Intelligent power supply system constructed by using artificial intelligence auxiliary tool

Abstract

The invention discloses an intelligent power supply system constructed by utilizing an artificial intelligence auxiliary tool, which is executed in a Windows system host and comprises a CPU (Central Processing Unit) and a PSU (Power Supply Unit), and the PSU is provided with a control chip and a cooling fan, is electrically connected with the CPU and outputs an output voltage. An artificial intelligence auxiliary tool with LLMs and Microsoft graphic data as a building framework is embedded in the CPU, and the CPU is provided with a dialogue interface. When the dialogue interface receives a dialogue input from the outside, the CPU analyzes the meaning of the dialogue to generate an action instruction, and at the moment, the CPU requests the control chip to check and report a real-time state of the PSU, so that when it is known that the meaning of the dialogue is different from the real-time state through comparison, the action instruction is sent to the control chip. And the action instruction is transmitted to the control chip to drive the PSU to adjust the voltage value of the output voltage or the working state of the cooling fan. In this way, the user-friendly, intuitive and convenient PSU control experience is provided.

Description

Technical Field

[0001] The present invention relates to a power supply system, and in particular to an intelligent power supply system constructed using artificial intelligence (AI) auxiliary tools. Background Art

[0002] For a computer, the power supply is as crucial a component as the human heart. Composed primarily of rectifier circuits, transformer circuits, power switching circuits, and voltage output circuits, it's responsible for converting mains electricity into the appropriate voltages (including +12V, +5V, +3.3V, and other voltages) and currents required by other components within the computer. Therefore, the power supply's performance directly impacts the computer's operation. To meet the demands of niche markets, high-performance and gaming computers capable of handling highly complex computations have been developed. As previously mentioned, these superior computing capabilities place higher demands on the power supply's output power and operational stability, as well as more sophisticated operational settings. Furthermore, with the increasing awareness of energy conservation, carbon reduction, and environmental protection, the energy-saving function of power supplies has been put on the development agenda and urged to be effectively implemented. However, although some power supplies now offer energy-saving functions, the user still has to manually switch on or off the function. Users unfamiliar with digital systems may forgo using the function due to the complex procedures. Alternatively, users who find the switching of the energy-saving function unintuitive may forgo using it, which defeats the purpose of establishing the energy-saving function. Furthermore, with the development of high-performance and high-function applications of computer hosts, the power consumed by the computer host during operation may vary greatly due to the increasing complexity of the types of programs executed. This requires the power supply to provide a power output function with a wider range of variations and a higher maximum power when the host is running. As a result, whether the power supply can maintain high and stable working quality has become more important and needs to be closely monitored.

[0003] On the other hand, Copilot is a code editor plug-in developed jointly by GitHub and OpenAI. It uses a large pre-trained language model and Microsoft graph data to understand code, that is, the context of the conversational statements we are accustomed to, and generates corresponding real-time code suggestions and auto-completion functions. Therefore, Copilot can help developers improve code productivity, reduce the workload of repetitive coding, and provide new solutions and ideas. In this regard, the present invention aims to further improve the comprehensiveness and real-time monitoring of the operating status of power supplies through the direct use of the Copilot code editor plug-in. While ensuring stable power output, users can easily control the power supply through natural language chat through a system dialogue interface with human-like intuitiveness, thereby replacing the complex manual operation process and improving the shortcomings of the above-mentioned existing technologies. Summary of the Invention

[0004] The primary objective of the present invention is to provide an intelligent power supply system with natural language dialogue capabilities, offering an intuitive system dialogue interface or window that enables control of the power supply unit (PSU) operating mode through natural language chat, thereby providing a more user-friendly and convenient computer operation experience.

[0005] To achieve the above objectives, the present invention provides an intelligent power supply system constructed using an artificial intelligence-assisted tool. The system runs on a Windows system host and includes a central processing unit (CPU) and a power supply. The power supply is equipped with a control chip and a cooling fan, electrically connected to the CPU, and outputs an output voltage. The CPU is embedded with the artificial intelligence-assisted tool, which uses Large Language Models (LLMs) and Microsoft Graph data as its architecture, and has a dialogue interface. When the dialogue interface receives a dialogue input from an external source, the CPU parses the meaning of the dialogue and generates an action command. At this point, the CPU requests the control chip to verify and report a real-time status of the power supply. If the comparison shows that the meaning of the dialogue is different from the real-time status, the CPU transmits the action command to the control chip to drive the power supply to adjust the output voltage or the operating status of the cooling fan.

[0006] The power supply includes a power manager electrically connected to the control chip for monitoring the load status of the power supply. The action command is a command for optimizing power output efficiency during a period. When the central processing unit (CPU) learns that the power supply has not enabled the power output efficiency optimization function during the period, the CPU transmits the action command to the control chip, causing the power supply to automatically increase the output voltage value in response to an increase in load during the period. The CPU then displays an action message through the dialogue interface to notify the CPU that the action command has been executed. The control chip periodically transmits the real-time status to the CPU, and the real-time status includes at least the operating efficiency, load status, and output voltage value of the power supply. When the CPU detects that the operating efficiency of the power supply is below an efficiency standard based on the real-time status, the CPU outputs an aging notification through the dialogue interface. When the CPU detects that the output voltage value is not within a specified range based on the real-time status, the CPU outputs a voltage anomaly notification through the dialogue interface. When the CPU detects that the load state of the power supply is greater than or equal to a maximum rated output based on the real-time status, the CPU outputs an overload notification through the dialogue interface. When the central processing unit detects that the power supply is in an extremely light-load working state, the control chip is driven to turn off the power factor correction (PFC) function of the power supply, and a PFC turn-off notification is sent through the dialogue interface.

[0007] The action command is a command for turning on / off the fan's eco-saving (Ecology, Conservation, Optimization, ECO) function. When the central processing unit checks the real-time status and finds that the power supply has not turned on / off the eco-saving function of the cooling fan, the action command is transmitted to the control chip to turn on / off the eco-saving function of the cooling fan, and an action message is displayed on the dialogue interface to inform that the action command has been executed. Conversely, when the central processing unit finds that the eco-saving function of the cooling fan has been turned on / off, the command is not executed and a no-action message is fed back. The cooling fan is an RGB fan having a plurality of LED lights, the plurality of LED lights being configured to emit lights of a plurality of colors. The action command is a command for turning on / off the fan's RGB function. When the central processing unit checks the real-time status and learns that the power supply has not turned on / off the plurality of LED lights of the cooling fan, the central processing unit transmits the action command to the control chip to turn on / off the plurality of LED lights of the cooling fan, and displays an action message through the dialogue interface to inform that the action command has been executed. Conversely, when the central processing unit checks and learns that the RGB function of the cooling fan has been turned on / off, the command is not executed and a no-action message is fed back. Furthermore, the action command is a command for switching the light colors of the plurality of LED lights. When the central processing unit checks the real-time status and learns that the light colors of the plurality of LED lights are different from the meaning of the dialogue, the central processing unit transmits the action command to the control chip to switch the light colors of the plurality of LED lights of the cooling fan.

[0008] The power supply is equipped with a thermal sensor element for sensing an internal operating temperature of the power supply, and the central processing unit stores a temperature warning value of the power supply reported by the control chip. When the cooling fan is in an energy-saving mode, the central processing unit detects the output wattage of the power supply and the operating temperature, and when it determines that the workload is between 0% and 30%, stops the cooling fan to reduce the operating noise to below 5dBA. The central processing unit periodically requests the control chip to report the operating temperature via the real-time status, and when it detects that the operating temperature exceeds the corresponding temperature warning value, transmits an emergency warning message via the dialogue interface to notify the user to perform emergency operations. The central processing unit stores a preset speed value of the cooling fan reported by the control chip, and the central processing unit requests the control chip to report the speed of the cooling fan. When it detects that the speed of the cooling fan does not meet the preset speed value, the central processing unit transmits a fan abnormality message via the dialogue interface to notify the user of the fan abnormality.

[0009] The central processing unit (CPU) performs a rolling analysis of the operating performance of the Windows system host to obtain at least one high-load period, and sequentially records energy consumption data of a plurality of operating programs in the Windows system host. The CPU analyzes the plurality of energy consumption data to obtain an operating probability of each operating program corresponding to the high-load period, and further summarizes the plurality of operating programs with a high probability of running during the high-load period to form a predicted execution table based on the plurality of operating probabilities. Accordingly, the CPU outputs a predicted execution query before the high-load period, for querying the execution intention of each operating program in the predicted execution table through the dialogue interface.

[0010] In summary, the present invention utilizes the Copilot AI-assisted tool built into the CPU to assist users in understanding and using the power supply. This allows users to intuitively control the power supply's hardware functions and perform internal monitoring operations through a conversational interface within the Windows system. This provides users with a more intelligent, streamlined, and efficient power supply system, thereby achieving a digital service experience that meets human needs. Furthermore, the present invention's intelligent power supply system includes the following features:

[0011] 1. Given the same load output, a higher current results in greater power loss and lower efficiency. However, when a Windows host executes software, the load on the host typically fluctuates, causing the power supply's output efficiency to fluctuate accordingly. To address this issue, the present invention allows users to control the Copilot AI through a dialogue interface to transmit power output efficiency optimization commands to the control chip. This dynamically increases the output voltage when the power supply is under heavy load, ensuring constant, optimized efficiency.

[0012] 2. Through real-time status transmission and communication between the control chip and the CPU, Copilot monitors the power supply's efficiency at load levels of 20%, 50%, and 100%. If the efficiency at any stage fails to meet the corresponding efficiency standard, it will output an aging notification. This provides real-time warnings to the user of possible power supply malfunctions or internal component degradation, ensuring safe power supply operation.

[0013] 3. Power Factor (PF) measures the efficient use of electricity. The power switch of a typical power supply is a capacitor-input circuit, resulting in power loss due to the phase difference between current and voltage. To address this, power supply circuits typically include a Power Factor Corrector (PFC) to improve PF. However, the inventors, through years of research, have learned that when the power supply is operating at very light loads, i.e., when the Windows system host is idle, PFC operation actually consumes more power than operation without PFC. Therefore, the inventors have designed Copilot to automatically disable PFC when it detects that the power supply is operating at very light loads and transmit a PFC shutdown notification to the user via the dialogue interface. This implements dynamic, interactive intelligent power management, allowing users to easily monitor real-time power usage. Furthermore, the communication of power usage status and corresponding management strategies further educates users about relevant knowledge and information.

[0014] 4. The present invention further enables the central processing unit to utilize Copilot to monitor the operating status of the power supply. When an abnormality occurs, the central processing unit outputs the aging notification, the voltage abnormality notification, and the overload notification through the dialogue interface, thereby providing real-time warnings to the user regarding power supply usage issues and enhancing electrical safety.

[0015] 5. The CPU monitors the operating temperature and cooling fan speed of the power supply in the background via Copilot. If it detects an impending high load or an abnormality, it proactively notifies the user through the dialogue interface, providing reminders, notifications, and early warnings. Further dialogue and guidance through the dialogue interface assists the user in adjusting running programs, troubleshooting, or performing emergency operations such as saving current files and closing currently executed programs. This prevents the high-intensity operation and high power output of the power supply from affecting the stability and smoothness of the overall system operation. It also prevents the power supply from shutting down in an emergency due to an abnormality, which could cause damage to important files, loss of files, or transaction errors due to incomplete program transactions. This ensures the security and smoothness of user operations within the computer system. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 This is a structural diagram of the first preferred embodiment of the present invention.

[0017] Figure 2 FIG. 4 is a structural diagram of a second preferred embodiment of the present invention.

[0018] Figure 3A 、 Figure 3B 、 Figure 3CFlowchart of the second preferred embodiment of the present invention.

[0019] Figure 4 FIG. 1 is a schematic diagram of a power supply according to a second preferred embodiment of the present invention.

[0020] Figure 5 Schematic diagram of the third preferred embodiment of the present invention.

[0021] Figure 6 Schematic diagram of the fourth preferred embodiment of the present invention.

[0022] Figure 7 Schematic diagram of the fifth preferred embodiment of the present invention.

[0023] Figure 8 Schematic diagram of the sixth preferred embodiment of the present invention.

[0024] Explanation of the accompanying symbols: 1-intelligent power supply system; 10-central processing unit; 100-dialogue interface; 1000-aging notification; 1001-voltage abnormality notification; 1002-overload notification; 1003-PFC shutdown notification; 1004-emergency warning message; 1005-fan abnormality information; 1006-predictive execution inquiry; 101-action instruction; 102-predictive execution form; 11-power supply; 110-rectifier; 111-transformer; 112-PFC; 113-power manager; 114-control chip; 1140-real-time status; 115-cooling fan; 116-thermal sensing element; 20-dialogue; 200-action instruction; S10~S151-step; S20~S25-step; S30~S35-step; S40~S45-step; S50~S55-step. DETAILED DESCRIPTION

[0025] To enable those skilled in the art to clearly understand the present invention, the following description is provided with reference to the drawings.

[0026] See also Figure 1 , which is an architectural diagram of the first preferred embodiment of the present invention. As shown, the intelligent power supply system 1, constructed using artificial intelligence-assisted tools, runs on a Windows system host (not shown) and includes a central processing unit 10 and a power supply 11. The power supply 11 is equipped with a control chip 114 and a cooling fan 115, electrically connected to the central processing unit 10 and outputting an output voltage. The central processing unit 10 is embedded with artificial intelligence-assisted tools based on a large language model and Microsoft graphics data, and has a conversational interface 100.

[0027] When the dialogue interface 100 receives a dialogue 20 inputted from an external source, the central processing unit 10 parses the meaning of the dialogue 20 and generates an action instruction 101. At this point, the central processing unit 10 requests the control chip 114 to check and report a real-time status 1140 of the power supply 11. If the comparison shows that the meaning of the dialogue 20 is different from the real-time status 1140, the control chip 114 transmits the action instruction 101 to drive the power supply 11 to adjust the output voltage or the operating status of the cooling fan 115.

[0028] See also Figures 2 to 4 , which are respectively an architecture diagram, a flow chart, and a power supply schematic diagram of the second preferred embodiment of the present invention. As shown in the figure, the intelligent power supply system 1 constructed using artificial intelligence-assisted tools is executed on a Windows system host (not shown) and includes a central processing unit 10 and a power supply 11. The central processing unit 10 is electrically connected to the power supply 11 and extracts operating power from the power supply 11. The central processing unit 10 is embedded with the Copilot artificial intelligence-assisted tool, which is a data computing auxiliary model built using a large language model and Microsoft graphics data. This provides the central processing unit 10 with a dialogue interface 100, allowing a user of the Windows system host to control the operating mode or action function of the power supply 11 through natural language dialogue, further realizing an intelligent and convenient computer operation experience.

[0029] The power supply 11 includes a rectifier 110, a transformer 111, a PFC (power factor corrector) 112, a power manager 113, a control chip 114, a cooling fan 115, and a thermal sensor 116. The transformer 111 is connected to the rectifier 110 and the PFC 112. The power manager 113 is electrically connected to the PFC 112, the control chip 114, and the cooling fan 115. The control chip 114 is connected to the thermal sensor 116. Accordingly, the power supply 11 processes AC power through power conditioning processes such as rectification, voltage transformation, power regulation, filtering, and voltage stabilization. The power manager 113 monitors the load status of the power supply 11 and the control chip 114 communicates with Copilot to ultimately generate various voltages that meet the requirements of the Windows system host and output an output voltage. The operation of the intelligent power supply system 1 may include the following steps. Incidentally, the combination of the central processing unit 10 and the control chip 114 enables information transmission and processing between the power supply 11 and the host computing processor within the Windows system host. That is, the combination of the central processing unit 10 and the control chip 114 forms a monitoring unit of the present invention installed in the Windows system host, which helps the user to easily understand and control the operation of the power supply through natural language.

[0030] When the Windows system host completes basic component assembly and initially operates, the control chip 114 reports a temperature warning value for the power supply 11, for example, a factory-set maximum critical temperature of 90°C and a temperature warning value of 80°C or above, to the CPU 10 for storage. Simultaneously, the control chip 114 reports a speed preset for the cooling fan 115, for example, a factory-set maximum speed of 2900 RPM and a minimum speed of 1200 RPM, to the CPU 10 for storage. The thermal sensor 116 senses an operating temperature within the power supply 11, which the control chip 114 reports to the CPU 10 in real time, periodically, or upon request.

[0031] The user can also open the dialogue interface 100 on their own. In step S10, when the dialogue interface 100 receives a dialogue 20 inputted from an external source, the CPU 10 parses the dialogue 20 to obtain a plurality of keywords. It then performs a logical operation on the keywords to determine the meaning of the dialogue 20 and generates an action instruction 101. In step S11, the CPU 10 requests the power supply 11 to report a real-time status 1140. This real-time status 1140 includes at least the output voltage value, the load status and operating efficiency of the power supply 11, the operating status and speed of the cooling fan 115, and the operating temperature sensed by the thermal sensor 116. In step S12, the CPU 10 compares the real-time status 1140 with the dialogue 20 to determine if they are identical. If not, the CPU 10 compares and finds that the meaning of the dialogue 20 is different from the real-time status 1140, then in step S120, the action instruction 101 is transmitted to the control chip 114 to drive the power supply 11 to adjust the output voltage value or the operating state of the cooling fan 115. After the power supply 11 executes the instruction, the CPU 10 can feedback an action message through the dialogue interface 100 to inform that the instruction has been executed. On the other hand, if the meaning of the dialogue 20 is the same as the real-time status 1140, that is, the power supply 11 is already operating in the power output adjustment function requested by the action instruction 101, or the cooling fan 115 is already in the operating state required by the user, then in step S121, the instruction is not executed and a non-action message is fed back.

[0032] In addition to the user independently communicating with the CPU 10 through the dialogue interface 100 , the CPU 10 further enables Copilot to periodically request the control chip 114 to report the real-time status 1140 of the power supply 11 through the real-time status 1140 . Accordingly, in step S13, when the CPU 10 detects that the operating efficiency of the power supply 11 is lower than an efficiency standard through the real-time status 1140, an aging notification 1000 is output through the dialogue interface 100. In step S130, when the CPU 10 detects that the output voltage value is not within a specification range, a voltage abnormality notification 1001 is output through the dialogue interface 100. In step S131, when the CPU 10 detects that the load state of the power supply 11 is greater than or equal to a maximum rated output, an overload notification 1002 is output through the dialogue interface 100. In step S132, when the CPU 10 detects that the power supply 11 is in an extremely light-load operating state, the CPU 10 drives the control chip 114 to disable the power factor correction function of the power supply 11 and transmits a PFC shutdown notification 1003 through the dialogue interface 100.

[0033] At the same time, the CPU 10 monitors the temperature status of the power supply 11 through the real-time status 1140. In step S14, when the CPU 10 detects that the operating temperature exceeds the corresponding temperature warning value, it actively sends an emergency warning message 1004 through the dialogue interface 100 to inform the user to perform emergency operations, such as saving the editing file as soon as possible. The CPU 10 monitors the operating status of the cooling fan 115 via the real-time status 1140. In step S140, when the CPU 10 detects that the energy-saving function of the cooling fan 115 is enabled, the CPU 10 detects the output wattage and the operating temperature of the power supply 11. When the CPU 10 determines that the workload is between 0% and 30%, the CPU 10 stops the cooling fan 115 to reduce the operating noise to below 5dBA. Simultaneously, the CPU 10 monitors the speed of the cooling fan 115. In step S141, when the CPU 10 detects that the speed of the cooling fan 115 does not meet the corresponding preset speed value, such as insufficient speed or the cooling fan 115 stops operating under light load even though the energy-saving function is not enabled, the CPU 10 proactively transmits a fan abnormality message 1005 via the dialogue interface 100 to inform the user of the fan abnormality. This provides the user with a direct, clear, and comprehensive method of obtaining information about the failure of the cooling fan 115, thereby helping the user accurately assess and determine whether to repair or replace the power supply 11.

[0034] Incidentally, the intelligent power supply system 1 further allows the user to engage in intelligent dialogue with the central processing unit 10 through the dialogue interface 100 to clarify the fault problem, thereby achieving the effect of using Copilot to guide the user in handling events such as overheating or fan failure. In this way, the central processing unit 10 uses the Copilot artificial intelligence auxiliary tool to monitor the operation of the cooling fan 115 to ensure that the fan's maximum and minimum speeds are maintained at the original factory settings, thereby preventing the fan from being affected by bearing or other component wear or failure, such as the maximum speed of 2900RPM dropping to below 2800RPM or the minimum speed of 1200RPM dropping to below 1100RPM, thereby avoiding problems such as damage to electronic components of the Windows system host due to untimely heat dissipation.

[0035] In step S15, the CPU 10 sequentially records energy consumption data for a plurality of running programs on the Windows system host. Simultaneously, the CPU 10 self-monitors and records its operating performance and temperature changes to calculate load variations. This allows for rolling analysis of the Windows system host's operating performance to identify at least one high-load period. In step S150, the CPU 10 analyzes the plurality of energy consumption data to determine a probability of each running program running during the high-load period. Based on the plurality of running probabilities, the CPU 10 aggregates the plurality of running programs with a high probability of running during the high-load period into a predicted execution table 102. In step S151, the CPU 10 outputs a predicted execution query 1006 before the high-load period arrives, for querying the execution intention of each of the running programs in the predicted execution form 102 through the dialogue interface 100, so as to calculate and estimate the system load level in combination with each of the running programs, and thus notify the user in advance through the dialogue interface 100 when the system may have insufficient computing resources and may cause transmission errors or delays.

[0036] In summary, the present invention utilizes the combination of the control chip 114 and the Copilot in the central processing unit 10 to enable the user and the power supply 11 to autonomously and directly communicate with each other through the dialogue interface 100 in a natural language dialogue to communicate power management demands, usage status, abnormal warnings and problem troubleshooting, etc., thereby helping the user to easily master, understand and control digital equipment while allowing the power supply 11 to autonomously and intelligently inform the user of relevant working conditions and possible problems, thereby effectively shortening the distance between digital technology and people. In this regard, the practical cases of the present invention are further listed for steps S10 to S121 for reference. Figure 5As shown. In step S20, when the dialogue interface 100 receives a dialogue 20 inputted from an external device, such as a request to enable / disable the PSU's ECO function, enable / disable the PSU's mute function, or enable / disable the PSU's semi-fanless mode, in step S21, the CPU 10 utilizes the Copilot artificial intelligence to parse the dialogue 20, understand the meaning, and generate the action instruction 101 for enabling / disabling the energy-saving (ECO) function. In step S22, the CPU 10 requests the control chip 114 to report the real-time status 1140 of the power supply 11 to verify whether the ECO function of the cooling fan 115 is enabled. If the real-time status 1140 is different from the meaning of the dialogue 20, meaning that the power supply 11 has not enabled / disabled the ECO function of the cooling fan 115, in step S23, the CPU 10 transmits the action instruction 101 to the control chip 114 via the Copilot tool to drive the power supply 11 to enable / disable the ECO function of the cooling fan 115. In step S24, after the cooling fan 115 has activated the ECO function, the CPU 10 displays the action information on the dialogue interface 100 via the Copilot tool, notifying the user that the system has executed the action instruction 101 to activate / deactivate the fan ECO function. Conversely, if the CPU 10 checks the real-time status 1140 and learns that the cooling fan 115's ECO function has already been activated / deactivated, the instruction is not executed in step S25, and the Copilot tool displays the inaction information via the dialogue interface 100, notifying the user that the cooling fan 115's ECO function has already been activated / deactivated, and the action will not be repeated. In this embodiment, the cooling fan 115 typically operates at 1000 RPM (Revolutions Per Minute) with a noise level of approximately 14 dBA (A-weighted decibels) when the workload is between 0% and 30%, i.e., between 0W and 300W. When the workload is between 30% and 100%, the fan operates at a power level of 300W to 1000W. Thus, when the cooling fan 115 has the ECO function enabled, the CPU 10 detects the output wattage and operating temperature of the power supply 11. If the workload is between 0% and 30%, the CPU 10 instructs the control chip 114 to stop the fan 115, thereby reducing the noise level to below 5 dBA. When the workload is between 30% and 100%, the fan continues to operate at a power level of 300W to 1000W. Thus, energy conservation is achieved by stopping the fan 115 during light load conditions.

[0037] For another example, the cooling fan 115 is an RGB fan having a plurality of LED lights, and the plurality of LED lights are configured to emit a plurality of colors, such as red, blue or yellow lights. Figure 6 As shown, in step S30, when the dialogue interface 100 receives the external input dialogue 20 for turning on / off the PSU's RGB function, in step S31, the central processing unit 10 parses the meaning of the dialogue 20 and generates the action instruction 101 for turning on / off the RGB function of the cooling fan 115. In step S32, the central processing unit 10 requests the control chip 114 to report the real-time status 1140. After confirming that the power supply 11 does not turn on / off the RGB function of the cooling fan 115, which is different from the meaning of the dialogue 20, in step S33, the action instruction 101 is transmitted to the control chip 114 to drive the power supply 11 to turn on / off the RGB function of the cooling fan 115 and turn the lighting of the plurality of LED lights on / off. In step S34, the Copilot tool feedbacks the action information through the dialogue interface 100. On the contrary, in step S35, when the CPU 10 checks and confirms that the on / off status of the plurality of LED lights is the same as the meaning of the dialogue 20, the CPU 10 feeds back the no action information through the dialogue interface 100 to inform the cooling fan 115 that the RGB function has been turned on / off, and therefore the action instruction 101 is not executed.

[0038] like Figure 7 As shown, in step S40, if the dialogue interface 100 receives the external input dialogue 20 to change the lighting color of the PSU fan to red, blue, or yellow, in step S41, the central processing unit 10 parses the meaning of the dialogue 20 and generates the action instruction 101 for switching the lighting color of the plurality of LED lights to red, blue, or yellow. At this time, in step S42, the central processing unit 10 requests the control chip 114 to report the real-time status 1140. After verifying that the lighting color of the plurality of LED lights of the cooling fan 115 is different from the meaning of the dialogue 20, in step S43, the action instruction 101 is transmitted to the control chip 114 to drive the power supply 11 to switch the lighting color of the plurality of LED lights of the cooling fan 115 to red, blue, or yellow. In step S44, the Copilot tool feedbacks the action information through the dialogue interface 100. On the contrary, in step S45, when the CPU 10 checks and confirms that the luminous colors of the plurality of LED lights are the same as the meaning of the dialogue 20, the CPU 10 feeds back the inaction information through the dialogue interface 100 to inform the cooling fan 115 that the light color has been displayed in red, blue or yellow.

[0039] In another example, Figure 7As shown, in step S50, the dialog interface 100 receives an external input, such as a dialog 20 requesting to optimize the PSU power output efficiency during a gaming session. In step S51, the CPU 10 utilizes the Copilot artificial intelligence to parse the dialog 20, understand the meaning, and generate the action instruction 101 for optimizing the power output efficiency during the gaming session. In step S52, the CPU 10 then requests the control chip 114 to report the real-time status 1140 of the power supply 11 to verify the operating status of the power supply 11 and confirm whether it is operating in a state where power output efficiency is optimized. If the real-time status 1140 is different from the meaning of the dialog 20, meaning that the power supply 11 has not enabled the power output efficiency optimization function during the gaming session, the CPU 10 then transmits the action instruction 101 to the control chip 114 via the Copilot tool, instructing the power supply 11 to automatically increase its output voltage in response to the increased load during the gaming session. In other words, after receiving the action command 101, the control chip 114 will start to automatically adjust the output voltage during the entire process of game execution. When the output of the power supply 11 reaches an overload and the output voltage drops from +12V to 11.6V, the control chip 114 will automatically adjust the output voltage to 12.4V. When it drops from +12V to 11.4V, the control chip 114 will automatically adjust the output voltage to 12.6V. When it drops from +12V to 11.12V, the control chip 114 will automatically adjust the output voltage to 12.8V. The increase in load causes the voltage to drop, and the voltage drop is used to linearly increase the voltage proportionally. This is used to predict a heavier load in the future and prevent a further increase in voltage due to a subsequent heavy load, thereby reducing current, reducing power loss, and improving efficiency. In this embodiment, the voltage is increased by 100% of the voltage drop, preferably by 50% to 150% of the voltage drop, and the maximum voltage increase is limited to 15V. In this way, regardless of the load changes during the entire gaming process, the high voltage and low current state can be maintained at all times, optimizing the power output efficiency of the power supply 11. In step S54, after the power supply 11 activates the power output efficiency optimization function during gaming, the central processing unit 10 displays the action information in the dialogue interface 100 through the Copilot tool to inform the user that the system has executed the action command 101. On the other hand, if the CPU 10 checks the real-time status 1140 and finds that the power supply 11 is already operating in the power output efficiency optimization state, in step S55, the command is not executed, and the Copilot tool feeds back the no-action information through the dialogue interface 100 to inform the user that the power supply 11 has turned on the power output efficiency optimization function and will no longer repeat the action.

[0040] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of the present invention. Therefore, equivalent changes and modifications that do not depart from the spirit and scope of the present invention are intended to be encompassed within the scope of the present invention. The monitoring unit is implemented using hardware or software supplemented by hardware. Its definition essentially refers to the integration of various hardware devices, such as a microprocessor, memory, or signal transmitter, with the support of software programs for implementation.

Claims

1. An intelligent power supply system constructed using artificial intelligence-assisted tools, executed on a Windows system host, comprising a central processing unit (CPU) and a power supply. The power supply includes a control chip and a cooling fan electrically connected to the CPU and outputting an output voltage. The system is characterized by: The central processing unit (CPU) is embedded with an artificial intelligence-assisted tool based on a large language model and Microsoft graphics data, and has a dialogue interface. When the dialogue interface receives a dialogue input from an external source, the CPU parses the meaning of the dialogue and generates an action instruction. At this time, the CPU requests the control chip to verify and report the real-time status of the power supply. If the comparison shows that the meaning of the dialogue is different from the real-time status, the CPU transmits the action instruction to the control chip to drive the power supply to adjust the output voltage value or the operating status of the cooling fan.

2. The intelligent power supply system according to claim 1, wherein: The power supply is provided with a power manager electrically connected to the control chip for monitoring the load status of the power supply. Furthermore, the action command is a command for optimizing the power output efficiency during a period. When the central processing unit learns that the power supply has not enabled the power output efficiency optimization function during a period, the central processing unit transmits the action command to the control chip, thereby driving the power supply to automatically increase the output voltage value in response to an increase in load during this period. The central processing unit also displays an action message on the dialogue interface to inform the user that the action command has been executed.

3. The intelligent power supply system according to claim 2, wherein: The control chip periodically transmits the real-time status to the central processing unit, and the real-time status at least includes the working efficiency, load status and the output voltage value of the power supply. When the central processing unit detects the real-time status and finds that the working efficiency of the power supply is lower than an efficiency standard, an aging notification is output through the dialogue interface.

4. The intelligent power supply system according to claim 3, wherein: When the central processing unit detects the real-time status and finds that the output voltage value is not within a specification range, a voltage abnormality notification is output through the dialogue interface; when the central processing unit detects the real-time status and finds that the load state of the power supply is greater than or equal to a maximum rated output, an overload notification is output through the dialogue interface.

5. The intelligent power supply system according to claim 3, wherein: When the central processing unit detects that the power supply is in an extremely light-load working state, the control chip is driven to turn off the power factor correction function of the power supply and a PFC turn-off notification is sent through the dialogue interface.

6. The intelligent power supply system according to claim 1, wherein: The action command is a command for turning on / off the fan energy-saving function. When the central processing unit checks the real-time status and finds that the power supply has not turned on / off the energy-saving function of the cooling fan, it transmits the action command to the control chip to turn on / off the energy-saving function of the cooling fan, and displays an action message on the dialogue interface to inform that the action command has been executed. Conversely, when the central processing unit finds that the energy-saving function of the cooling fan has been turned on / off, it does not execute the command and returns a no-action message.

7. The intelligent power supply system according to claim 1, wherein: The cooling fan is an RGB fan having a plurality of LED lights, the plurality of LED lights being configured to emit lights of a plurality of colors. The action command is a command for turning on / off the fan's RGB function. When the central processing unit checks the real-time status and learns that the power supply has not turned on / off the plurality of LED lights of the cooling fan, the central processing unit transmits the action command to the control chip to turn on / off the plurality of LED lights of the cooling fan, and displays an action message through the dialogue interface to inform that the action command has been executed. Conversely, when the central processing unit checks and learns that the RGB function of the cooling fan has been turned on / off, the command is not executed and a no-action message is fed back. Furthermore, the action command is a command for switching the light colors of the plurality of LED lights. When the central processing unit checks the real-time status and learns that the light colors of the plurality of LED lights are different from the meaning of the dialogue, the central processing unit transmits the action command to the control chip to switch the light colors of the plurality of LED lights of the cooling fan.

8. The intelligent power supply system according to claim 1, wherein: The power supply is equipped with a thermal sensor for sensing an internal operating temperature of the power supply, and the central processing unit stores a temperature warning value of the power supply reported by the control chip. When the cooling fan is in an energy-saving mode, the central processing unit detects the output wattage of the power supply and the operating temperature. When it determines that the workload is between 0% and 30%, it stops the cooling fan to reduce the operating noise to below 5dBA. The central processing unit periodically requests the control chip to report the operating temperature via the real-time status. When it detects that the operating temperature exceeds the corresponding temperature warning value, it transmits an emergency warning message via the dialogue interface to notify the user to perform emergency operations.

9. The intelligent power supply system according to claim 1, wherein: The central processing unit stores a preset speed value of the cooling fan reported by the control chip, and the central processing unit requests the control chip to report the speed of the cooling fan, so that when it is detected that the speed of the cooling fan does not meet the preset speed value, a fan abnormality message is transmitted through the dialogue interface to inform the fan abnormality.

10. The intelligent power supply system according to claim 1, wherein: The central processing unit (CPU) performs a rolling analysis of the operating performance of the Windows system host to obtain at least one high-load period, and sequentially records energy consumption data of a plurality of operating programs in the Windows system host. The CPU analyzes the plurality of energy consumption data to obtain an operating probability of each operating program corresponding to the high-load period, and further summarizes the plurality of operating programs with a high probability of running during the high-load period to form a predicted execution table based on the plurality of operating probabilities. Accordingly, the CPU outputs a predicted execution query before the high-load period, for querying the execution intention of each operating program in the predicted execution table through the dialogue interface.

Images