WO2024032655A1 - 一种风扇监控方法、系统、装置、服务器及可读存储介质 - Google Patents

一种风扇监控方法、系统、装置、服务器及可读存储介质 Download PDF

Info

Publication number
WO2024032655A1
WO2024032655A1 PCT/CN2023/111986 CN2023111986W WO2024032655A1 WO 2024032655 A1 WO2024032655 A1 WO 2024032655A1 CN 2023111986 W CN2023111986 W CN 2023111986W WO 2024032655 A1 WO2024032655 A1 WO 2024032655A1
Authority
WO
WIPO (PCT)
Prior art keywords
fan
blade passing
duty cycle
passing frequency
pwm duty
Prior art date
Application number
PCT/CN2023/111986
Other languages
English (en)
French (fr)
Inventor
王羽茜
刘广志
吴安
黄家明
Original Assignee
苏州元脑智能科技有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 苏州元脑智能科技有限公司 filed Critical 苏州元脑智能科技有限公司
Publication of WO2024032655A1 publication Critical patent/WO2024032655A1/zh

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F11/00—Error detection; Error correction; Monitoring
    • G06F11/30—Monitoring
    • G06F11/3058—Monitoring arrangements for monitoring environmental properties or parameters of the computing system or of the computing system component, e.g. monitoring of power, currents, temperature, humidity, position, vibrations
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F11/00—Error detection; Error correction; Monitoring
    • G06F11/30—Monitoring
    • G06F11/32—Monitoring with visual or acoustical indication of the functioning of the machine
    • G06F11/321—Display for diagnostics, e.g. diagnostic result display, self-test user interface
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management

Definitions

  • This application relates to the field of fan monitoring, and in particular to a fan monitoring method, system, device, server and readable storage medium.
  • the fan is an important heat dissipation component in the server. Once it fails, it will cause the server to make abnormal noise, system errors, and even shut down due to overheating protection.
  • the fan in the server can only feed back the rotational speed signal to the system through the Tach terminal, thereby adjusting and monitoring the fan rotational speed.
  • the rotational speed signal cannot provide good feedback on the health status of the fan. Therefore, in order to monitor the health of the fan, additional voltage and current measurement circuits, clock circuits, AD (Digital to Analog, analog-to-digital conversion) sampling calibration circuits, humidity-sensitive capacitors, etc. are needed to make the already crowded server motherboard It becomes more difficult to lay out and design, and the hardware requires excessive system resources.
  • the purpose of this application is to provide a fan monitoring method, system, device, server and readable storage medium that can improve the comprehensiveness of fan fault diagnosis. Fan fault diagnosis can be completed only through the noise signal collected by the microphone.
  • the hardware architecture is simple and does not require excessive hardware resources.
  • the fan monitoring method includes:
  • the fan monitoring method further includes:
  • the process of generating fan status diagnostic prompt information based on diagnostic data and fan blade passing frequency includes:
  • the fan monitoring method further includes:
  • the process of calculating the deviation percentage between the fan blade passing frequency and the reference fan blade passing frequency includes:
  • the process of determining the reference fan blade passing frequency includes:
  • the first target PWM duty cycle and the second target PWM duty cycle are adjacent, the first target PWM duty cycle is smaller than the current PWM duty cycle, and the second target PWM duty cycle is greater than the current PWM duty cycle.
  • the reference fan blade pass frequency corresponding to the current PWM duty cycle is calculated based on the initial fan blade pass frequency corresponding to the first target PWM duty cycle and the initial fan blade pass frequency corresponding to the second target PWM duty cycle.
  • Frequency processes include:
  • a difference calculation is performed on the first initial fan blade passing frequency and the second initial fan blade passing frequency to obtain the reference fan blade passing frequency corresponding to the current PWM duty cycle.
  • the signal characteristic data includes time domain characteristic data, frequency domain characteristic data and timely frequency domain characteristic data.
  • the process of obtaining the passing frequency of the fan blades based on the noise signal includes:
  • the fan blade passing frequency is calculated based on the spectrum data.
  • FFT processing is performed on the noise signal to obtain spectrum data, including:
  • the diagnostic data includes health status, or fault status and fault causes corresponding to the fault status.
  • the cause of the failure includes one or more of blade eccentricity, bearing wear, winding performance degradation, lubricating oil shortage or depletion, and IC component resistance changes.
  • the fan monitoring method further includes:
  • the fan speed is adjusted by the frequency of fan blade passage.
  • the frequency of fan blade passage is proportional to the rotational speed of the fan.
  • the process of pre-building a state diagnosis model includes:
  • the noise samples include fault noise samples and non-fault noise samples, and add corresponding labels to the fault noise samples and non-fault noise samples;
  • Extract feature data from each noise sample combine the feature data into matrix samples, and divide the matrix samples into first matrix samples and second matrix samples;
  • corresponding labels are added to fault noise samples and non-fault noise samples, including:
  • the microphone is provided on a side of the server motherboard close to the fan.
  • the fan monitoring system includes:
  • the acquisition module is used to acquire the fan noise signal collected by the microphone
  • the extraction module is used to obtain the signal characteristic data and the fan blade passing frequency based on the noise signal
  • the diagnostic module is used to input signal characteristic data into a pre-built state diagnostic model to obtain diagnostic data;
  • An information generation module is used to generate fan status diagnostic prompt information based on diagnostic data and fan blade passing frequency.
  • this application also provides a fan monitoring device, including:
  • Memory used to store computer programs
  • a processor is used to implement the steps of any of the above fan monitoring methods when executing a computer program.
  • this application also provides a server, including the fan monitoring device as above.
  • this application also provides a non-volatile readable storage medium.
  • the non-volatile readable storage medium stores a computer program.
  • the computer program is executed by the processor, any one of the above is implemented. Steps of fan monitoring method.
  • This application provides a fan monitoring method, which is applied to BMC.
  • the microphone is used to obtain the noise signal of the fan. Based on the noise signal, the signal characteristic data and the fan blade passing frequency of the fan are extracted.
  • the signal characteristic data is input into the preset diagnostic model to obtain the fan's Diagnostic data, based on the diagnostic data and fan blade passing frequency, jointly determines the status diagnosis result of the fan, making the diagnosis more comprehensive. Fault diagnosis can be completed only through the noise signal collected by the microphone.
  • the hardware architecture is simple and does not require excessive hardware resources.
  • This application also provides a fan monitoring system, device, server and readable storage medium, which has the same beneficial effects as the above fan monitoring method.
  • Figure 1 is a step flow chart of a fan monitoring method provided by this application.
  • FIG. 2 is a schematic structural diagram of a BMC provided by this application.
  • Figure 3 is a flow chart provided by this application for obtaining the reference fan blade passing frequency corresponding to the current PWM duty cycle
  • FIG. 4 is a schematic structural diagram of a fan monitoring system provided by this application.
  • Figure 5 is a schematic structural diagram of a non-volatile readable storage medium provided by this application.
  • the core of this application is to provide a fan monitoring method, system, device, server and readable storage medium, which can improve the comprehensiveness of fan fault diagnosis. Fan fault diagnosis can be completed only through the noise signal collected by the microphone.
  • the hardware architecture is simple and does not require excessive hardware resources.
  • FIG 1 is a step flow chart of a fan monitoring method provided by this application.
  • the fan monitoring method can be applied to PCs (Personal Computers, personal computers), edge servers and other electronic products that use fans as heat dissipation devices.
  • PCs Personal Computers, personal computers
  • edge servers and other electronic products that use fans as heat dissipation devices.
  • the above fan monitoring method can be implemented through BMC (Baseboard Management Controller, baseboard management controller).
  • BMC Baseboard Management Controller, baseboard management controller
  • the structure of BMC is shown in Figure 2, including Fourier transform unit, storage unit, feature extraction unit, BPF (Blade Passing Frequency, fan blade Through frequency) extraction unit, rotational speed deviation calculation unit, feature matching and analysis unit.
  • This fan monitoring method includes:
  • the server motherboard is integrated with one or more microphones for collecting noise signals from the fan.
  • the microphone can be placed on the side of the server motherboard close to the fan to facilitate the collection of noise signals.
  • the microphone collects the noise signal of the fan according to a preset period.
  • the preset period can be set to 1 hour, and the sampling time can be set to 10 seconds.
  • the preset period and sampling time can be set according to actual needs, and this application does not impose specific restrictions here.
  • BMC After BMC obtains the noise signal, it inputs the noise signal into the Fourier transform unit and the feature extraction unit respectively, and extracts signal features of the noise signal through the feature extraction unit to obtain signal feature data.
  • the signal feature data includes but is not limited to time. Domain feature data, frequency domain feature data and timely frequency domain feature data, specifically including but not limited to kurtosis indicators, PSD (Power Spectral Density, power spectral density) discrete peaks and frequencies, PSD and time waterfall charts, etc.
  • the noise signal collected by the microphone is a time domain signal
  • the noise signal is sent to the Fourier transform unit to perform FFT (Fast Fourier Transform, Fast Fourier Transform) processing on the noise signal, and the noise signal is transformed into a time domain signal.
  • the domain is converted into the frequency domain to obtain spectrum data, and the fan blade passing frequency is calculated based on the spectrum data.
  • the spectrum data with a frequency resolution of df is an array of N rows and 2 columns [df, p 1 ; 2df, p 2 ; 3df, p 3 ;...;Ndf, p N ].
  • S103 Input signal characteristic data into the pre-built state diagnosis model to obtain diagnosis data
  • the process of pre-building a state diagnosis model includes:
  • Noise samples include fault noise samples and non-fault noise samples. and add corresponding labels to fault noise samples and non-fault noise samples;
  • Extract feature data from each noise sample combine the feature data into matrix samples, and divide the matrix samples into first matrix samples and second matrix samples;
  • the label of the non-fault sample is non-fault, and the label of the fault sample indicates the specific cause of the fault, such as blade eccentricity, bearing wear, winding performance degradation, insufficient or dry lubricating oil.
  • signal feature extraction is performed, including time domain features, frequency domain features and time-frequency domain features; the kurtosis index, PSD discrete peak and frequency, PSD and Waterfall chart of time, these feature data are combined into matrices respectively; input most of the matrix samples to the classifier algorithm for training, and a series of models will be output after the model training is completed; load the remaining matrix samples into the model generated in the previous step, Carry out testing; select the best-performing model based on the test results, and save the model code in a dedicated storage unit of BMC.
  • the signal characteristic data obtained based on the noise signal is input into the above-trained preset diagnosis model.
  • the preset diagnosis model can output the diagnostic data of the fan.
  • the diagnostic data includes health status, or fault status and fault.
  • the cause of the failure corresponding to the status.
  • the cause of the failure includes but is not limited to one or more of blade eccentricity, bearing wear, winding performance degradation, insufficient or depleted lubricating oil, and changes in IC component resistance, so that staff can know the status of the fan in a timely manner. , and handle it in a timely manner.
  • S104 Generate fan status diagnostic prompt information based on the diagnostic data and fan blade passing frequency.
  • this application obtains the status diagnosis prompt information of the fan based on the diagnostic data and the fan blade passing frequency, thereby making the diagnosis results more accurate.
  • Comprehensive, including status diagnosis prompt information including alarm information, fan failure logs, etc.
  • the fan monitoring method uses a microphone to obtain the noise signal of the fan, extracts the signal characteristic data and the fan blade passing frequency of the fan based on the noise signal, and inputs the signal characteristic data into the preset diagnostic model to obtain the fan's Diagnostic data, based on the diagnostic data and fan blade passing frequency, jointly determines the status diagnosis result of the fan, making the diagnosis more comprehensive. Fault diagnosis can be completed only through the noise signal collected by the microphone.
  • the hardware architecture is simple and does not require excessive hardware resources.
  • the fan monitoring method further includes:
  • the process of generating fan status diagnostic prompt information based on diagnostic data and fan blade passing frequency includes:
  • the deviation of the fan blade passing frequency is the deviation percentage of the current fan blade passing frequency extracted from the current noise signal and the reference fan blade passing frequency under the same PWM duty cycle. It can be understood that when the deviation percentage is small, it means that the fan is under normal control. When the deviation percentage is greater than the preset threshold, it means that the fan is out of control for some reasons. At this time, the fan is abnormal. Therefore, the deviation percentage can also be used as a fan. The basis for the failure. Corresponding prompt information can be generated based on the deviation percentage and diagnostic data. It is understandable that the prompt information includes fault status and corresponding alarm information when the deviation percentage is too large.
  • the fan monitoring method further includes:
  • the process of calculating the deviation percentage between the fan blade passing frequency and the reference fan blade passing frequency includes:
  • the sensitivity of the microphone is verified after the server product is assembled and before leaving the factory; after the microphone sensitivity verification passes, the noise signal of the fan collected by the microphone is acquired.
  • the sensitivity of the microphone needs to be calibrated at a suitable temperature and humidity. Suitable here means as close as possible to the actual working environment.
  • it is necessary to start the step frequency sweep of the fan collect the time domain signals collected by the microphone under different PWM duty cycles, extract the fan blade passing frequency BPF in its initial state, and save the PWM-BPF mapping table to the BMC's dedicated in the storage unit.
  • mapping table is saved in the dedicated storage unit of the BMC.
  • the process of determining the reference fan blade passing frequency includes:
  • the first target PWM duty cycle and the second target PWM duty cycle are adjacent, the first target PWM duty cycle is smaller than the current PWM duty cycle, and the second target PWM duty cycle is greater than the current PWM duty cycle.
  • the reference fan blade pass frequency corresponding to the current PWM duty cycle is calculated based on the initial fan blade pass frequency corresponding to the first target PWM duty cycle and the initial fan blade pass frequency corresponding to the second target PWM duty cycle.
  • Frequency processes include:
  • a difference calculation is performed on the first initial fan blade passing frequency and the second initial fan blade passing frequency to obtain the reference fan blade passing frequency corresponding to the current PWM duty cycle.
  • the PWM duty cycle may not be in the PWM-BPF mapping table. Therefore, when determining the current PWM duty After comparison, first determine whether the current PWM duty cycle exists in the mapping table. If it exists, directly match the initial blade passing frequency with the same PWM duty cycle as the reference blade passing frequency corresponding to the current PWM duty cycle. If the current PWM duty cycle obtained in actual work is not in the mapping table, use the two initial fan blades with PWM duty cycles that are closest to the current PWM duty cycle in the mapping table to calculate the frequency difference between them and the current PWM duty cycle.
  • the reference fan blade passing frequency corresponding to the duty cycle Assuming that the current PWM duty cycle is 34%, 34% can be obtained by differential calculation between the initial fan blade passing frequency corresponding to 30% and the initial fan blade passing frequency corresponding to 40% in the mapping table. Corresponding reference fan blade passing frequency.
  • the current PWM duty cycle value is retrieved from the BMC, and the corresponding BPF value is retrieved from the mapping table in the storage unit. If the current PWM duty cycle is not in the pre-stored parameter table, the BPF can be calculated differentially.
  • the current PWM duty cycle (replaced by W) is 34, and the rounding function can be used to calculate the closest PWM in the mapping table, that is, the first target PWM duty cycle and the second target PWM duty cycle, using the first target
  • the BPF corresponding to the current PWM duty cycle is obtained by difference calculation between the first initial fan blade passing frequency corresponding to the PWM duty cycle and the second initial fan blade passing frequency corresponding to the second target PWM duty cycle.
  • the fan monitoring method further includes:
  • the fan speed is adjusted by the frequency of fan blade passage.
  • the fan blade passing frequency is proportional to the rotation speed, and the deviation of the fan blade passing frequency is the deviation of the rotation speed. Therefore, this step can monitor and adjust the fan speed through the fan blade passing frequency. Using this solution, There is no need to install a speed control circuit or FG signal terminal in the fan, and the hardware structure is simple.
  • the solution of this application is used to integrate one or more microphones on the server motherboard, collect the fan noise, and judge the health status of the fan through noise characteristic analysis, and pre-store the model parameters and fault labels in the BMC.
  • the status of the fan is analyzed online, and an alarm is issued after identifying a fault or fault trend.
  • the rotation speed can be calculated through the signal collected by the microphone, replacing the reading component on the fan PCB.
  • the hardware structure is simple and does not require too much system resources. , and at the same time, it can locate the specific cause of the failure and provide accurate suggestions for server maintenance.
  • Figure 4 is a schematic structural diagram of a fan monitoring system provided by this application and applied to BMC.
  • the fan monitoring system includes:
  • Acquisition module 1 is used to acquire the fan noise signal collected by the microphone
  • Extraction module 2 is used to obtain signal characteristic data and the fan blade passing frequency based on the noise signal
  • Diagnosis module 3 is used to input signal characteristic data into a pre-built state diagnosis model to obtain diagnostic data;
  • the information generation module 4 is used to generate status diagnosis prompt information of the fan based on the diagnostic data and the fan blade passing frequency.
  • the fan monitoring system uses a microphone to obtain the noise signal of the fan, extracts the signal characteristic data and the fan blade passing frequency based on the noise signal, and inputs the signal characteristic data into the preset diagnostic model to obtain the fan's Diagnostic data, based on the diagnostic data and fan blade passing frequency, jointly determines the status diagnosis result of the fan, making the diagnosis more comprehensive. Fault diagnosis can be completed only through the noise signal collected by the microphone.
  • the hardware architecture is simple and does not require excessive hardware resources.
  • the fan monitoring system also includes:
  • a calculation module used to calculate the deviation percentage between the fan blade passing frequency and the reference fan blade passing frequency
  • the process of generating fan status diagnostic prompt information based on diagnostic data and fan blade passing frequency includes:
  • the fan monitoring system also includes:
  • the preprocessing module is used to perform a step frequency sweep on the fan according to preset rules to obtain the fan's initial blade passing frequency under each PWM duty cycle; based on all PWM duty cycles corresponding to the step frequency sweep and each Initial fan blade passing frequency under PWM duty cycle rate, construct a mapping table between PWM duty cycle and fan blade passing frequency;
  • the process of calculating the deviation percentage between the fan blade passing frequency and the reference fan blade passing frequency includes:
  • the process of determining the reference fan blade passing frequency includes:
  • the first target PWM duty cycle and the second target PWM duty cycle are adjacent, the first target PWM duty cycle is smaller than the current PWM duty cycle, and the second target PWM duty cycle is greater than the current PWM duty cycle.
  • the reference fan blade pass frequency corresponding to the current PWM duty cycle is calculated based on the initial fan blade pass frequency corresponding to the first target PWM duty cycle and the initial fan blade pass frequency corresponding to the second target PWM duty cycle.
  • Frequency processes include:
  • a difference calculation is performed on the first initial fan blade passing frequency and the second initial fan blade passing frequency to obtain the reference fan blade passing frequency corresponding to the current PWM duty cycle.
  • the signal characteristic data includes time domain characteristic data, frequency domain characteristic data and timely frequency domain characteristic data.
  • the process of obtaining the passing frequency of the fan blades based on the noise signal includes:
  • the fan blade passing frequency is calculated based on the spectrum data.
  • the diagnostic data includes health status, or fault status and fault causes corresponding to the fault status.
  • the cause of the failure includes one or more of blade eccentricity, bearing wear, winding performance degradation, lubricating oil shortage or depletion, and IC component resistance changes.
  • the fan monitoring system also includes:
  • a control module used to adjust the fan speed based on the fan blade passing frequency.
  • the process of pre-building a state diagnosis model includes:
  • the noise samples include fault noise samples and non-fault noise samples, and add corresponding labels to the fault noise samples and non-fault noise samples;
  • Extract feature data from each noise sample combine the feature data into matrix samples, and divide the matrix samples into first matrix samples and second matrix samples;
  • this application also provides a fan monitoring device, including:
  • Memory used to store computer programs
  • a processor configured to implement the steps of the fan monitoring method described in any of the above embodiments when executing a computer program.
  • the memory includes non-volatile storage media and internal memory.
  • the non-volatile storage medium stores an operating system and computer-readable instructions
  • the internal memory provides an environment for the execution of the operating system and computer-readable instructions in the non-volatile storage medium.
  • the processor provides computing and control capabilities for the fan monitoring device.
  • the following steps can be implemented: obtain the fan noise signal collected by the microphone; obtain signal characteristic data and the fan blade passing frequency based on the noise signal; Input signal characteristic data into the pre-built status diagnosis model to obtain diagnostic data; generate fan status diagnosis prompt information based on the diagnostic data and fan blade passing frequency.
  • the fan monitoring device uses a microphone to obtain the noise signal of the fan, extracts the signal characteristic data and the fan blade passing frequency based on the noise signal, and inputs the signal characteristic data into the preset diagnostic model to obtain the fan's Diagnostic data, based on the diagnostic data and fan blade passing frequency, jointly determines the status diagnosis result of the fan, making the diagnosis more comprehensive. Fault diagnosis can be completed only through the noise signal collected by the microphone.
  • the hardware architecture is simple and does not require excessive hardware resources.
  • the processor executes the computer subroutine stored in the memory, the following steps can be implemented: calculate the deviation percentage of the fan blade passing frequency and the reference fan blade passing frequency; generate status diagnostic prompt information of the fan based on the diagnostic data and the deviation percentage.
  • the processor executes the computer subroutine stored in the memory
  • the following steps can be implemented: determine whether the current PWM duty cycle corresponding to the fan blade passing frequency exists in the mapping table; if so, compare the mapping table with The initial fan blade passing frequency corresponding to the current PWM duty cycle is used as the reference fan blade passing frequency; if not, the first target PWM duty cycle and the second target PWM duty cycle in the mapping table are determined based on the current PWM duty cycle; according to Calculate the reference fan blade passing frequency corresponding to the current PWM duty cycle from the initial fan blade passing frequency corresponding to the first target PWM duty cycle and the initial fan blade passing frequency corresponding to the second target PWM duty cycle; where, the first target PWM duty cycle The duty cycle is adjacent to the second target PWM duty cycle, the first target PWM duty cycle is smaller than the current PWM duty cycle, and the second target PWM duty cycle is greater than the current PWM duty cycle.
  • the processor executes the computer subroutine stored in the memory, the following steps can be implemented: determine the first initial fan blade passing frequency corresponding to the first target PWM duty cycle and the second target PWM duty cycle. The corresponding second initial fan blade passing frequency; perform a differential calculation on the first initial fan blade passing frequency and the second initial fan blade passing frequency to obtain the reference fan blade passing frequency corresponding to the current PWM duty cycle.
  • the processor when the processor executes the computer subroutine stored in the memory, the following steps can be implemented: obtaining signal characteristic data based on the noise signal; the signal characteristic data includes time domain characteristic data, frequency domain characteristic data, and frequency domain characteristics. data.
  • the processor when the processor executes the computer subroutine stored in the memory, the following steps can be implemented: perform FFT processing on the noise signal to obtain spectrum data; and calculate the fan blade passing frequency based on the spectrum data.
  • the processor executes the computer subroutine stored in the memory, the following steps can be implemented: input signal characteristic data into a pre-built state diagnosis model to obtain diagnostic data; the diagnostic data includes health status, or fault status and the fault cause corresponding to the fault status.
  • the processor executes the computer subroutine stored in the memory
  • the following steps can be implemented: input signal characteristic data into a pre-built state diagnosis model to obtain diagnostic data;
  • the diagnostic data includes health status, or fault status and the cause of the failure corresponding to the failure state;
  • the cause of the failure includes one or more of blade eccentricity, bearing wear, winding performance degradation, insufficient or depleted lubricating oil, and changes in IC component resistance.
  • the processor when the processor executes the computer subroutine stored in the memory, the following steps can be implemented: adjusting the rotational speed of the fan based on the frequency of fan blade passage.
  • the processor executes the computer subroutine stored in the memory
  • the following steps can be implemented: obtain noise samples of the fan in the target electronic device.
  • the noise samples include fault noise samples and non-fault noise samples, and analyze Add corresponding labels to the fault noise samples and non-fault noise samples; extract the characteristic data in each noise sample, combine the characteristic data into matrix samples, and divide the matrix samples into the first matrix sample and the second matrix sample; divide the first matrix sample into the first matrix sample and the second matrix sample.
  • the matrix samples are input into the classifier for training to obtain multiple models; the second matrix samples are loaded into multiple models for testing, and the optimal model is selected as the state diagnosis model based on the test results.
  • the fan monitoring device further includes:
  • the input interface is connected to the processor and is used to obtain externally imported computer programs, parameters and instructions, and save them to the memory under the control of the processor.
  • the input interface can be connected to an input device to receive parameters or instructions manually input by the user.
  • the input device may be a touch layer covered on the display screen, or may be a button, trackball or touch pad provided on the terminal housing.
  • the display unit is connected to the processor and used to display data sent by the processor.
  • the display unit may be a liquid crystal display or an electronic ink display.
  • the network port is connected to the processor and is used to communicate with external terminal devices.
  • the communication technology used in the communication connection can be wired communication technology or wireless communication technology, such as Mobile High Definition Link Technology (MHL), Universal Serial Bus (USB), High Definition Multimedia Interface (HDMI), Wireless Fidelity Technology (WiFi), Bluetooth communication technology, low-power Bluetooth communication technology, communication technology based on IEEE802.11s, etc.
  • this application also provides a server, including the fan monitoring device as described in the above embodiment.
  • a server provided by this application has the same beneficial effects as the above-mentioned fan monitoring device.
  • this application also provides a non-volatile readable storage medium.
  • the non-volatile readable storage medium stores a computer program 51.
  • the computer program 51 When the computer program 51 is executed by the processor, it implements any of the above embodiments. Steps of the described fan monitoring method.
  • the non-volatile readable storage medium may include: U disk, mobile hard disk, read-only memory (Read-Only Memory, ROM), random access memory (Random Access Memory, RAM), magnetic disk or optical disk, etc.
  • the storage medium stores a computer program. When the computer program is executed by the processor, the following steps are implemented: obtain the noise signal of the fan collected by the microphone; obtain the signal characteristic data and the fan blade passing frequency based on the noise signal; input the signal characteristic data in advance Construct a state diagnosis model to obtain diagnostic data; based on the diagnostic data and fan blades generate fan status diagnostic prompt information based on frequency.
  • this embodiment uses a microphone to obtain the noise signal of the fan, extracts the signal characteristic data and the fan blade passing frequency based on the noise signal, and inputs the signal characteristic data into the preset diagnostic model to obtain the diagnostic data of the fan. According to the diagnostic data and the fan blade The fan status diagnosis results are jointly determined by frequency, making the diagnosis more comprehensive. Fault diagnosis can be completed only through the noise signal collected by the microphone.
  • the hardware architecture is simple and does not require excessive hardware resources.
  • the following steps can be implemented: calculate the deviation percentage between the fan blade passing frequency and the reference fan blade passing frequency; based on The diagnostic data and deviation percentage generate status diagnostic prompts for the fan.
  • the following steps can be implemented: stepwise sweep the fan according to preset rules, and obtain the frequency of the fan every time.
  • the initial fan blade passing frequency under a PWM duty cycle; based on all PWM duty cycles corresponding to the step sweep and the initial fan blade passing frequency under each PWM duty cycle, the PWM duty cycle and the fan blade passing frequency are constructed mapping table; determine the current PWM duty cycle corresponding to the fan blade passing frequency; determine the reference fan blade passing frequency based on the current PWM duty cycle and the initial fan blade passing frequency in the mapping table; calculate the fan blade passing frequency and the reference fan blade The deviation percentage from the passing frequency.
  • the following steps can be implemented: Determine whether the current PWM duty corresponding to the fan blade passing frequency exists in the mapping table Ratio; if yes, use the initial blade passing frequency corresponding to the current PWM duty cycle in the mapping table as the reference blade passing frequency; if not, determine the first target PWM duty cycle and sum in the mapping table based on the current PWM duty cycle The second target PWM duty cycle; calculate the reference fan blade pass frequency corresponding to the current PWM duty cycle based on the initial fan blade pass frequency corresponding to the first target PWM duty cycle and the initial fan blade pass frequency corresponding to the second target PWM duty cycle. frequency; wherein, the first target PWM duty cycle and the second target PWM duty cycle are adjacent, the first target PWM duty cycle is smaller than the current PWM duty cycle, and the second target PWM duty cycle is greater than the current PWM duty cycle.
  • the following steps can be implemented: Determine the first initial fan blade pass corresponding to the first target PWM duty cycle The second initial fan blade passing frequency corresponding to the frequency and the second target PWM duty cycle; perform a differential calculation on the first initial fan blade passing frequency and the second initial fan blade passing frequency to obtain the reference fan blade corresponding to the current PWM duty cycle. pass frequency.
  • the following steps can be implemented: obtaining signal characteristic data based on the noise signal; the signal characteristic data includes time domain characteristic data , frequency domain feature data and timely frequency domain feature data.
  • the following steps can be implemented: perform FFT processing on the noise signal to obtain spectrum data; calculate based on the spectrum data Fan blade passing frequency.
  • the following steps can be implemented: input signal characteristic data into a pre-built state diagnosis model to obtain diagnostic data; Diagnostic data includes health status, or fault status and fault causes corresponding to the fault status.
  • the following steps can be implemented: input signal characteristic data into a pre-built state diagnosis model to obtain diagnostic data; Diagnostic data includes health status, or fault status and fault causes corresponding to the fault status; fault causes include one or more of blade eccentricity, bearing wear, winding performance degradation, lubricating oil shortage or depletion, and IC component resistance changes.
  • the following steps can be implemented: adjusting the rotation speed of the fan based on the fan blade passing frequency.
  • the following steps can be implemented: Obtain the noise sample of the fan in the target electronic device, and the noise sample includes fault noise samples and non-fault noise samples, and add corresponding labels to the fault noise samples and non-fault noise samples; extract the characteristic data in each noise sample, combine the characteristic data into matrix samples, and divide the matrix samples into first matrix samples and the second matrix sample; input the first matrix sample into the classifier for training to obtain multiple models; load the second matrix sample into multiple models for testing, and select the optimal model as the state diagnosis model based on the test results.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Quality & Reliability (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Human Computer Interaction (AREA)
  • Computing Systems (AREA)
  • Control Of Positive-Displacement Air Blowers (AREA)

Abstract

一种风扇监控方法、系统、装置、服务器及可读存储介质，涉及风扇监控领域，风扇监控方法应用于BMC，包括：获取麦克风采集的风扇的噪音信号（S101）；基于噪音信号得到信号特征数据及风扇的扇叶通过频率(S102)；将信号特征数据输入预先构建的状态诊断模型，得到诊断数据(S103)；基于诊断数据和扇叶通过频率生成风扇的状态诊断提示信息(S104)。提高对风扇的故障诊断的全面性，仅通过麦克风采集到的噪音信号即可完成风扇的故障诊断，硬件架构简单，无需占用过多硬件资源。

Description

一种风扇监控方法、系统、装置、服务器及可读存储介质
相关申请的交叉引用
本申请要求于2022年08月09日提交中国专利局，申请号为202210947593.7，申请名称为“一种风扇监控方法、系统、装置、服务器及可读存储介质”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请涉及风扇监控领域，特别涉及一种风扇监控方法、系统、装置、服务器及可读存储介质。
背景技术
风扇作为服务器中的重要散热部件，一旦发生故障会导致服务器的噪声异常、系统报错甚至因过热保护而关机。目前服务器中的风扇只能通过Tach端子向系统反馈转速信号，从而实现对风扇转速的调节与监控，但转速信号不能很好地反馈风扇的健康状况。因此，为了监控风扇的健康状况，还需要额外增加电压、电流的测量电路、时钟电路、AD(Digital to Analog，模数转换)采样校准电路、湿敏电容等，使本来就拥挤不堪的服务器主板变得更加难以布局和设计，且硬件需要占用过多的系统资源。
因此，如何提供一种解决上述技术问题的方案是本领域技术人员目前需要解决的问题。
发明内容
本申请的目的是提供一种风扇监控方法、系统、装置、服务器及可读存储介质，能够提高对风扇的故障诊断的全面性，仅通过麦克风采集到的噪音信号即可完成风扇的故障诊断，硬件架构简单，无需占用过多硬件资源。
为解决上述技术问题，本申请提供了一种风扇监控方法，应用于BMC，该风扇监控方法包括：
获取麦克风采集的风扇的噪音信号；
基于噪音信号得到信号特征数据及风扇的扇叶通过频率；
将信号特征数据输入预先构建的状态诊断模型，得到诊断数据；
基于诊断数据和扇叶通过频率生成风扇的状态诊断提示信息。
在本申请一些实施例中，该风扇监控方法还包括：
计算扇叶通过频率和参考扇叶通过频率的偏差百分比；
基于诊断数据和扇叶通过频率生成风扇的状态诊断提示信息的过程包括：
基于诊断数据和偏差百分比生成风扇的状态诊断提示信息。
在本申请一些实施例中，该风扇监控方法还包括：
按预设规则对风扇进行步进扫频，获取风扇在每一PWM占空比下的初始扇叶通过频率；
基于步进扫频对应的所有PWM占空比及每一PWM占空比下的初始扇叶通过频率，构建PWM占空比与扇叶通过频率的映射表；
计算扇叶通过频率和参考扇叶通过频率的偏差百分比的过程包括：
确定扇叶通过频率对应的当前PWM占空比；
基于当前PWM占空比及映射表中的初始扇叶通过频率，确定参考扇叶通过频率；
计算扇叶通过频率和参考扇叶通过频率的偏差百分比。
在本申请一些实施例中，基于当前PWM占空比及映射表中的初始扇叶通过频率，确定参考扇叶通过频率的过程包括：
判断映射表中是否存在扇叶通过频率对应的当前PWM占空比；
若是，将映射表中与当前PWM占空比对应的初始扇叶通过频率作为参考扇叶通过频率；
若否，基于当前PWM占空比确定映射表中的第一目标PWM占空比和第二目标PWM占空比；
根据第一目标PWM占空比对应的初始扇叶通过频率及第二目标PWM占空比对应的初始扇叶通过频率计算当前PWM占空比对应的参考扇叶通过频率；
其中，第一目标PWM占空比和第二目标PWM占空比相邻，第一目标PWM占空比小于当前PWM占空比，第二目标PWM占空比大于当前PWM占空比。
在本申请一些实施例中，根据第一目标PWM占空比对应的初始扇叶通过频率及第二目标PWM占空比对应的初始扇叶通过频率计算当前PWM占空比对应的参考扇叶通过频率的过程包括：
确定第一目标PWM占空比对应的第一初始扇叶通过频率及第二目标PWM占空比对应的第二初始扇叶通过频率；
对第一初始扇叶通过频率和第二初始扇叶通过频率进行差分计算，得到当前PWM占空比对应的参考扇叶通过频率。
在本申请一些实施例中，信号特征数据包括时域特征数据、频域特征数据及时频域特征数据。
在本申请一些实施例中，基于噪音信号得到风扇的扇叶通过频率的过程包括：
对噪音信号进行FFT处理，得到频谱数据；
基于频谱数据计算扇叶通过频率。
在本申请一些实施例中，对噪音信号进行FFT处理，得到频谱数据，包括：
对噪音信号进行FFT处理，以将噪音信号由时域转换为频域，得到频谱数据。
在本申请一些实施例中，诊断数据包括健康状态，或，故障状态及故障状态对应的故障原因。
在本申请一些实施例中，故障原因包括叶片偏心、轴承磨损、绕组性能退化、润滑油不足或耗干、IC元件电阻变化中的一项或多项。
在本申请一些实施例中，该风扇监控方法还包括：
通过扇叶通过频率对风扇的转速进行调整。
在本申请一些实施例中，扇叶通过频率与风扇的转速呈正比。
在本申请一些实施例中，预先构建状态诊断模型的过程包括：
获取风扇在目标电子设备中的噪音样本，噪音样本包括故障噪音样本及非故障噪音样本，并对故障噪音样本和非故障噪音样本添加各自对应的标签；
提取各个噪音样本中的特征数据，将特征数据组合成矩阵样本，并将矩阵样本划分为第一矩阵样本和第二矩阵样本；
将第一矩阵样本输入分类器进行训练，得到多个模型；
将第二矩阵样本载入多个模型中进行测试，根据测试结果选择最优模型作为状态诊断模型。
在本申请一些实施例中，对故障噪音样本和非故障噪音样本添加各自对应的标签，包括：
将非故障噪音样本的标签设置为非故障；
将故障噪音样本的标签设置为故障原因。
在本申请一些实施例中，还包括：
在服务器产品完成组装后、出厂前对麦克风的灵敏度进行校验；
在麦克风灵敏度校验通过后，执行获取麦克风采集的风扇的噪音信号。
在本申请一些实施例中，麦克风设置在服务器主板上靠近风扇的一侧。
为解决上述技术问题，本申请还提供了一种风扇监控系统，应用于BMC，该风扇监控系统包括：
获取模块，用于获取麦克风采集的风扇的噪音信号；
提取模块，用于基于噪音信号得到信号特征数据及风扇的扇叶通过频率；
诊断模块，用于将信号特征数据输入预先构建的状态诊断模型，得到诊断数据；
信息生成模块，用于基于诊断数据和扇叶通过频率生成风扇的状态诊断提示信息。
为解决上述技术问题，本申请还提供了一种风扇监控装置，包括：
存储器，用于存储计算机程序；
处理器，用于执行计算机程序时实现如上文任意一项的风扇监控方法的步骤。
为解决上述技术问题，本申请还提供了一种服务器，包括如上文的风扇监控装置。
为解决上述技术问题，本申请还提供了一种非易失性可读存储介质，非易失性可读存储介质上存储有计算机程序，计算机程序被处理器执行时实现如上文任意一项的风扇监控方法的步骤。
本申请提供了一种风扇监控方法，应用于BMC，利用麦克风获取风扇的噪音信号，基于噪音信号提取信号特征数据及风扇的扇叶通过频率，将信号特征数据输入预设诊断模型可得到风扇的诊断数据，根据诊断数据及扇叶通过频率共同确定风扇的状态诊断结果，使得诊断更全面，仅通过麦克风采集到的噪音信号即可完成故障诊断，硬件架构简单，无需占用过多硬件资源。本申请还提供了一种风扇监控系统、装置、服务器及可读存储介质，具有和上述风扇监控方法相同的有益效果。
附图说明
为了更清楚地说明本申请实施例，下面将对实施例中所需要使用的附图做简单的介绍，显而易见地，下面描述中的附图仅仅是本申请的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。
图1为本申请所提供的一种风扇监控方法的步骤流程图；
图2为本申请所提供的一种BMC的结构示意图；
图3为本申请所提供的一种获取当前PWM占空比对应的参考扇叶通过频率的流程图；
图4为本申请所提供的一种风扇监控系统的结构示意图；
图5为本申请所提供的一种非易失性可读存储介质的结构示意图。
具体实施方式
本申请的核心是提供一种风扇监控方法、系统、装置、服务器及可读存储介质，能够提高对风扇的故障诊断的全面性，仅通过麦克风采集到的噪音信号即可完成风扇的故障诊断，硬件架构简单，无需占用过多硬件资源。
为使本申请实施例的目的、技术方案和优点更加清楚，下面将结合本申请实施例中的附 图，对本申请实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例是本申请一部分实施例，而不是全部的实施例。基于本申请中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都属于本申请保护的范围。
请参照图1，图1为本申请所提供的一种风扇监控方法的步骤流程图，该风扇监控方法可应用到PC(Personal Computer，个人计算机)、边缘服务器等使用风扇作为散热装置的电子产品中，为便于理解本申请的方案，以应用在服务器中为例进行说明。上述风扇监控方法可通过BMC(Baseboard Management Controller，基板管理控制器)实现，BMC的结构参照图2所示，包括傅里叶变换单元、存储单元、特征提取单元、BPF(Blade Passing Frequency，扇叶通过频率)提取单元、转速偏差计算单元、特征匹配与分析单元。该风扇监控方法包括：
S101：获取麦克风采集的风扇的噪音信号；
具体的，在服务器主板内集成有一颗或多颗用于采集风扇的噪音信号的麦克风，在本申请一些实施例中，可将麦克风设置在服务器主板上靠近风扇的一侧，便于噪音信号的采集。
具体的，麦克风按预设周期采集风扇的噪音信号，预设周期可以设置为1小时，采样时间可以设置为10s。
其中，预设周期和采样时间可以根据实际需要设置，本申请在此不做具体的限制。
S102：基于噪音信号得到信号特征数据及风扇的扇叶通过频率；
具体的，BMC获取到噪音信号后，将噪音信号分别输入傅里叶变换单元和特征提取单元，通过特征提取单元对噪音信号进行信号特征提取，得到信号特征数据，信号特征数据包括但不限于时域特征数据、频域特征数据及时频域特征数据，具体包括但不限于峭度指标、PSD(Power Spectral Density，功率谱密度)离散峰值和频率、PSD和时间的瀑布图等。
可以理解的是，麦克风采集到的噪音信号为时域信号，将该噪音信号发送到傅里叶变换单元对噪音信号进行FFT(Fast Fourier Transform，快速傅里叶变换)处理，将噪音信号由时域转换为频域，得到频谱数据，基于频谱数据计算扇叶通过频率，其中，频率分辨率为df的频谱数据为N行2列数组[df，p1；2df，p2；3df，p3；…；Ndf，pN]。
S103：将信号特征数据输入预先构建的状态诊断模型，得到诊断数据；
在执行本步骤之前还包括预先训练预设诊断模型的操作，在本申请的一些实施例中，预先构建状态诊断模型的过程包括：
获取风扇在目标电子设备中的噪音样本，噪音样本包括故障噪音样本及非故障噪音样 本，并对故障噪音样本和非故障噪音样本添加各自对应的标签；
提取各个噪音样本中的特征数据，将特征数据组合成矩阵样本，并将矩阵样本划分为第一矩阵样本和第二矩阵样本；
将第一矩阵样本输入分类器进行训练，得到多个模型；
将第二矩阵样本载入多个模型中进行测试，根据测试结果选择最优模型作为状态诊断模型。
具体的，收集足够多的故障与非故障的噪音样本，非故障样本的标签为非故障，故障样本的标签指示具体故障原因，如叶片偏心、轴承磨损、绕组性能退化、润滑油不足或耗干、IC元件电阻变化等；在标记好故障标签后进行信号特征提取，包括时域特征、频域特征和时频域特征；本实施例中应用的峭度指标、PSD离散峰值和频率、PSD和时间的瀑布图，这些特征数据分别组合成矩阵；将大部分矩阵样本输入给分类器算法进行训练，模型训练完成后会输出一系列模型；将其余的矩阵样本载入上一步生成的模型中，进行测试；根据测试结果挑选出表现最好的模型，将模型代码保存在BMC的专用存储单元中。
在实际应用过程中，将基于噪音信号得到的信号特征数据输入上述训练好的预设诊断模型，该预设诊断模型即可输出风扇的诊断数据，诊断数据包括健康状态，或，故障状态及故障状态对应的故障原因，其中故障原因包括但不限于叶片偏心、轴承磨损、绕组性能退化、润滑油不足或耗干、IC元件电阻变化中的一项或多项，以便工作人员及时了解风扇的状态，并及时做出处理。
S104：基于诊断数据和扇叶通过频率生成风扇的状态诊断提示信息。
可以理解的是，通过扇叶通过频率可以确定风扇是否存在通信故障，如通信线故障等，因此，本申请基于诊断数据和扇叶通过频率共同得到风扇的状态诊断提示信息，从而使诊断结果更全面，其中状态诊断提示信息包括报警信息风扇故障日志等。
可见，本实施例所提供的一种风扇监控方法，利用麦克风获取风扇的噪音信号，基于噪音信号提取信号特征数据及风扇的扇叶通过频率，将信号特征数据输入预设诊断模型可得到风扇的诊断数据，根据诊断数据及扇叶通过频率共同确定风扇的状态诊断结果，使得诊断更全面，仅通过麦克风采集到的噪音信号即可完成故障诊断，硬件架构简单，无需占用过多硬件资源。
在上述实施例的基础上：
在本申请一些实施例中，该风扇监控方法还包括：
计算扇叶通过频率和参考扇叶通过频率的偏差百分比；
基于诊断数据和扇叶通过频率生成风扇的状态诊断提示信息的过程包括：
基于诊断数据和偏差百分比生成风扇的状态诊断提示信息。
其中，扇叶通过频率的偏差即从当前噪音信号中提取到的当前扇叶通过频率与相同PWM占空比下的参考扇叶通过频率的偏差百分比。可以理解的是，偏差百分比较小时，说明风扇正常受控，偏差百分比大于预设阈值时，说明风扇已经由于一些原因不受控了，此时风扇存在异常，因此，偏差百分比也可以作为一个风扇故障的依据。可根据偏差百分比和诊断数据生成对应的提示信息，可以理解的是，提示信息中包括故障状态及偏差百分比过大时对应的报警信息。
在本申请一些实施例中，该风扇监控方法还包括：
按预设规则对风扇进行步进扫频，获取风扇在每一PWM(Pulse Width Modulation，脉宽调制技术)占空比下的初始扇叶通过频率；
基于步进扫频对应的所有PWM占空比及每一PWM占空比下的初始扇叶通过频率，构建PWM占空比与扇叶通过频率的映射表；
计算扇叶通过频率和参考扇叶通过频率的偏差百分比的过程包括：
确定扇叶通过频率对应的当前PWM占空比；
基于当前PWM占空比及映射表中的初始扇叶通过频率，确定参考扇叶通过频率；
计算扇叶通过频率和参考扇叶通过频率的偏差百分比。
在本申请一些实施例中，在服务器产品完成组装后、出厂前对麦克风的灵敏度进行校验；在麦克风灵敏度校验通过后，执行获取麦克风采集的风扇的噪音信号。
具体的，在服务器产品完成组装后、出厂前需要在适宜的温湿度下校准麦克风的灵敏度，这里的适宜是指尽量接近实际工作环境。另外需要启动风扇的步进扫频，采集不同PWM占空比下麦克风采集到的时域信号，提取其初始状态下的扇叶通过频率BPF，并将PWM-BPF的映射表保存到BMC的专用存储单元中。
具体的，启动风扇从10％到100％PWMduty的步进扫频，采集不同PWM占空比对应的时域信号，提取其初始状态下的BPF(扇叶通过频率)，并将PWM-BPF的映射表保存到BMC的专用存储单元中。映射表为一个10行2列的数组MatrixBPF＝[10％，F1；20％，F2；…，100％，F10]。
在本申请一些实施例中，基于当前PWM占空比及映射表中的初始扇叶通过频率，确定参考扇叶通过频率的过程包括：
判断映射表中是否存在扇叶通过频率对应的当前PWM占空比；
若是，将映射表中与当前PWM占空比对应的初始扇叶通过频率作为参考扇叶通过频率；
若否，基于当前PWM占空比确定映射表中的第一目标PWM占空比和第二目标PWM占空比；
根据第一目标PWM占空比对应的初始扇叶通过频率及第二目标PWM占空比对应的初始扇叶通过频率计算当前PWM占空比对应的参考扇叶通过频率；
其中，第一目标PWM占空比和第二目标PWM占空比相邻，第一目标PWM占空比小于当前PWM占空比，第二目标PWM占空比大于当前PWM占空比。
在本申请一些实施例中，根据第一目标PWM占空比对应的初始扇叶通过频率及第二目标PWM占空比对应的初始扇叶通过频率计算当前PWM占空比对应的参考扇叶通过频率的过程包括：
确定第一目标PWM占空比对应的第一初始扇叶通过频率及第二目标PWM占空比对应的第二初始扇叶通过频率；
对第一初始扇叶通过频率和第二初始扇叶通过频率进行差分计算，得到当前PWM占空比对应的参考扇叶通过频率。
具体的，考虑到构建PWM-BPF的映射表时，是按照步进扫频的方式进行的，实际工作中PWM占空比可能不在PWM-BPF的映射表中，因此，在确定当前PWM占空比后，首先判断映射表中是否存在当前PWM占空比，如果存在，则直接匹配相同的PWM占空比的初始扇叶通过频率作为当前PWM占空比对应的参考扇叶通过频率。若实际工作中获取到的当前PWM占空比不在映射表中，用映射表中与当前PWM占空比最接近的两个PWM占空比的初始扇叶通过频率做差分计算得到与当前PWM占空比对应的参考扇叶通过频率，假设当前PWM占空比为34％，则通过映射表中30％对应的初始扇叶通过频率和40％对应的初始扇叶通过频率进行差分计算得到34％对应的参考扇叶通过频率。
具体的，首先从BMC调取当前PWM占空比数值，在存储单元中的映射表中检索到对应的BPF数值。若当前PWM占空比不在预存参数表中,可差分计算出BPF。比如当前PWM占空比(用W代替)为34,可用取整函数计算其在映射表中最接近的PWM，即第一目标PWM占空比和第二目标PWM占空比，利用第一目标PWM占空比对应的第一初始扇叶通过频率及第二目标PWM占空比对应的第二初始扇叶通过频率作差分计算求出当前PWM占空比对应的BPF。参照图3所示，图3为本申请所提供的一种获取当前PWM占空比对应的参考扇叶通过频率的流程图，假设W＝34，那么通过S201计算得到n为3，通过S202计算得到的delta为4，可以理解的是，如果S203中的delta＝0，说明当前PWM占空比位于映射表中，可直接匹配，即当前PWM占空比对应的BPF＝Matrix_BPF(n，2)，如果S203中的delta≠0，说明当前PWM占空比不在映射表中，通过 S205中的差分计算方案计算当前PWM占空比对应的参考扇叶通过频率。
在本申请一些实施例中，该风扇监控方法还包括：
通过扇叶通过频率对风扇的转速进行调整。
可以理解的是，扇叶通过频率与转速呈正比，扇叶通过频率的偏差就是转速的偏差，因此，本步骤可通过扇叶通过频率来对风扇的转速进行监控和调整，采用这种方案，风扇内不需要在设置转速控制电路，也不需要FG信号端子，硬件架构简单。
综上，采用本申请的方案在服务器主板上集成一颗或多颗麦克风，采集风扇的噪音，并通过噪音的特征分析来判断风扇的健康状况，并在BMC中预存储模型参数和故障标签，对风扇的状态进行在线分析，识别到故障或故障趋势后发出警报，并且可以通过麦克风采集到的信号计算转速，替代风扇PCB板上的读取元件，硬件架构简单不需占用过多系统内资源，同时能定位具体的故障原因，为服务器的维护提供准确建议。
另一方面，请参照图4，图4为本申请所提供的一种风扇监控系统的结构示意图，应用于BMC，该风扇监控系统包括：
获取模块1，用于获取麦克风采集的风扇的噪音信号；
提取模块2，用于基于噪音信号得到信号特征数据及风扇的扇叶通过频率；
诊断模块3，用于将信号特征数据输入预先构建的状态诊断模型，得到诊断数据；
信息生成模块4，用于基于诊断数据和扇叶通过频率生成风扇的状态诊断提示信息。
可见，本实施例所提供的一种风扇监控系统，利用麦克风获取风扇的噪音信号，基于噪音信号提取信号特征数据及风扇的扇叶通过频率，将信号特征数据输入预设诊断模型可得到风扇的诊断数据，根据诊断数据及扇叶通过频率共同确定风扇的状态诊断结果，使得诊断更全面，仅通过麦克风采集到的噪音信号即可完成故障诊断，硬件架构简单，无需占用过多硬件资源。
在本申请一些实施例中，该风扇监控系统还包括：
计算模块，用于计算扇叶通过频率和参考扇叶通过频率的偏差百分比；
基于诊断数据和扇叶通过频率生成风扇的状态诊断提示信息的过程包括：
基于诊断数据和偏差百分比生成风扇的状态诊断提示信息。
在本申请一些实施例中，该风扇监控系统还包括：
预处理模块，用于按预设规则对风扇进行步进扫频，获取风扇在每一PWM占空比下的初始扇叶通过频率；基于步进扫频对应的所有PWM占空比及每一PWM占空比下的初始扇叶通过频 率，构建PWM占空比与扇叶通过频率的映射表；
计算扇叶通过频率和参考扇叶通过频率的偏差百分比的过程包括：
确定扇叶通过频率对应的当前PWM占空比；
基于当前PWM占空比及映射表中的初始扇叶通过频率，确定参考扇叶通过频率；
计算扇叶通过频率和参考扇叶通过频率的偏差百分比。
在本申请一些实施例中，基于当前PWM占空比及映射表中的初始扇叶通过频率，确定参考扇叶通过频率的过程包括：
判断映射表中是否存在扇叶通过频率对应的当前PWM占空比；
若是，将映射表中与当前PWM占空比对应的初始扇叶通过频率作为参考扇叶通过频率；
若否，基于当前PWM占空比确定映射表中的第一目标PWM占空比和第二目标PWM占空比；
根据第一目标PWM占空比对应的初始扇叶通过频率及第二目标PWM占空比对应的初始扇叶通过频率计算当前PWM占空比对应的参考扇叶通过频率；
其中，第一目标PWM占空比和第二目标PWM占空比相邻，第一目标PWM占空比小于当前PWM占空比，第二目标PWM占空比大于当前PWM占空比。
在本申请一些实施例中，根据第一目标PWM占空比对应的初始扇叶通过频率及第二目标PWM占空比对应的初始扇叶通过频率计算当前PWM占空比对应的参考扇叶通过频率的过程包括：
确定第一目标PWM占空比对应的第一初始扇叶通过频率及第二目标PWM占空比对应的第二初始扇叶通过频率；
对第一初始扇叶通过频率和第二初始扇叶通过频率进行差分计算，得到当前PWM占空比对应的参考扇叶通过频率。
在本申请一些实施例中，信号特征数据包括时域特征数据、频域特征数据及时频域特征数据。
在本申请一些实施例中，基于噪音信号得到风扇的扇叶通过频率的过程包括：
对噪音信号进行FFT处理，得到频谱数据；
基于频谱数据计算扇叶通过频率。
在本申请一些实施例中，诊断数据包括健康状态，或，故障状态及故障状态对应的故障原因。
在本申请一些实施例中，故障原因包括叶片偏心、轴承磨损、绕组性能退化、润滑油不足或耗干、IC元件电阻变化中的一项或多项。
在本申请一些实施例中，该风扇监控系统还包括：
控制模块，用于通过扇叶通过频率对风扇的转速进行调整。
在本申请一些实施例中，预先构建状态诊断模型的过程包括：
获取风扇在目标电子设备中的噪音样本，噪音样本包括故障噪音样本及非故障噪音样本，并对故障噪音样本和非故障噪音样本添加各自对应的标签；
提取各个噪音样本中的特征数据，将特征数据组合成矩阵样本，并将矩阵样本划分为第一矩阵样本和第二矩阵样本；
将第一矩阵样本输入分类器进行训练，得到多个模型；
将第二矩阵样本载入多个模型中进行测试，根据测试结果选择最优模型作为状态诊断模型。
另一方面，本申请还提供了一种风扇监控装置，包括：
存储器，用于存储计算机程序；
处理器，用于执行计算机程序时实现如上文任意一个实施例所描述的风扇监控方法的步骤。
具体的，存储器包括非易失性存储介质、内存储器。该非易失性存储介质存储有操作系统和计算机可读指令，该内存储器为非易失性存储介质中的操作系统和计算机可读指令的运行提供环境。处理器为风扇监控装置提供计算和控制能力，执行存储器中保存的计算机程序时，可以实现以下步骤：获取麦克风采集的风扇的噪音信号；基于噪音信号得到信号特征数据及风扇的扇叶通过频率；将信号特征数据输入预先构建的状态诊断模型，得到诊断数据；基于诊断数据和扇叶通过频率生成风扇的状态诊断提示信息。
可见，本实施例所提供的一种风扇监控装置，利用麦克风获取风扇的噪音信号，基于噪音信号提取信号特征数据及风扇的扇叶通过频率，将信号特征数据输入预设诊断模型可得到风扇的诊断数据，根据诊断数据及扇叶通过频率共同确定风扇的状态诊断结果，使得诊断更全面，仅通过麦克风采集到的噪音信号即可完成故障诊断，硬件架构简单，无需占用过多硬件资源。
处理器执行存储器中保存的计算机子程序时，可以实现以下步骤：计算扇叶通过频率和参考扇叶通过频率的偏差百分比；基于诊断数据和偏差百分比生成风扇的状态诊断提示信息。
在本申请一些实施例中，处理器执行存储器中保存的计算机子程序时，可以实现以下步 骤：按预设规则对风扇进行步进扫频，获取风扇在每一PWM占空比下的初始扇叶通过频率；基于步进扫频对应的所有PWM占空比及每一PWM占空比下的初始扇叶通过频率，构建PWM占空比与扇叶通过频率的映射表；确定扇叶通过频率对应的当前PWM占空比；基于当前PWM占空比及映射表中的初始扇叶通过频率，确定参考扇叶通过频率；计算扇叶通过频率和参考扇叶通过频率的偏差百分比。
在本申请一些实施例中，处理器执行存储器中保存的计算机子程序时，可以实现以下步骤：判断映射表中是否存在扇叶通过频率对应的当前PWM占空比；若是，将映射表中与当前PWM占空比对应的初始扇叶通过频率作为参考扇叶通过频率；若否，基于当前PWM占空比确定映射表中的第一目标PWM占空比和第二目标PWM占空比；根据第一目标PWM占空比对应的初始扇叶通过频率及第二目标PWM占空比对应的初始扇叶通过频率计算当前PWM占空比对应的参考扇叶通过频率；其中，第一目标PWM占空比和第二目标PWM占空比相邻，第一目标PWM占空比小于当前PWM占空比，第二目标PWM占空比大于当前PWM占空比。
在本申请一些实施例中，处理器执行存储器中保存的计算机子程序时，可以实现以下步骤：确定第一目标PWM占空比对应的第一初始扇叶通过频率及第二目标PWM占空比对应的第二初始扇叶通过频率；对第一初始扇叶通过频率和第二初始扇叶通过频率进行差分计算，得到当前PWM占空比对应的参考扇叶通过频率。
在本申请一些实施例中，处理器执行存储器中保存的计算机子程序时，可以实现以下步骤：基于噪音信号得到信号特征数据；信号特征数据包括时域特征数据、频域特征数据及时频域特征数据。
在本申请一些实施例中，处理器执行存储器中保存的计算机子程序时，可以实现以下步骤：对噪音信号进行FFT处理，得到频谱数据；基于频谱数据计算扇叶通过频率。
在本申请一些实施例中，处理器执行存储器中保存的计算机子程序时，可以实现以下步骤：将信号特征数据输入预先构建的状态诊断模型，得到诊断数据；诊断数据包括健康状态，或，故障状态及故障状态对应的故障原因。
在本申请一些实施例中，处理器执行存储器中保存的计算机子程序时，可以实现以下步骤：将信号特征数据输入预先构建的状态诊断模型，得到诊断数据；诊断数据包括健康状态，或，故障状态及故障状态对应的故障原因；故障原因包括叶片偏心、轴承磨损、绕组性能退化、润滑油不足或耗干、IC元件电阻变化中的一项或多项。
在本申请一些实施例中，处理器执行存储器中保存的计算机子程序时，可以实现以下步骤：通过扇叶通过频率对风扇的转速进行调整。
在本申请一些实施例中，处理器执行存储器中保存的计算机子程序时，可以实现以下步骤：获取风扇在目标电子设备中的噪音样本，噪音样本包括故障噪音样本及非故障噪音样本，并对故障噪音样本和非故障噪音样本添加各自对应的标签；提取各个噪音样本中的特征数据，将特征数据组合成矩阵样本，并将矩阵样本划分为第一矩阵样本和第二矩阵样本；将第一矩阵样本输入分类器进行训练，得到多个模型；将第二矩阵样本载入多个模型中进行测试，根据测试结果选择最优模型作为状态诊断模型。
在上述实施例的基础上，在本申请一些实施方式中，该风扇监控装置还包括：
输入接口，与处理器相连，用于获取外部导入的计算机程序、参数和指令，经处理器控制保存至存储器中。该输入接口可以与输入装置相连，接收用户手动输入的参数或指令。该输入装置可以是显示屏上覆盖的触摸层，也可以是终端外壳上设置的按键、轨迹球或触控板。
显示单元，与处理器相连，用于显示处理器发送的数据。该显示单元可以为液晶显示屏或者电子墨水显示屏等。
网络端口，与处理器相连，用于与外部各终端设备进行通信连接。该通信连接所采用的通信技术可以为有线通信技术或无线通信技术，如移动高清链接技术(MHL)、通用串行总线(USB)、高清多媒体接口(HDMI)、无线保真技术(WiFi)、蓝牙通信技术、低功耗蓝牙通信技术、基于IEEE802.11s的通信技术等。
另一方面，本申请还提供了一种服务器，包括如上文实施例所描述的风扇监控装置。
对于本申请所提供的一种服务器的介绍请参照上述实施例，本申请在此不再赘述。
本申请所提供的一种服务器具有和上述风扇监控装置相同的有益效果。
另一方面，本申请还提供了一种非易失性可读存储介质，非易失性可读存储介质上存储有计算机程序51，计算机程序51被处理器执行时实现如上文任意一个实施例所描述的风扇监控方法的步骤。
具体的，该非易失性可读存储介质可以包括：U盘、移动硬盘、只读存储器(Read-Only Memory，ROM)、随机存取存储器(Random Access Memory，RAM)、磁碟或者光盘等各种可以存储程序代码的介质。该存储介质上存储有计算机程序，计算机程序被处理器执行时实现以下步骤：获取麦克风采集的风扇的噪音信号；基于噪音信号得到信号特征数据及风扇的扇叶通过频率；将信号特征数据输入预先构建的状态诊断模型，得到诊断数据；基于诊断数据 和扇叶通过频率生成风扇的状态诊断提示信息。
可见，本实施例利用麦克风获取风扇的噪音信号，基于噪音信号提取信号特征数据及风扇的扇叶通过频率，将信号特征数据输入预设诊断模型可得到风扇的诊断数据，根据诊断数据及扇叶通过频率共同确定风扇的状态诊断结果，使得诊断更全面，仅通过麦克风采集到的噪音信号即可完成故障诊断，硬件架构简单，无需占用过多硬件资源。
在本申请一些实施例中，非易失性可读存储介质中存储的计算机子程序被处理器执行时，具体可以实现以下步骤：计算扇叶通过频率和参考扇叶通过频率的偏差百分比；基于诊断数据和偏差百分比生成风扇的状态诊断提示信息。
在本申请一些实施例中，非易失性可读存储介质中存储的计算机子程序被处理器执行时，具体可以实现以下步骤：按预设规则对风扇进行步进扫频，获取风扇在每一PWM占空比下的初始扇叶通过频率；基于步进扫频对应的所有PWM占空比及每一PWM占空比下的初始扇叶通过频率，构建PWM占空比与扇叶通过频率的映射表；确定扇叶通过频率对应的当前PWM占空比；基于当前PWM占空比及映射表中的初始扇叶通过频率，确定参考扇叶通过频率；计算扇叶通过频率和参考扇叶通过频率的偏差百分比。
在本申请一些实施例中，非易失性可读存储介质中存储的计算机子程序被处理器执行时，具体可以实现以下步骤：判断映射表中是否存在扇叶通过频率对应的当前PWM占空比；若是，将映射表中与当前PWM占空比对应的初始扇叶通过频率作为参考扇叶通过频率；若否，基于当前PWM占空比确定映射表中的第一目标PWM占空比和第二目标PWM占空比；根据第一目标PWM占空比对应的初始扇叶通过频率及第二目标PWM占空比对应的初始扇叶通过频率计算当前PWM占空比对应的参考扇叶通过频率；其中，第一目标PWM占空比和第二目标PWM占空比相邻，第一目标PWM占空比小于当前PWM占空比，第二目标PWM占空比大于当前PWM占空比。
在本申请一些实施例中，非易失性可读存储介质中存储的计算机子程序被处理器执行时，具体可以实现以下步骤：确定第一目标PWM占空比对应的第一初始扇叶通过频率及第二目标PWM占空比对应的第二初始扇叶通过频率；对第一初始扇叶通过频率和第二初始扇叶通过频率进行差分计算，得到当前PWM占空比对应的参考扇叶通过频率。
在本申请一些实施例中，非易失性可读存储介质中存储的计算机子程序被处理器执行时，具体可以实现以下步骤：基于噪音信号得到信号特征数据；信号特征数据包括时域特征数据、频域特征数据及时频域特征数据。
在本申请一些实施例中，非易失性可读存储介质中存储的计算机子程序被处理器执行时，具体可以实现以下步骤：对噪音信号进行FFT处理，得到频谱数据；基于频谱数据计算 扇叶通过频率。
在本申请一些实施例中，非易失性可读存储介质中存储的计算机子程序被处理器执行时，具体可以实现以下步骤：将信号特征数据输入预先构建的状态诊断模型，得到诊断数据；诊断数据包括健康状态，或，故障状态及故障状态对应的故障原因。
在本申请一些实施例中，非易失性可读存储介质中存储的计算机子程序被处理器执行时，具体可以实现以下步骤：将信号特征数据输入预先构建的状态诊断模型，得到诊断数据；诊断数据包括健康状态，或，故障状态及故障状态对应的故障原因；故障原因包括叶片偏心、轴承磨损、绕组性能退化、润滑油不足或耗干、IC元件电阻变化中的一项或多项。
在本申请一些实施例中，非易失性可读存储介质中存储的计算机子程序被处理器执行时，具体可以实现以下步骤：通过扇叶通过频率对风扇的转速进行调整。
在本申请一些实施例中，非易失性可读存储介质中存储的计算机子程序被处理器执行时，具体可以实现以下步骤：获取风扇在目标电子设备中的噪音样本，噪音样本包括故障噪音样本及非故障噪音样本，并对故障噪音样本和非故障噪音样本添加各自对应的标签；提取各个噪音样本中的特征数据，将特征数据组合成矩阵样本，并将矩阵样本划分为第一矩阵样本和第二矩阵样本；将第一矩阵样本输入分类器进行训练，得到多个模型；将第二矩阵样本载入多个模型中进行测试，根据测试结果选择最优模型作为状态诊断模型。
还需要说明的是，在本说明书中，诸如第一和第二等之类的关系术语仅仅用来将一个实体或者操作与另一个实体或操作区分开来，而不一定要求或者暗示这些实体或操作之间存在任何这种实际的关系或者顺序。而且，术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含，从而使得包括一系列要素的过程、方法、物品或者设备不仅包括那些要素，而且还包括没有明确列出的其他要素，或者是还包括为这种过程、方法、物品或者设备所固有的要素。在没有更多限制的状况下，由语句“包括一个……”限定的要素，并不排除在包括要素的过程、方法、物品或者设备中还存在另外的相同要素。
对所公开的实施例的上述说明，使本领域专业技术人员能够实现或使用本申请。对这些实施例的多种修改对本领域的专业技术人员来说将是显而易见的，本文中所定义的一般原理可以在不脱离本申请的精神或范围的情况下，在其他实施例中实现。因此，本申请将不会被限制于本文所示的这些实施例，而是要符合与本文所公开的原理和新颖特点相一致的最宽的范围。

Claims (20)

  1. 一种风扇监控方法，其特征在于，应用于BMC，该风扇监控方法包括：
    获取麦克风采集的风扇的噪音信号；
    基于所述噪音信号得到信号特征数据及所述风扇的扇叶通过频率；
    将所述信号特征数据输入预先构建的状态诊断模型，得到诊断数据；
    基于所述诊断数据和所述扇叶通过频率生成所述风扇的状态诊断提示信息。
  2. 根据权利要求1所述的风扇监控方法，其特征在于，该风扇监控方法还包括：
    计算所述扇叶通过频率和参考扇叶通过频率的偏差百分比；
    基于所述诊断数据和所述扇叶通过频率生成所述风扇的状态诊断提示信息的过程包括：
    基于所述诊断数据和所述偏差百分比生成所述风扇的状态诊断提示信息。
  3. 根据权利要求2所述的风扇监控方法，其特征在于，该风扇监控方法还包括：
    按预设规则对所述风扇进行步进扫频，获取所述风扇在每一PWM占空比下的初始扇叶通过频率；
    基于所述步进扫频对应的所有所述PWM占空比及每一所述PWM占空比下的所述初始扇叶通过频率，构建PWM占空比与扇叶通过频率的映射表；
    计算所述扇叶通过频率和参考扇叶通过频率的偏差百分比的过程包括：
    确定所述扇叶通过频率对应的当前PWM占空比；
    基于当前PWM占空比及所述映射表中的所述初始扇叶通过频率，确定参考扇叶通过频率；
    计算所述扇叶通过频率和所述参考扇叶通过频率的偏差百分比。
  4. 根据权利要求3所述的风扇监控方法，其特征在于，所述基于当前PWM占空比及所述映射表中的所述初始扇叶通过频率，确定参考扇叶通过频率的过程包括：
    判断所述映射表中是否存在所述扇叶通过频率对应的当前PWM占空比；
    若是，将所述映射表中与当前PWM占空比对应的初始扇叶通过频率作为参考扇叶通过频率；
    若否，基于当前PWM占空比确定所述映射表中的第一目标PWM占空比和第二目标PWM占空比；
    根据所述第一目标PWM占空比对应的初始扇叶通过频率及所述第二目标PWM占空比对应的初始扇叶通过频率计算当前PWM占空比对应的参考扇叶通过频率；
    其中，所述第一目标PWM占空比和所述第二目标PWM占空比相邻，所述第一目标PWM占 空比小于当前PWM占空比，所述第二目标PWM占空比大于当前PWM占空比。
  5. 根据权利要求4所述的风扇监控方法，其特征在于，根据所述第一目标PWM占空比对应的初始扇叶通过频率及所述第二目标PWM占空比对应的初始扇叶通过频率计算当前PWM占空比对应的参考扇叶通过频率的过程包括：
    确定所述第一目标PWM占空比对应的第一初始扇叶通过频率及所述第二目标PWM占空比对应的第二初始扇叶通过频率；
    对所述第一初始扇叶通过频率和所述第二初始扇叶通过频率进行差分计算，得到当前PWM占空比对应的参考扇叶通过频率。
  6. 根据权利要求1所述的风扇监控方法，其特征在于，所述信号特征数据包括时域特征数据、频域特征数据及时频域特征数据。
  7. 根据权利要求1所述的风扇监控方法，其特征在于，基于所述噪音信号得到所述风扇的扇叶通过频率的过程包括：
    对所述噪音信号进行FFT处理，得到频谱数据；
    基于所述频谱数据计算扇叶通过频率。
  8. 根据权利要求7所述的风扇监控方法，其特征在于，所述对所述噪音信号进行FFT处理，得到频谱数据，包括：
    对所述噪音信号进行FFT处理，以将所述噪音信号由时域转换为频域，得到频谱数据。
  9. 根据权利要求1所述的风扇监控方法，其特征在于，所述诊断数据包括健康状态，或，故障状态及所述故障状态对应的故障原因。
  10. 根据权利要求9所述的风扇监控方法，其特征在于，所述故障原因包括叶片偏心、轴承磨损、绕组性能退化、润滑油不足或耗干、IC元件电阻变化中的一项或多项。
  11. 根据权利要求1所述的风扇监控方法，其特征在于，该风扇监控方法还包括：
    通过所述扇叶通过频率对所述风扇的转速进行调整。
  12. 根据权利要求11所述的风扇监控方法，其特征在于，所述扇叶通过频率与所述风扇的转速呈正比。
  13. 根据权利要求1-12任意一项所述的风扇监控方法，其特征在于，预先构建状态诊断模型的过程包括：
    获取所述风扇在目标电子设备中的噪音样本，所述噪音样本包括故障噪音样本及非故障噪音样本，并对所述故障噪音样本和所述非故障噪音样本添加各自对应的标签；
    提取各个所述噪音样本中的特征数据，将所述特征数据组合成矩阵样本，并将所述矩阵样本划分为第一矩阵样本和第二矩阵样本；
    将所述第一矩阵样本输入分类器进行训练，得到多个模型；
    将所述第二矩阵样本载入多个所述模型中进行测试，根据测试结果选择最优模型作为所述状态诊断模型。
  14. 根据权利要求1-12任意一项所述的风扇监控方法，其特征在于，所述对所述故障噪音样本和所述非故障噪音样本添加各自对应的标签，包括：
    将所述非故障噪音样本的标签设置为非故障；
    将所述故障噪音样本的标签设置为故障原因。
  15. 根据权利要求1所述的风扇监控方法，其特征在于，还包括：
    在服务器产品完成组装后、出厂前对所述麦克风的灵敏度进行校验；
    在所述麦克风灵敏度校验通过后，执行所述获取麦克风采集的风扇的噪音信号。
  16. 根据权利要求1或15所述的风扇监控方法，其特征在于，所述麦克风设置在服务器主板上靠近风扇的一侧。
  17. 一种风扇监控系统，其特征在于，应用于BMC，该风扇监控系统包括：
    获取模块，用于获取麦克风采集的风扇的噪音信号；
    提取模块，用于基于所述噪音信号得到信号特征数据及所述风扇的扇叶通过频率；
    诊断模块，用于将所述信号特征数据输入预先构建的状态诊断模型，得到诊断数据；
    信息生成模块，用于基于所述诊断数据和所述扇叶通过频率生成所述风扇的状态诊断提示信息。
  18. 一种风扇监控装置，其特征在于，包括：
    存储器，用于存储计算机程序；
    处理器，用于执行所述计算机程序时实现如权利要求1-16任意一项所述的风扇监控方法的步骤。
  19. 一种服务器，其特征在于，包括如权利要求18所述的风扇监控装置。
  20. 一种非易失性可读存储介质，其特征在于，所述非易失性可读存储介质上存储有计算机程序，所述计算机程序被处理器执行时实现如权利要求1-16任意一项所述的风扇监控方法的步骤。
PCT/CN2023/111986 2022-08-09 2023-08-09 一种风扇监控方法、系统、装置、服务器及可读存储介质 WO2024032655A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202210947593.7 2022-08-09
CN202210947593.7A CN115033461B (zh) 2022-08-09 2022-08-09 一种风扇监控方法、系统、装置、服务器及可读存储介质

Publications (1)

Publication Number Publication Date
WO2024032655A1 true WO2024032655A1 (zh) 2024-02-15

Family

ID=83130359

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/111986 WO2024032655A1 (zh) 2022-08-09 2023-08-09 一种风扇监控方法、系统、装置、服务器及可读存储介质

Country Status (2)

Country Link
CN (1) CN115033461B (zh)
WO (1) WO2024032655A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN118030588A (zh) * 2024-04-10 2024-05-14 永联科技(常熟)有限公司 风扇控制方法、装置、电子设备及存储介质

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115033461B (zh) * 2022-08-09 2023-04-18 苏州浪潮智能科技有限公司 一种风扇监控方法、系统、装置、服务器及可读存储介质

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102521088A (zh) * 2011-11-15 2012-06-27 浪潮电子信息产业股份有限公司 一种基于声学探测的服务器风扇状态检测方法
CN103423186A (zh) * 2012-05-21 2013-12-04 联想(北京)有限公司 一种检测风扇故障的方法及一种电子设备
CN105136435A (zh) * 2015-07-15 2015-12-09 北京汉能华科技股份有限公司 一种风力发电机组叶片故障诊断的方法和装置
JP2018025841A (ja) * 2016-08-08 2018-02-15 株式会社明電舎 ファン監視システム，監視方法，監視プログラム
CN113886156A (zh) * 2021-09-09 2022-01-04 苏州浪潮智能科技有限公司 一种风扇噪声对硬盘的干扰的测试方法、系统及相关组件
CN115033461A (zh) * 2022-08-09 2022-09-09 苏州浪潮智能科技有限公司 一种风扇监控方法、系统、装置、服务器及可读存储介质

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101105501A (zh) * 2006-07-14 2008-01-16 鸿富锦精密工业（深圳）有限公司 风扇转速测试系统及方法
CN110985425A (zh) * 2019-11-29 2020-04-10 联想(北京)有限公司 一种信息检测方法、电子设备及计算机可读存储介质
CN114153693B (zh) * 2021-11-26 2023-11-14 苏州浪潮智能科技有限公司 一种服务器风扇状态监控方法、装置及存储介质

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102521088A (zh) * 2011-11-15 2012-06-27 浪潮电子信息产业股份有限公司 一种基于声学探测的服务器风扇状态检测方法
CN103423186A (zh) * 2012-05-21 2013-12-04 联想(北京)有限公司 一种检测风扇故障的方法及一种电子设备
CN105136435A (zh) * 2015-07-15 2015-12-09 北京汉能华科技股份有限公司 一种风力发电机组叶片故障诊断的方法和装置
JP2018025841A (ja) * 2016-08-08 2018-02-15 株式会社明電舎 ファン監視システム，監視方法，監視プログラム
CN113886156A (zh) * 2021-09-09 2022-01-04 苏州浪潮智能科技有限公司 一种风扇噪声对硬盘的干扰的测试方法、系统及相关组件
CN115033461A (zh) * 2022-08-09 2022-09-09 苏州浪潮智能科技有限公司 一种风扇监控方法、系统、装置、服务器及可读存储介质

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN118030588A (zh) * 2024-04-10 2024-05-14 永联科技(常熟)有限公司 风扇控制方法、装置、电子设备及存储介质
CN118030588B (zh) * 2024-04-10 2024-06-11 永联科技(常熟)有限公司 风扇控制方法、装置、电子设备及存储介质

Also Published As

Publication number Publication date
CN115033461B (zh) 2023-04-18
CN115033461A (zh) 2022-09-09

Similar Documents

Publication Publication Date Title
WO2024032655A1 (zh) 一种风扇监控方法、系统、装置、服务器及可读存储介质
US10935458B2 (en) Facility condition monitoring device and facility condition monitoring method
Preece et al. Assessing the applicability of uncertainty importance measures for power system studies
Peng et al. Regression modeling of semicompeting risks data
US20050228875A1 (en) System for estimating processing requirements
CN108173670A (zh) 检测网络的方法和装置
CN112882796A (zh) 异常根因分析方法和装置，及存储介质
JP2002342128A (ja) ホストマシンからサービスの健全性を抽出する方法
CN109783324B (zh) 系统运行预警方法及装置
US20200249281A1 (en) Information processing apparatus and computer-readable recording medium storing battery deterioration diagnosis program
KR101953558B1 (ko) 스마트 기기 결함 관리 장치 및 방법
JP2019095822A (ja) パラメータ設定方法、データ分析装置、データ分析システム及びプログラム
CN112285586B (zh) Bms测试方法、装置、系统、仿真测试设备及存储介质
Ardito et al. Creating and evaluating a software power model for linux single board computers
CN117291690B (zh) 一种基于数据分析的产品销售智能推荐方法
CN117115937B (zh) 设备运行状态监测方法、装置、云端设备及存储介质
CN113688564A (zh) 一种预测ssd硬盘剩余寿命的方法、装置、终端及存储介质
Hoque et al. Full charge capacity and charging diagnosis of smartphone batteries
CN102650263B (zh) 对风力涡轮机设备进行测试的方法和装置
CN101957436B (zh) 一种it服务器类设备的电源质量测试仪和测试方法
Liu et al. Research on online moisture detector in grain drying process based on V/F conversion
CN105093157B (zh) 基于iec61850采样值的参数算法的测试方法
CN113608953B (zh) 测试数据生成方法、装置、电子设备及可读存储介质
CN110632519A (zh) 燃料电池故障诊断方法和装置
CN109800114B (zh) 一种bmc可视化测试方法、装置、终端及存储介质

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 23851876

Country of ref document: EP

Kind code of ref document: A1