Ai direct-drive water-cooling heat dissipation system
Patent Information
- Application Number
- TW115205706
- Authority / Receiving Office
- TW · TW
- Patent Type
- Utility models
- Current Assignee / Owner
- Filing Date
- 2026-06-22
- Publication Date
- 2026-09-01
- Estimated Expiration
- 2036-06-21
Smart Images
Figure TWG2TB001909595_001 
Figure TWG2TB001909595_002 
Figure TWG2TB001909595_003
Abstract
Claims
1. An AI direct-drive water cooling system, comprising: A water-cooling radiator, disposed in a heat dissipation area, is configured to dissipate heat from a plurality of circulating pipes; a fan, configured to cool the water-cooling radiator; a water pump, connected to the circulating pipes and configured to drive coolant to flow in the circulating pipes, the circulating pipes passing through a heat source of a computer host and the heat dissipation area; at least one direct-drive motor, disposed at at least one of the fan and the water pump, each direct-drive motor including a plurality of coil groups; and a water-cooling control system, including: a storage device storing an edge machine learning model; a communication interface configured to connect to the computer host; a controller electrically connected to the at least one direct-drive motor, the storage device, and the communication interface, including: a data receiving module configured to receive thermal data from the computer host through the communication interface; and a command generation module configured to input the thermal data into the edge machine learning model to generate a drive command for each direct-drive motor, the drive command corresponding to a target speed. A status detection module is configured to detect a current state of each of the direct drive motors; and a motor control module is configured to control a drive current of each of the direct drive motors based on the current state and the drive command, and to control the conduction state of the coil groups of each of the direct drive motors, so as to make each of the direct drive motors operate to the target speed.
2. The AI direct-drive water-cooling system as claimed in claim 1, wherein the direct-drive motor is a three-phase brushless DC motor, and the coil groups include a U-phase coil group, a V-phase coil group, and a W-phase coil group; the system further includes a three-phase full-bridge circuit electrically connecting the controller and the direct-drive motor, including a plurality of switching elements, which are used to control the conduction state of the coil groups respectively.
3. The AI direct-drive water cooling system as described in claim 2, wherein the switching elements include three high-side switches and three low-side switches, and each of the U-phase coil group, the V-phase coil group and the W-phase coil group corresponds to one high-side switch and one low-side switch respectively; The controller further includes: A commutation control module is configured to continuously execute a commutation control based on the current state to sequentially switch the current direction of the U-phase coil group, the V-phase coil group and the W-phase coil group, wherein the commutation control includes a U-phase to V-phase step, a U-phase to W-phase step, a V-phase to W-phase step, a V-phase to U-phase step, a W-phase to U-phase step and a W-phase to V-phase step; And a first delay module is configured to delay a commutation delay time after setting one of the high-side switch and the low-side switch of the same coil to be open and before setting the other to be closed, so as to prevent the high-side switch and the low-side switch from forming a direct circuit.
4. The AI direct-drive water-cooling system as described in claim 2 further includes a plurality of magnetic sensing elements electrically connected to the controller, configured to detect the magnetic pole position of a rotor of the direct-drive motor and output a corresponding magnetic pole state as the current state; the controller further includes: A lookup table module is configured to look up a commutation table based on the magnetic pole state to determine the next state of each switching element, and set the switching element as closed or open based on the next state.
5. The AI direct-drive water cooling system as described in claim 4, wherein the magnetic sensing elements include three Hall elements disposed at three different positions of a stator of the direct-drive motor to sense the magnetic poles of the rotor and generate three magnetic sensing signals at the three positions as the magnetic pole state.
6. The AI direct-drive water cooling system as described in claim 2, wherein the controller further comprises: An electromotive force detection module is configured to detect the back electromotive force of the non-conducting coil in the U-phase coil group, the V-phase coil group and the W-phase coil group; And a second delay module is configured to obtain a commutation delay time when the back electromotive force of the non-conducting device meets a zero-crossing condition, and after the commutation delay time has elapsed, set the switching elements to perform commutation.
7. The AI direct-drive water cooling system as described in claim 2, wherein the controller further comprises: A prepositioning module is configured to turn on two of the coil groups for at least a prepositioning time; And a back EMF trigger module is configured to gradually increase the drive current and perform commutation until the back EMF of the coil groups is detected.
8. The AI direct-drive water-cooling system as described in claim 3 further includes an analog-to-digital converter electrically connected to the controller; wherein, The controller further includes: a first conversion module configured to acquire the three-phase current value of the direct drive motor and convert the three-phase current value into two-dimensional coordinates in a planar coordinate system; and a second conversion module configured to calculate a direct-axis current and a quadrature-axis current based on the two-dimensional coordinates and a current electrical angle of a rotor of the direct drive motor; wherein the controller is configured to minimize the direct-axis current and make the quadrature-axis current correspond to a target torque corresponding to the drive command.
9. The AI direct-drive water-cooling system as described in claim 3, wherein the edge machine learning model is configured to determine one of a plurality of operating modes corresponding to different cooling strategies based on the thermal data, and the drive command corresponds to the cooling strategy corresponding to the determined operating mode; wherein, The multiple operating modes include a performance mode and a silent mode. In the performance mode, the controller directly switches the conduction state of the switching elements through the commutation control module to perform the commutation control without applying carrier modulation to the switching elements. In the silent mode, the controller controls the drive current of each direct drive motor by applying carrier modulation based on the direct axis current and the quadrature axis current.
10. The AI direct-drive water cooling system as claimed in claim 1, wherein the at least one direct-drive motor includes a water pump motor and a fan motor; the controller is configured to coordinately control the target speed of each of the water pump motor and the fan motor, and under the same heat dissipation requirements, when the target speed of the water pump motor decreases, increase the target speed of the fan motor to dynamically maintain the heat dissipation capacity.