Automatic meteorological station five-element signal comprehensive fault detector and system

Abstract

The invention relates to the technical field of weather station detectors, in particular to an automatic weather station five-element signal comprehensive fault detector and system, which comprises a power supply box, a control circuit board and a voltmeter, and is characterized in that the voltmeter and the power supply box are electrically coupled with the control circuit board; the control circuit board is electrically coupled with a rainfall sensor, a wind speed sensor, a wind direction sensor, a temperature sensor and a humidity sensor; in the invention, the requirements of field maintenance work are combined, the functions of main maintenance tools are replaced, the main maintenance tools comprise a mechanical multimeter, an electric multimeter, a rainfall counter, a maintenance power supply, an oscilloscope, a frequency meter and the like, the maintenance operation and fault judgment are very simple and direct, independent pulse signal fault maintenance can be carried out on each sensor separated from a collector, and the maintenance efficiency is greatly improved. And pulse signal data of each element can be measured on line, and a fault part can be well determined by comparing according to a detection result, so that the maintenance is more convenient, faster and more accurate, the consumed time is less, and the efficiency is high.

Description

Technical Field

[0001] The invention relates to the technical field of weather station fault detectors, in particular to an automatic weather station five-element signal comprehensive fault detector and system. Background Art

[0002] Automatic weather stations are important infrastructure for meteorological observation. The accuracy and reliability of the five elements of rainfall, temperature, humidity, wind speed and wind direction measured by them are crucial to many fields such as meteorological research, weather forecasting, military, agricultural production, aerospace, etc. However, in the long-term operation of existing automatic weather stations, due to the influence of various natural environmental factors (such as high temperature, high humidity, strong wind, heavy rain, sandstorm, lightning, etc.) and other problems, automatic weather stations often fail.

[0003] The application of five-element automatic weather stations in meteorological departments accounts for 90%, and the number reaches tens of thousands. With the continuous upgrading and updating of automatic stations, their guarantee, maintenance, and repair means and professional tools have not been well valued and developed, and there have been many shortcomings and deficiencies for a long time. In addition to multimeters, there are almost no other maintenance auxiliary and fault detection equipment tools available for field sites. This has greatly restricted the efficiency of on-site support of meteorological equipment, and the meteorological department has extremely strict requirements on the timeliness of equipment support. Our main and commonly used maintenance tools are digital multimeters, mechanical multimeters, rain counters, maintenance power supplies, frequency meters, oscilloscopes, etc. These tools are very inconvenient to use and carry during field maintenance, and the operation is complicated, and the application conditions are completely unavailable; and meteorological observation equipment is distributed in different positions of various heights and distances. Therefore, traditional maintenance tools and instruments have problems such as carrying and operation that do not meet the requirements of field equipment maintenance, and maintenance has low efficiency and high work intensity. The fault detector is easy to operate, fast, accurate, takes less time, and has high efficiency, filling the gap in the application of meteorological ground detection equipment support and maintenance and repair technology in the meteorological department. Summary of the invention

[0004] Traditional meteorological observation equipment is distributed in different locations of high and low, far and near, so traditional segmented maintenance has technical problems of low efficiency and high physical exertion of personnel. The present invention provides an automatic weather station five-element signal comprehensive fault detector and system.

[0005] The technical solution adopted by the present invention is: an automatic weather station five-element signal comprehensive fault detector, including a power box, a control circuit board and a voltmeter, the voltmeter and the power box are electrically coupled to the control circuit board, and the control circuit board is electrically coupled to a rainfall sensor, a wind speed sensor, a wind direction sensor, a temperature sensor and a humidity sensor.

[0006] In one embodiment, it includes a signal acquisition unit, a data preprocessing and feature extraction unit.

[0007] In one of the embodiments, the signal acquisition unit is a sensor group, and the sensor group includes a rainfall sensor, a wind speed sensor, a wind direction sensor, a temperature sensor and a humidity sensor.

[0008] In one embodiment, the output resistance value RT of the temperature sensor and the temperature T generally satisfy the following functional relationship: RT=RO(1+αT) (RO is the reference resistor and α is the temperature coefficient); a voltage divider circuit is formed by a precision resistor and a temperature sensor; Then, the resistance change is converted into a voltage change VT, which is amplified by an amplifier with an amplification factor A to obtain an amplified voltage signal VOUT; Where, VOUT = A×VT. The analog voltage signal is input into the analog-to-digital converter (ADC) and converted into a digital signal. According to the Nyquist sampling theorem, the sampling frequency fs of the ADC should be at least twice the highest frequency fmax of the signal, that is, fs ≥ 2fmax. For signals with relatively slow changes such as temperature and humidity, fs can be set to 1Hz. For signals with fast changes such as wind speed and wind direction, fs can be increased to above 10Hz; Among them, a low-pass filter, a high-pass filter and a band-pass filter are added to the signal conditioning circuit.

[0009] In one embodiment, a clock calibration unit is further included, wherein the clock calibration unit uses a clock chip as a clock source of the system and performs clock calibration through GPS or Beidou satellite signals, and during the signal acquisition process, the acquisition time of each sensor is triggered by a unified clock signal, and the measurement data of different elements have a strict corresponding relationship in time; Assume that the collection time is tn, for the nth data Tn collected by the temperature sensor, the nth data Hn collected by the humidity sensor, the nth data Pn collected by the air pressure sensor, the nth data Wn collected by the wind speed sensor and the nth data Dn collected by the wind direction sensor, the collection time error △t satisfies |△t|≤1μs.

[0010] In one embodiment, the data preprocessing and feature extraction unit includes the following: Data cleaning and outlier removal: The sliding window median filter algorithm is used to clean the collected data to remove noise and outliers. Assume that the sliding window size is m. For the measurement value Xi at a certain moment, the output value yi after filtering is the median of the data in the window. That is, yi=median{Xi-(m-1)÷2, ..., Xi, ..., Xi+(m-1)÷2}; At the same time, the 3σ criterion in statistical methods is combined to further identify and eliminate outliers; for a data sequence of a certain meteorological element {X1, X2, ..., Xn}, its mean X and standard deviation σ are first calculated. If a certain measurement value Xj satisfies |Xj -X|>3σ, it is judged as an outlier and eliminated; Feature extraction and parameter calculation: The key characteristic parameters that can reflect the change characteristics of meteorological elements and signal quality are extracted from the preprocessed data. For temperature data, the daily change amplitude △T24h within 24 hours is calculated. The formula is △T24h=Tmax-Tmin (where Tmax and Tmin are the highest and lowest temperatures within 24 hours, respectively), and the temperature change rate rT, rT=△T÷△t (△T is the temperature change within a certain time interval △t); For rainfall data, the core component reed switch is used for cumulative calculation, with the unit of measurement being 1 mm. For humidity data, calculate the daily average relative humidity H and the standard deviation of humidity variation σH; For wind speed data, calculate the gust coefficient Gw of wind speed, Gw=Wgust÷Wavg (Wgust is the wind speed, Wavg is the average wind speed) and the spectrum characteristics of wind speed. Obtain the spectrum distribution of wind speed signal through fast Fourier transform (FFT), analyze the frequency range where its energy is concentrated, and determine whether there is abnormal frequency component. For wind direction data, the distribution probability of wind direction in each direction P(Di), i=1, 2, ..., 16 (wind direction is divided into 16 directions), and the frequency of wind direction change fD are obtained by calculating the number of times the wind direction changes in unit time.

[0011] The beneficial effects of the present invention are: 1. Compared with the prior art, the present invention combines the requirements of field maintenance work and replaces the functions of the main maintenance tools, including an electromechanical multimeter, a rain gauge counter, a maintenance power supply, an oscilloscope, a frequency meter, etc. The maintenance operation and fault judgment are also extremely simple and direct. Each sensor can be separated from the collector for separate pulse signal fault maintenance, and the pulse signal data of each element can also be measured online. According to the test results, comparison can be made to better and more clearly determine the fault location, making maintenance more convenient and not consuming too much physical strength.

[0012] 2. Compared with the prior art, in the present invention, through automated signal acquisition, real-time data processing and intelligent fault diagnosis technology, the system can quickly detect and diagnose faults in the five-element signals of the automatic weather station, and the fault detection time is shortened from several hours or even days to within a few minutes, which greatly improves the efficiency of troubleshooting, reduces the time of missing meteorological data, ensures the continuity, timeliness and availability of meteorological data, and provides a strong guarantee for the normal development of meteorological services. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 It is a schematic diagram of the structure of the present invention; Figure 2 The circuit principle of the present invention is Figure 1 .

[0014] Figure 3 The circuit principle of the present invention is Figure 2 .

[0015] The following are marked in the figure: 1. Power box; 2. Control circuit board; 3. Voltmeter; 4. LED light; 5. Collector; 6. Sensor group. DETAILED DESCRIPTION

[0016] In the description of the present invention, it should be noted that the orientations or positional relationships indicated by terms such as “front”, “up”, “down”, “left”, “right”, “vertical” and “horizontal” are based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention.

[0017] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a direct connection, or an indirect connection through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0018] The following is combined with Figure 1-3 The present invention is further described. Example

[0019] In order to solve the problems existing in the background technology, the present application proposes the following technical solutions: an automatic weather station five-element signal comprehensive fault detector, comprising a power box 1, a control circuit board 2 and a voltmeter 3, the voltmeter 3 and the power box 1 are electrically coupled to the control circuit board 2, the control circuit board 2 is electrically coupled with a rainfall sensor, a wind speed sensor, a wind direction sensor, a temperature sensor and a humidity sensor, and an LED light 4 is connected to the outside of the power box 1.

[0020] The maintenance instrument in this embodiment combines the requirements of field maintenance work and better replaces the functions of the main maintenance tools, including an electromechanical multimeter, a rain counter, a maintenance power supply, an oscilloscope, a frequency meter, etc., and provides a corresponding maintenance power supply for each element. At the same time, the digital tube display replaces the function of the oscilloscope frequency meter counter. The maintenance operation and fault judgment are also extremely simple and direct. Each sensor can be separated from the collector for separate pulse signal fault maintenance, and the pulse signal data of each element can also be measured online. According to the test results, the comparison can better and clearly determine the fault location. Among them, for the detection of the two elements of wind direction and wind speed, the detector greatly reduces the difficulty and intensity of the maintenance work; The actual usage of the function selection switch is as follows: 1: Twist the shape selection to the first gear for digital voltmeter 3, which is used to measure the voltage value of the signal port and the power supply port. Rotate to the corresponding gear to perform the corresponding measurement.

[0021] There are many types of signal port voltages, the normal voltage is usually: 1V, 5V, a few are 12V. The power supply voltage is usually 5V, 12V. (For details, please refer to the user manual of the relevant equipment).

[0022] This detector applies quantitative measurement. The numbers displayed on the voltmeter 3 can be used to determine whether the sensor has the power supply conditions required for normal operation and whether the normal voltage of the signal port is within the range specified by the product.

[0023] 2: Turn clockwise to the second gear to get electromagnetic voltmeter 3, which is used to measure the voltage value of signal ports such as temperature and humidity. The quantitative measurement method is used to measure temperature, the humidity signal output voltage is 0~5V, and the temperature signal output voltage is 0~1V.

[0024] For simple test, you can rub the temperature sensor probe with your hands and observe the pointer of electromagnetic voltmeter 3 slowly changing and rising. When measuring humidity, blow on the humidity sensor probe, and the humidity signal output voltage slowly rises. At the same time, in order to judge and distinguish the collector failure and sensor failure, the detector can perform a separate performance test on the temperature and humidity sensor (the method is the same as above), so that the collector failure and sensor failure can be well judged and distinguished.

[0025] 3: Turn clockwise to the third gear to measure the pulse signal, which is used to measure pulse signals such as wind speed and rainfall to determine whether there is a fault in the signal circuit.

[0026] The pulse signal voltage of the wind speed signal output varies between 0V and 5V. By counting the pulse signal in combination with the digital tube counter, it is possible to determine whether the wind speed sensor is faulty. The blue LED4 turns off once after every 500 pulses, and then turns on once after another 500 pulses. This is more intuitive when measuring the height pulse signal (at high wind speed). This detector can be separated from the collector and the sensor can be directly tested. The fault judgment method is the same as above. Through the above detection process, the fault of the sensor and the fault of the collector wind speed signal circuit can be clearly judged.

[0027] 4: When measuring the quality of the reed switch alone, you need to press the bias power switch to provide working voltage for the reed switch test. At this time, the counter counts the on and off of the reed switch. When measuring other pulse signal ports, the switch should be in the off state.

[0028] 5: Press the counter reset button to clear the numbers on the counter digital tube at any time during the measurement process.

[0029] 2. Auxiliary circuit function 1: This instrument uses two 18650 batteries (7.4V internally stabilized to 5V) as the working power source, and has its own USB lithium battery charging circuit. To a certain extent, it can be used as an emergency 5V power supply.

[0030] 2: This instrument is equipped with a small LED auxiliary lighting lamp, which can provide local lighting during on-site maintenance.

[0031] Among them, meteorological ground observation equipment is widely distributed in rural areas at different altitudes and longitudes. For a long time, the meteorological department has lacked professional field maintenance tools corresponding to automatic weather stations. The carrying and use environment of maintenance tools in laboratories and maintenance rooms and the complexity of operation cannot meet the requirements of field maintenance. There have long been problems such as poor timeliness, low maintenance efficiency, and high work intensity. The maintenance instrument is highly integrated, comprehensive and targeted, and easy to carry and operate. It well guarantees the timeliness requirements of maintenance of meteorological ground observation equipment and improves the overall maintenance efficiency.

[0032] 2: Solve the diversity maintenance problem of collector output signals.

[0033] Facing the diversity of collector output signals, for example, the sampling outputs of temperature, humidity, and wind direction are 1V to 5V slowly changing voltage signals. Rainfall and wind speed are TTL pulse signals with changing rates. Other fixed rate signal acquisition devices are also widely used (such as RS232TTL pulse signals, RS485 differential bus signals, CAN differential bus signals). 1: For RS232TTL pulse signals, RS485 differential bus signals, and CAN differential bus signals, the manufacturer has set up working status indicator lights in the collector. Pay attention to this part when repairing the equipment.

[0034] 2: Position 3 is mainly for the measurement of TTL pulse signals. For rainfall and wind speed, which are TTL pulse signals, asynchronous counting method is used to count pulses because the signal is rate-varying. The working principle is: when there is a pulse signal in such a line, the counter counts the number of pulses, and the number that changes in the digital tube is confirmed as a TTL pulse signal. The counter stops counting when there is no pulse signal in the line. While measuring the TTL pulse signal, measure the voltage value of the signal, and combine the two measurement results to comprehensively judge whether the signal channel is normal (for example: 1 port voltage is normal and there is no pulse signal, the sensor has no pulse output, and it can be judged as a fault at this time. 2 There is a pulse signal port voltage is low, which may be due to internal sensor problems, insufficient line power supply, or aging of the line and poor contact) 3: For analog voltage signals such as temperature and humidity, the characteristics are: 1 Slowly changing voltage signal, 2 The value range of electromagnetic voltmeter 3 is usually between 1 volt and 5 volts. Since the digital multimeter currently used adopts the integral sampling method, the display result is the average voltage of a certain period of time, and the display result is not real-time, and it is not possible to observe the linear change voltage output by sensors such as temperature and humidity. Therefore, in view of this problem, the traditional electromagnetic voltmeter 3 can better meet the measurement requirements of linear voltage. For the detection of temperature and humidity, its normal measurement voltage change value is between zero volt and one volt. If the measurement shows negative pressure, zero volt or exceeds this measurement voltage, it is determined that the sensor is faulty and needs to be replaced.

[0035] 4: Wind direction measurement is divided into two output states, among which: 1 When the output is Gray code, the TTL pulse signal is measured, and the number of the digital tube changes to confirm that it is a TTL pulse signal. At the same time, the detector uses a combination of qualitative measurement and quantitative measurement. The Edl lamp, which is lit, represents the output of the wind speed signal. At the same time, combined with the speed of change of the digital tube number, it can be judged whether the performance of the sensor is intact. 2: When the signal output is 0 to 5 volt analog voltage, an electromagnetic voltmeter 3 voltage measurement is used.

[0036] 5: Digital voltmeter 3 is mainly used to measure universal voltages, such as 5V, 12V, and 24V power supply circuits.

[0037] Note: Wind direction and speed, temperature, humidity and rainfall sensors can be separated from the collector for individual performance tests, so that it is easy to distinguish between collector failure and sensor failure.

[0038] refer to Figure 2 , to explain the circuit schematic enter: J1 and J2 are the red and black test pen insertion holes, respectively, to complete the input of the signal to be tested. SW is a double-pole three-gear input signal selection switch. One set of SW1-2 switches, gear 1, is mainly used to measure DC power supply voltage, gear 2 is mainly used to measure weak voltage and slowly changing DC voltage, and gear 3 is mainly used to measure TTL pulse signals. The other set of switches SW1-1, gear 1 and gear 2 are vacant gears. Gear 3 connects the digital voltmeter 3 to the circuit when measuring the pulse signal, and completes the measurement of the signal voltage while measuring the pulse signal.

[0039] Functional description of the core part of the circuit: U1, U2, and U3 are CD4026 decimal counting and 7-segment LED digital tube driving logic chips. Three of these chips form a three-digit counting circuit. The maximum display number of this group of LED digital tubes is 999, which is enough to meet the needs of observing pulse signals. When the number of pulses reaches 500 times, LED4 will cycle on and off at this frequency, which still has a prompt display function for high-speed pulse signals.

[0040] U4 is a 74HC14 six-NOT gate logic chip, which is mainly used to isolate and shape the input pulse signal, isolate the impact on the measured signal during online measurement, and shape it to provide a clean counting square wave (CLK) for the subsequent CD4026 decimal counting.

[0041] When measuring the quality of the reed switch alone, pressing the self-locking switch K1 can provide the reed switch with a bias voltage of about 3.5V for measurement. When measuring the rainfall and wind speed pulse signals online, K1 should be in the disconnected state.

[0042] K2 is the manual reset button for the counter, which can reset the counter at any time; It should also be noted that the five sensors are: rainfall sensor, wind speed sensor, wind direction sensor, temperature sensor and humidity sensor.

[0043] The rainfall signal is provided by the rainfall signal R and G ports located below the collector. The signal is shaped into square waves by the internal integration circuit of the collector 5 and removes some interference. Its function is to allow the counter to count normally and reflect the quality of the rainfall signal. It is used for qualitative analysis and judgment of rainfall signal faults. The fault is the pulse signal generated by the on-off of the rain sensor reed switch. The detection method is online measurement or disconnecting the signal input line to measure the presence or absence of pulses in the signal line separately.

[0044] The temperature signal is provided by the temperature signal 4-wire port at the bottom of the collector. The signal is a DC voltage difference signal generated by a temperature-sensitive platinum resistor as the temperature changes. The detection method is online measurement or disconnecting the signal input line to directly measure the voltage in the signal line.

[0045] The wind speed signal of the collector is provided by the R and G ports of the wind speed signal of the collector. This signal is a pulse signal generated by the on-off of the grating switch inside the wind speed sensor. The detection method can be online measurement or disconnecting the signal input line to directly measure the presence of pulses in the signal line. The wind direction signal is provided by the wind direction signal A0~A8 ports of the collector. This signal is an 8-channel pulse signal generated by the Gray code encoder switch inside the sensor. The detection method is to measure the presence of pulses at each port online.

[0046] The humidity signal is provided by the humidity signal 5V, H1, G port of the collector. The signal is converted from the analog signal inside the humidity sensor to a pulse signal. The detection method can be online measurement or disconnecting the signal input line to directly measure the presence or absence of pulses in the signal line.

[0047] DC signal voltage detection: The present invention has an analog voltmeter and a digital voltmeter for voltage measurement. The two measurement methods are complementary to each other to expand the scope of use of the instrument and can be selected according to needs.

[0048] The pulse counter used in the present invention includes an integration circuit, and has the ability to shape signals and eliminate signal interference for signals within the test range, and can complete online measurement or offline measurement.

[0049] The instrument of the present invention can provide a 7.4V DC voltage for use by an external sensor, such as providing a working power supply for the external sensor during offline measurement.

[0050] Functional description of the core part of the circuit: U4 is a 74HC14 six-NOT gate logic chip, which is mainly used to isolate and shape the input pulse signal, isolate the impact on the measured signal during online measurement, and shape it to provide a clean counting square wave (CLK) for the subsequent CD4026 decimal counting.

[0051] U1, U2, and U3 are CD4026 decimal counting and 7-segment LED digital tube driving logic chips. Three of these chips and 7-segment LED digital tubes form a three-digit counting circuit and display. The maximum display number of this group of LED digital tubes is 999. When the total number of pulses exceeds 999 times, LED4 will be lit, and the LED4 light will be controlled to turn on and off every time the count reaches 500 times.

[0052] Detailed workflow The pulse signal to be tested input from the J1 end is sent to the RC circuit composed of R2 and C2 through the third pin of the SW1-2 switch for filtering. After the interference caused by the transmission process is filtered out, the pulse signal is sent to the 1A pin of the U4 NOT gate circuit for signal waveform shaping. The shaped signal is output from the 1Y pin to the 2A input end and is subjected to a second signal shaping by the internal circuit of the U4 chip. The pulse signal after the secondary signal shaping is output from the 2Y pin to the 1 pin (CLK) of the U3 counter chip. The internal circuit of the chip completes the counting of the pulse signal and converts the pulse signal of each ten digit into the driving level of the 7-segment LED digital tube. The 6, 7, 9, 10, 11, 12, and 13 pins of the U3 counting chip output the seven-segment levels of FGEADBC to the corresponding segment input pins of the LED1 digital tube respectively, completing the display of the single-digit pulse signal 0 to 9. The tenth-digit pulse signal is output from the 5th pin of the U3 chip to the 1st pin of the U2 counter as the carry signal of the counter. Similarly, LED2 digital tube completes the display of the tens digit pulse signal 0 to 9, and LED3 digital tube completes the display of the hundreds digit pulse signal 0 to 9, thus realizing the requirement of counting and displaying the pulse signal 999. LED4 lamp is only used as the carry signal output by U1 to display the thousands digit. For the pulse signal that needs to be counted and observed, 999 counts are sufficient.

[0053] When measuring the quality of the reed switch alone, pressing the self-locking switch K1 can provide the reed switch with a bias voltage of about 3.5V for measurement. When measuring the rainfall and wind speed pulse signals online, K1 should be in the disconnected state.

[0054] K2 is the manual reset button for the counter. By providing 5V voltage to the 15th (RST) pin of cd4062, the cd4062 counter can be reset and the display of the LED digital tube can be cleared at any time. Example

[0055] The present embodiment is different from the first embodiment in that a detection system of a five-element signal comprehensive fault detector for an automatic weather station is provided, including a signal acquisition unit, a data preprocessing and feature extraction unit.

[0056] The signal acquisition unit is a sensor group 6, and the sensor group 6 includes a rainfall sensor, a wind speed sensor, a wind direction sensor, a temperature sensor and a humidity sensor.

[0057] Among them, the output resistance value RT of the temperature sensor and the temperature T usually satisfy the functional relationship, as follows: RT=RO(1+αT) (RO is the reference resistor and α is the temperature coefficient); a voltage divider circuit is formed by a precision resistor and a temperature sensor; Then, the resistance change is converted into a voltage change VT, which is amplified by an amplifier with an amplification factor A to obtain an amplified voltage signal VOUT; Where, VOUT = A×VT. The analog voltage signal is input into the analog-to-digital converter (ADC) and converted into a digital signal. According to the Nyquist sampling theorem, the sampling frequency fs of the ADC should be at least twice the highest frequency fmax of the signal, that is, fs ≥ 2fmax. For signals with relatively slow changes such as temperature and humidity, fs can be set to 1Hz. For signals with fast changes such as wind speed and wind direction, fs can be increased to above 10Hz; Among them, a low-pass filter, a high-pass filter and a band-pass filter are added to the signal conditioning circuit.

[0058] It also includes a clock calibration unit, which uses a clock chip as the clock source of the system and performs clock calibration through GPS or Beidou satellite signals. In the signal acquisition process, the acquisition time of each sensor is triggered by a unified clock signal, and the measurement data of different elements have a strict correspondence in time; Assume that the collection time is tn, for the nth data Tn collected by the temperature sensor, the nth data Hn collected by the humidity sensor, the nth data Pn collected by the air pressure sensor, the nth data Wn collected by the wind speed sensor and the nth data Dn collected by the wind direction sensor, the collection time error △t satisfies |△t|≤1μs.

[0059] The data preprocessing and feature extraction unit includes the following: Data cleaning and outlier removal: The sliding window median filter algorithm is used to clean the collected data to remove noise and outliers. Assume that the sliding window size is m. For the measurement value Xi at a certain moment, the output value yi after filtering is the median of the data in the window. That is, yi=median{Xi-(m-1)÷2, ..., Xi, ..., Xi+(m-1)÷2}; At the same time, the 3σ criterion in statistical methods is combined to further identify and eliminate outliers; for a data sequence of a certain meteorological element {X1, X2, ..., Xn}, its mean X and standard deviation σ are first calculated. If a certain measurement value Xj satisfies |Xj -X|>3σ, it is judged as an outlier and eliminated; Feature extraction and parameter calculation: The key characteristic parameters that can reflect the change characteristics of meteorological elements and signal quality are extracted from the preprocessed data. For temperature data, the daily change amplitude △T24h within 24 hours is calculated. The formula is △T24h=Tmax-Tmin (where Tmax and Tmin are the highest and lowest temperatures within 24 hours, respectively), and the temperature change rate rT, rT=△T÷△t (△T is the temperature change within a certain time interval △t); For rainfall data, the core component reed switch is used for cumulative calculation, with the unit of measurement being 1 mm. For humidity data, calculate the daily average relative humidity H and the standard deviation of humidity variation σH; For wind speed data, calculate the gust coefficient Gw of wind speed, Gw=Wgust÷Wavg (Wgust is the wind speed, Wavg is the average wind speed) and the spectrum characteristics of wind speed. Obtain the spectrum distribution of wind speed signal through fast Fourier transform (FFT), analyze the frequency range where its energy is concentrated, and determine whether there is abnormal frequency component. For wind direction data, the distribution probability of wind direction in each direction P(Di), i=1, 2, ..., 16 (wind direction is divided into 16 directions), and the frequency of wind direction change fD are obtained by calculating the number of times the wind direction changes in unit time.

[0060] By setting up the above technical solution, the correlation and change trend between the data of each factor can be accurately analyzed in the future, and the accuracy of fault diagnosis can be improved. It is not only convenient to judge the fault, but also through this rule-based reasoning diagnosis method, common fault conditions can be quickly preliminarily judged and located, providing direction for further accurate diagnosis.

[0061] While the embodiments of the present invention have been shown and described, it will be apparent to those skilled in the art that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An automatic weather station five-element signal comprehensive fault detector and system, characterized in that: The invention comprises a power supply box (1), a control circuit board (2) and a voltmeter (3); the voltmeter (3) and the power supply box (1) are electrically coupled to the control circuit board (2); and a rain sensor, a wind speed sensor, a wind direction sensor, a temperature sensor and a humidity sensor are electrically coupled to the control circuit board (2).

2. A detection system for an automatic weather station five-element signal comprehensive fault detector and system as claimed in claim 1, characterized in that: It includes a signal acquisition unit, a data preprocessing and feature extraction unit.

3. The automatic weather station five-element signal comprehensive fault detector and system according to claim 2, characterized in that: The signal acquisition unit is a sensor group (6), and the sensor group (6) comprises a rainfall sensor, a wind speed sensor, a wind direction sensor, a temperature sensor and a humidity sensor.

4. The automatic weather station five-element signal comprehensive fault detector and system according to claim 3, characterized in that: in, The output resistance value RT of the temperature sensor usually satisfies the functional relationship with the temperature T, as follows: RT=RO(1+αT) (RO is the reference resistor and α is the temperature coefficient); a voltage divider circuit is formed by a precision resistor and a temperature sensor; Then, the resistance change is converted into a voltage change VT, which is amplified by an amplifier with an amplification factor A to obtain an amplified voltage signal VOUT; Where, VOUT = A×VT. The analog voltage signal is input into the analog-to-digital converter (ADC) and converted into a digital signal. According to the Nyquist sampling theorem, the sampling frequency fs of the ADC should be at least twice the highest frequency fmax of the signal, that is, fs ≥ 2fmax. For signals with relatively slow changes such as temperature and humidity, fs can be set to 1Hz. For signals with fast changes such as wind speed and wind direction, fs can be increased to above 10Hz; Among them, a low-pass filter, a high-pass filter and a band-pass filter are added to the signal conditioning circuit.

5. The automatic weather station five-element signal comprehensive fault detector and system according to claim 4 is characterized in that: It also includes a clock calibration unit, which uses a clock chip as the clock source of the system and performs clock calibration through GPS or Beidou satellite signals. In the signal acquisition process, the acquisition time of each sensor is triggered by a unified clock signal, and the measurement data of different elements have a strict correspondence in time; Assume that the collection time is tn, for the nth data Tn collected by the temperature sensor, the nth data Hn collected by the humidity sensor, the nth data Pn collected by the air pressure sensor, the nth data Wn collected by the wind speed sensor and the nth data Dn collected by the wind direction sensor, the collection time error △t satisfies |△t|≤1μs.

6. The automatic weather station five-element signal comprehensive fault detector and system according to claim 5, characterized in that: The data preprocessing and feature extraction unit includes the following: Data cleaning and outlier removal: The sliding window median filter algorithm is used to clean the collected data to remove noise and outliers. Assume that the sliding window size is m. For the measurement value Xi at a certain moment, the output value yi after filtering is the median of the data in the window. That is, yi=median{Xi-(m-1)÷2, ..., Xi, ..., Xi+(m-1)÷2}; At the same time, the 3σ criterion in statistical methods is combined to further identify and eliminate outliers; for a data sequence of a certain meteorological element {X1, X2, ..., Xn}, its mean X and standard deviation σ are first calculated. If a certain measurement value Xj satisfies |Xj -X|>3σ, it is judged as an outlier and eliminated; Feature extraction and parameter calculation: The key characteristic parameters that can reflect the change characteristics of meteorological elements and signal quality are extracted from the preprocessed data. For temperature data, the daily change amplitude △T24h within 24 hours is calculated. The formula is △T24h=Tmax-Tmin (where Tmax and Tmin are the highest and lowest temperatures within 24 hours, respectively), and the temperature change rate rT, rT=△T÷△t (△T is the temperature change within a certain time interval △t); For rainfall data, the core component reed switch is used for cumulative calculation, with the unit of measurement being 1 mm. For humidity data, calculate the daily average relative humidity H and the standard deviation of humidity variation σH; For wind speed data, calculate the gust coefficient Gw of wind speed, Gw=Wgust÷Wavg (Wgust is the wind speed, Wavg is the average wind speed) and the spectrum characteristics of wind speed. Obtain the spectrum distribution of wind speed signal through fast Fourier transform (FFT), analyze the frequency range where its energy is concentrated, and determine whether there is abnormal frequency component. For wind direction data, the distribution probability of wind direction in each direction P(Di), i=1, 2, ..., 16 (wind direction is divided into 16 directions), and the frequency of wind direction change fD are obtained by calculating the number of times the wind direction changes in unit time.

Images