Control the first phase unit in the multiphase power supply to enter the on state, and control the multiphase power supply to perform a preset operation, wherein the preset operation includes at least one of the following operations: Increase the width of the first pulse signal of the first phase unit; Adjust the compensator parameters in the voltage feedback loop of the multiphase power supply; Increase the reference voltage of the voltage feedback loop in the multiphase power supply.

[0121] In some embodiments, the control module is further configured to control the first phase unit in the multiphase power supply to enter the on state, and to control the multiphase power supply to perform a preset operation, wherein the preset operation includes at least one of the following operations: Increase the width of the first pulse signal of the first phase unit; Adjust the compensator parameters in the voltage feedback loop of the multiphase power supply; Increase the reference voltage of the voltage feedback loop in the multiphase power supply.

[0122] In some embodiments, the processor is further configured to control the first phase unit in the multiphase power supply to enter an on state, and to control the multiphase power supply to perform a preset operation, wherein the preset operation includes at least one of the following operations: Increase the width of the first pulse signal of the first phase unit; Adjust the compensator parameters in the voltage feedback loop of the multiphase power supply; Increase the reference voltage of the voltage feedback loop in the multiphase power supply.

[0123] Specifically, the preset operation is a control operation that is pre-set to ensure the stable output of the power supply after phase addition, including increasing the width of the first pulse signal, adjusting the compensator parameters, and increasing the reference voltage.

[0124] The width of the first pulse signal is the duration of the first pulse output by the switching transistor after the first phase unit is turned on. The pulse width can affect the rise rate of the inductor current. Increasing the pulse width can speed up the establishment of the inductor current.

[0125] The voltage feedback loop is a closed-loop control system in a multiphase power supply used to stabilize the output voltage. It monitors the deviation between the output voltage and the reference voltage and adjusts the control signal to maintain the stability of the output voltage.

[0126] The compensator parameters are the relevant parameters of the compensator in the voltage feedback loop. Adjusting these parameters can change the characteristics of the voltage feedback loop, such as bandwidth, gain, and phase margin, to improve the dynamic response speed and stability of the voltage feedback loop.

[0127] The reference voltage is the target voltage value set in the voltage feedback loop. Increasing the reference voltage can, to some extent, offset the voltage drop caused by the increase in load.

[0128] In practical applications, when the execution chip receives the phase increase command and controls the first phase unit to enter the turn-on state, it will also select to execute one or more preset operations based on the load prediction and the status of the power supply system to ensure stable power output.

[0129] The first preset operation is to increase the width of the first pulse signal of the first phase unit. Compared with the pulse width in steady state, increasing the width of the first pulse signal can provide stronger excitation to the inductor, causing its current to rise rapidly and reach the average current level of other working phase units in a short time. This can quickly make up for the current gap that may be caused by a sudden increase in load and suppress the drop in output voltage to a certain extent.

[0130] The second preset operation is to adjust the compensator parameters of the voltage feedback loop. By increasing parameters such as loop bandwidth, gain, or phase margin, the dynamic response speed of the voltage feedback loop is improved, enabling it to detect small changes in the output voltage more quickly and adjust the control signal in a timely manner, thereby enhancing the transient performance of the power supply system and ensuring the rapid stabilization of the output voltage.

[0131] The third preset operation is to increase the reference voltage of the voltage feedback loop. When the load current rises and the output voltage tends to drop, the target voltage value is appropriately increased. This can provide a buffer for voltage adjustment in advance, offset some of the voltage drop, shorten the time required for voltage stabilization, and avoid voltage fluctuations.

[0132] Typically, the feedforward voltage offset can be calculated using the formula: Feedforward voltage offset = ΔI_predicted × R_droop, where ΔI_predicted is the predicted change in load current and R_droop is set to match the expected output impedance of the power supply loop. This offset is usually in the range of 1%-3%.

[0133] For example, if the system predicts that the load current will increase by ΔI=10A and the preset virtual impedance R_droop=2mΩ, then the calculated feedforward voltage offset = 10A × 0.002Ω = 0.02V, which can increase the reference voltage by 0.02V.

[0134] In this way, by performing the corresponding preset operation while activating the first phase unit, the stable output of the multiphase power supply after phase addition is effectively ensured. This solves the problems of slow current build-up and unstable voltage that exist in the newly added phase unit, enabling the multiphase power supply to quickly enter a stable working state after phase addition. To a certain extent, this enhances the dynamic response capability and stability of the power supply system and improves the operational reliability of the gateway equipment.

[0135] Please see Figure 6 In some implementations, the phase number setting instruction includes a phase reduction instruction for reducing the number of phases in a multi-phase power supply. Step 03 further includes: 034: According to the phase reduction command, reduce the duty cycle of the second phase unit in the multiphase power supply, and control the second phase unit to enter the off state when the inductor current of the second phase unit is less than or equal to the preset value.

[0136] In some implementations, the control module is also used to reduce the duty cycle of the second phase unit in the multiphase power supply according to the phase reduction command, and control the second phase unit to enter the off state when the inductor current of the second phase unit is less than or equal to a preset value.

[0137] In some implementations, the processor is also configured to reduce the duty cycle of the second phase unit in the multiphase power supply according to the phase reduction instruction, and control the second phase unit to enter the off state when the inductor current of the second phase unit is less than or equal to a preset value.

[0138] Specifically, the phase reduction instruction is one of the phase number setting instructions, which is used to instruct a multiphase power supply to reduce the number of working phase units in order to improve the power supply operating efficiency and avoid energy waste when power demand decreases.

[0139] The second phase unit is the working phase unit that needs to be shut down according to the phase reduction command. Its selection is based on the phase management logic of the multi-phase power supply. It is usually shut down in a preset order to ensure the load balance of the remaining working phase units. Generally, it is not possible to specify which phase unit to shut down.

[0140] The duty cycle is the ratio of the on-time of the switch in the phase unit to the switching period. The magnitude of the duty cycle can affect the output current of the phase unit. Reducing the duty cycle can gradually reduce the inductor current of the phase unit.

[0141] The inductor current is the current flowing through the inductor in the second phase unit, and its changes can affect the stability of the power supply output.

[0142] The preset value is a pre-set current threshold used to determine whether the second phase unit can be turned off. It is usually set to zero or a very small negative value to ensure that the inductor current is basically zero when it is turned off, thus avoiding sudden current changes.

[0143] In the off state, the switching transistor of the phase unit stops conducting and turning off, no longer participates in the total current output of the multiphase power supply, and is completely disconnected from the power supply circuit.

[0144] During implementation, when the instruction generation module determines that the power demand of the gateway device is reduced based on the power consumption status prediction results and the number of phases of the multiphase power supply needs to be reduced, it will generate a corresponding phase reduction instruction and transmit it to the execution chip of the multiphase power supply through the serial communication bus.

[0145] After receiving the phase reduction command, the execution chip first identifies the second phase unit that needs to be turned off, and then gradually reduces the duty cycle of the second phase unit. By adjusting the control signal, the execution chip gradually reduces the on-time of the switching transistor of the phase unit, so that the current flowing through the inductor decreases slowly.

[0146] During the reduction of the duty cycle, the execution chip monitors the inductor current change of the second phase unit in real time, obtains accurate current data through the current sampling circuit, and continuously compares it with the preset value.

[0147] When the inductor current is detected to drop to less than or equal to the preset value, it indicates that the current of the phase unit has been basically cleared to zero. At this time, the execution chip sends a shutdown signal to completely shut down the drive circuit of the second phase unit and put it into the shutdown state.

[0148] In this way, by gradually reducing the duty cycle of the second phase unit and then turning it off under appropriate current conditions, the problem of sudden changes in inductor current and voltage spikes caused by directly turning off the phase unit can be effectively avoided to a certain extent. This ensures the stability of the multi-phase power supply output voltage and prevents voltage fluctuations from affecting the gateway device.

[0149] Please see Figure 7 The following explanation uses the control flow of a multi-phase power supply in a high-performance wired router that handles concurrent services such as multi-enterprise video conferencing and data synchronization as an example to illustrate the control method of the multi-phase power supply in this application: When a high-performance wired router in a commercial park handles concurrent services such as multi-enterprise video conferencing and data synchronization, Real-time queue depth monitoring at the hardware platform layer accurately captures the rapid growth of the queue of pending data packets on the port; routing update event detection detects the frequent changes in the routing table caused by the access of new enterprises; priority traffic identification clearly distinguishes high-priority traffic such as video conferencing from ordinary data traffic. This network status data, along with hardware performance events collected by the performance counting unit and memory bandwidth data obtained by the memory controller, are synchronously transmitted to the multi-dimensional feature extractor.

[0150] The multi-dimensional feature extractor filters, integrates, and refines the aforementioned multi-source data to extract key features that can accurately characterize the load change trend of the router's core processor. These features are then input into the intelligent load prediction engine of the power conversion layer. Driven by an efficient predictive control mechanism, this engine, in conjunction with real-time power supply data acquired by high-precision current sampling and output voltage sampling modules, and combined with a decision tree model pre-trained during system idle periods, quickly analyzes the mapping relationship between features and historical load data, generates a predicted load curve and estimates the load current. Simultaneously, the accuracy of the prediction results is verified through a load current observer, and the management unit configures early warning strategies to promptly output load warning signals.

[0151] The SoC power model predictive controller calculates the optimal control quantity for the multi-phase power supply based on the prediction results output by the intelligent load prediction engine and the load warning signal, combined with the power output voltage data fed back by Vout (output voltage). After the optimal control quantity is processed by the intelligent power control module, on the one hand, it generates a compensation signal through the adaptive compensator to optimize the power output stability, and on the other hand, it drives the phase number management state machine and the current balance engine to operate, and sends control commands such as phase number adjustment, working mode switching, and switching frequency setting to each phase power level.

[0152] Ultimately, the power stages of phase 1, phase 2 to phase N work together according to the received control commands. Under the scheduling of the phase number management state machine, the number of working phases is dynamically adjusted. Under the action of the current balancing engine, the current of each phase is balanced, realizing the precise matching between the power supply status of the multi-phase power supply and the power demand of the router's core processor. This effectively copes with the rapid load transients caused by business concurrency and ensures that the router operates stably and efficiently under high load scenarios.

[0153] This application also provides a control device for a multiphase power supply, wherein the multiphase power supply supplies power to a gateway device, and the device includes: The data acquisition module is used to acquire historical network traffic data from the gateway device; The instruction generation module is used to generate state setting instructions for multi-phase power supply based on network traffic data and a pre-trained prediction model. The prediction model predicts the power consumption status of the gateway device based on network traffic data, obtains the power consumption status prediction result, and generates the state setting instructions for multi-phase power supply based on the power consumption status prediction result. The control module is used to control the current operating state of the multiphase power supply according to the status setting instructions.

[0154] Specifically, during the operation of the multiphase power supply, the power supply system will continuously collect network traffic data related to the gateway device within a preset time period. By capturing various network transmission data during the operation of the gateway device, it provides comprehensive and reliable data support for subsequent power consumption status prediction.

[0155] In addition, the preset duration needs to balance effectiveness and real-time performance. It should be long enough to capture meaningful network traffic patterns and ensure that the collected data accurately reflects the load change characteristics of the gateway device, but it should not be too long to avoid introducing irrelevant noise or causing calculation delays. It is usually in the range of hundreds of microseconds to tens of milliseconds. In practical applications, it can be flexibly adjusted according to the working characteristics of the gateway device and the network environment.

[0156] Then, the instruction generation module inputs the acquired historical network traffic data into the pre-trained prediction model. This prediction model learns the mapping relationship between historical network traffic data and the power consumption status of the gateway device, and can accurately predict the power consumption status of the gateway device based on the input historical network traffic data, thus obtaining the power consumption status prediction result.

[0157] Subsequently, based on the obtained power consumption status prediction results, the instruction generation module generates corresponding multiphase power supply status setting instructions, so that the instruction content matches the predicted power consumption demand, ensuring that the adjusted multiphase power supply can adapt to the upcoming power consumption changes.

[0158] After receiving the status setting command, the multiphase power supply adjusts its own operating status, such as the number of phases and working mode, according to the command requirements, so that the power supply capacity of the multiphase power supply is precisely matched with the power demand of the gateway device.

[0159] In this way, by generating state setting instructions based on historical network traffic data and predictive models, the operating status of multiphase power supplies can be proactively adjusted to a certain extent. This changes the passive response mode of feedback control, effectively solves the problem of response lag, and enables multiphase power supplies to adjust their operating status in advance when gateway devices face rapid load changes such as packet burst processing and route updates. To a certain extent, this ensures a precise match between power supply capacity and power demand, avoids fluctuations in output voltage, and guarantees the stable operation of gateway devices.

[0160] This application also provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, it implements the methods of some of the above-described embodiments.

[0161] This application also provides a computer-readable storage medium storing a computer program that, when executed by one or more processors, implements the methods of some of the above-described embodiments.

[0162] This application also provides a computer program product, including a computer program / instructions, characterized in that the computer program / instructions, when executed by a processor, implement the methods of some of the above-described embodiments.

[0163] It is understood that a computer program includes computer program code. Computer program code can be in the form of source code, object code, executable files, or some intermediate form. Computer-readable storage media can include: any entity or device capable of carrying computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), and software distribution media, etc.

[0164] In this specification, the terms "specifically," "furthermore," "particularly," "understandably," etc., refer to specific features, structures, materials, or characteristics described in connection with embodiments or examples that are included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0165] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of executable request code comprising one or more steps for implementing a particular logical function or process, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order according to the functions involved, as should be understood by those skilled in the art to which embodiments of this application pertain.

[0166] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of this application.

Claims

1. A control method for a multiphase power supply, characterized in that, The multiphase power supply supplies power to the gateway device, and the method includes: Obtain historical network traffic data from the gateway device; Based on the network traffic data and the pre-trained prediction model, the status setting instruction for the multiphase power supply is generated. The prediction model predicts the power consumption status of the gateway device based on the network traffic data, obtains the power consumption status prediction result, and generates the status setting instruction for the multiphase power supply based on the power consumption status prediction result. The current operating state of the multiphase power supply is controlled according to the state setting instructions.

2. The control method for a multiphase power supply according to claim 1, characterized in that, The historical network traffic data includes at least one of the following within a preset time period: packet rate, byte rate, queue depth, and route update events.

3. The control method for a multiphase power supply according to claim 1, characterized in that, The power consumption prediction results include expected current prediction and / or current change rate prediction.

4. The control method for a multiphase power supply according to claim 1, characterized in that, The prediction model includes a decision tree.

5. The control method for a multiphase power supply according to claim 1, characterized in that, The step of controlling the current operating state of the multiphase power supply according to the state setting instruction includes: The status setting instruction is written into the execution chip of the multiphase power supply via a serial communication bus, wherein the execution chip of the multiphase power supply sets the current operating state according to the status setting instruction.

6. The control method for a multiphase power supply according to claim 1, characterized in that, The status setting instructions include at least one of the following: phase number setting instructions, operating mode setting instructions, and switching frequency setting instructions.

7. The control method for a multiphase power supply according to claim 6, characterized in that, The phase number setting instruction includes a phase increment instruction for increasing the phase number of the multiphase power supply, and the step of controlling the operating state of the multiphase power supply according to the state setting instruction includes: According to the phase-increase command, the first phase unit in the multiphase power supply is controlled to enter the on state.

8. The control method for a multiphase power supply according to claim 7, characterized in that, The phase number setting instruction includes a phase increment instruction for increasing the phase number of the multiphase power supply, and the step of controlling the operating state of the multiphase power supply according to the state setting instruction further includes: The first phase unit in the multiphase power supply is controlled to enter the on state, and the multiphase power supply is controlled to perform a preset operation, wherein the preset operation includes at least one of the following operations: Increase the width of the first pulse signal of the first phase unit; Adjust the compensator parameters of the voltage feedback loop in the multiphase power supply; Increase the reference voltage of the voltage feedback loop in the multiphase power supply.

9. The control method for a multiphase power supply according to claim 6, characterized in that, The phase number setting instruction includes a phase reduction instruction for reducing the number of phases in the multiphase power supply, and the step of controlling the operating state of the multiphase power supply according to the state setting instruction includes: According to the phase reduction command, the duty cycle of the second phase unit in the multiphase power supply is reduced, and the second phase unit is controlled to enter the off state when the inductor current of the second phase unit is less than or equal to a preset value.

10. A control device for a multiphase power supply, characterized in that, The multiphase power supply supplies power to the gateway device, and the device includes: The data acquisition module is used to acquire historical network traffic data of the gateway device; The instruction generation module is used to generate the status setting instruction of the multiphase power supply based on the network traffic data and the pre-trained prediction model. The prediction model predicts the power consumption status of the gateway device based on the network traffic data, obtains the power consumption status prediction result, and generates the status setting instruction of the multiphase power supply based on the power consumption status prediction result. The control module is used to control the current operating state of the multiphase power supply according to the state setting instructions.

11. An electronic device, characterized in that, The method includes a memory and a processor, wherein the memory stores a computer program, which, when executed by the processor, implements the method according to any one of claims 1-9.

12. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by one or more processors, implements the method of any one of claims 1-9.

13. A computer program product comprising a computer program / instructions, characterized in that, When the computer program / instruction is executed by the processor, it implements the method described in any one of claims 1-9.

Images