WO2016165600A1 - 用于识别无感无刷电机初始位置的时变信号采样方法 - Google Patents

用于识别无感无刷电机初始位置的时变信号采样方法 Download PDF

Info

Publication number
WO2016165600A1
WO2016165600A1 PCT/CN2016/079011 CN2016079011W WO2016165600A1 WO 2016165600 A1 WO2016165600 A1 WO 2016165600A1 CN 2016079011 W CN2016079011 W CN 2016079011W WO 2016165600 A1 WO2016165600 A1 WO 2016165600A1
Authority
WO
WIPO (PCT)
Prior art keywords
sampling
rotor
time
sum
brushless motor
Prior art date
Application number
PCT/CN2016/079011
Other languages
English (en)
French (fr)
Inventor
曾菊阳
郭江龙
徐翔
陈赖平
Original Assignee
曾菊阳
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 曾菊阳 filed Critical 曾菊阳
Publication of WO2016165600A1 publication Critical patent/WO2016165600A1/zh

Links

Images

Classifications

    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02P—CONTROL OR REGULATION OF ELECTRIC MOTORS, ELECTRIC GENERATORS OR DYNAMO-ELECTRIC CONVERTERS; CONTROLLING TRANSFORMERS, REACTORS OR CHOKE COILS
    • H02P6/00—Arrangements for controlling synchronous motors or other dynamo-electric motors using electronic commutation dependent on the rotor position; Electronic commutators therefor
    • H02P6/14—Electronic commutators
    • H02P6/16—Circuit arrangements for detecting position
    • H02P6/18—Circuit arrangements for detecting position without separate position detecting elements
    • H02P6/183—Circuit arrangements for detecting position without separate position detecting elements using an injected high frequency signal
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02P—CONTROL OR REGULATION OF ELECTRIC MOTORS, ELECTRIC GENERATORS OR DYNAMO-ELECTRIC CONVERTERS; CONTROLLING TRANSFORMERS, REACTORS OR CHOKE COILS
    • H02P6/00—Arrangements for controlling synchronous motors or other dynamo-electric motors using electronic commutation dependent on the rotor position; Electronic commutators therefor
    • H02P6/14—Electronic commutators
    • H02P6/16—Circuit arrangements for detecting position
    • H02P6/18—Circuit arrangements for detecting position without separate position detecting elements
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02P—CONTROL OR REGULATION OF ELECTRIC MOTORS, ELECTRIC GENERATORS OR DYNAMO-ELECTRIC CONVERTERS; CONTROLLING TRANSFORMERS, REACTORS OR CHOKE COILS
    • H02P2203/00—Indexing scheme relating to controlling arrangements characterised by the means for detecting the position of the rotor
    • H02P2203/11—Determination or estimation of the rotor position or other motor parameters based on the analysis of high-frequency signals

Definitions

  • the invention relates to the technical field of motor drive, and in particular to a time-varying signal sampling method for identifying an initial position of a non-inductive brushless motor, wherein the non-inductive brushless motor comprises a brushless DC motor and a permanent magnet synchronous motor.
  • non-inductive brushless motors including brushless DC motors and permanent magnet synchronous motors have been used in industrial control, home appliance and new energy fields. The more widely used.
  • a non-sensorless brushless motor without a position sensor has significant advantages such as small size, low cost, difficulty in interference of the control system, and high reliability.
  • the initial position of the motor rotor can be obtained by the sensor, while the initial position of the rotor without the position sensorless brushless motor can usually be determined by a pre-positioning method or position recognition.
  • the method for determining the initial position of the rotor of the non-inductive brushless motor is mainly the pre-positioning method.
  • the pre-positioning method has relatively high requirements on the power capacity of the driver, and has shortcomings such as slow start, high probability of motor reversal during pre-positioning, and oscillation, which are not allowed in some cases.
  • Rotor initial position identification is currently based on theoretical studies based on the inductive method.
  • the principle of the inductive method is that the inductance of the stator winding of the motor is related to the position of the rotor.
  • the UVW three-phase for the non-inductive brushless motor is sequentially U-V negative, U-W Negative, V positive W negative, V positive U negative, W positive U negative, W positive V negative total six power-up modes), measure and compare the size of the six current sampling signals generated to identify the initial position of the rotor.
  • Inductive method does not require pre-positioning relative to the pre-positioning method, but is initiated directly after identifying the initial position of the rotor. Under the premise of accurately identifying the initial position of the rotor, the motor has significant advantages such as large starting torque, fast starting, no reversal at startup, and no oscillation at startup.
  • the initial position recognition of the rotor based on the inductance method has the following disadvantages: A) the applied voltage pulse time is short, so the current sampling signal is small; B) there is power supply interference, CPU core interference and I/O port. Equal interference; C) The difference in inductance of the motor at different positions is small; D) The resolution of the ADC is limited. Due to the objective existence of the above unfavorable factors, it is difficult to obtain accurate sampling values when using the traditional inductance method for time-varying signal AD sampling, so that the initial position of the rotor cannot be accurately identified, which leads to the method of inductively identifying the initial position of the rotor. Did not get the general Apply all over.
  • the technical problem to be solved by the present invention is: for the above technical problem of the prior art, it is provided that there is no need to increase the hardware circuit, there is no need to improve the resolution of the AD converter, the anti-interference is good, and the number of resolution bits is high, A time-varying signal sampling method for identifying the initial position of a non-inductive brushless motor that produces significant motor noise that does not cause the motor rotor to rotate or dither.
  • the technical solution adopted by the present invention is:
  • a time varying signal sampling method for identifying an initial position of a non-inductive brushless motor comprising:
  • the high frequency filtering in the step 2) specifically refers to performing high frequency filtering by using a low pass filter, and adjusting the cutoff frequency f 0 of the low pass filter so that the signal filtered by the low pass filter is included
  • the white noise amplitude is greater than the 1 least significant bit LSB of the AD sampler.
  • the extraction in the step 4) specifically refers to one of the following three to three methods: 1.
  • the cumulative and SUM i are right shifted by lgN/lg4 bits as the extraction result, where N is 4 Positive integer power; 2.
  • Divide the sum and SUM i by the sum of the number of sampling results contained in the sum and S i and retain the specified precision as the extraction result; 3.
  • the step of extracting the summation and SUM i in the step 4) further includes the step of performing data filtering on all the obtained sampling results, where the data screening specifically refers to: first, by recording the sampling result in step 2), Then, for the number of loops N times, the number of loops obtained by step 2) is N times, the maximum value is selected and the maximum value is removed from the summation SUM i , or the minimum value is selected and the sum SUM i is extracted from The minimum value is removed, or both the maximum and minimum values are selected and the maximum and minimum values are removed from the accumulation sum SUM i .
  • the step 1) before the setting of the number of cycles N further comprises the step of calibrating the number N of cycles, the detailed steps comprising:
  • the short-time voltage pulse of the specified amplitude is sequentially applied to the non-inductive brushless motor according to the six power-on modes, and the pulse width ⁇ T 0 of the short-time voltage pulse is greater than or equal to the sampling time ⁇ T when the sample-and-hold circuit is not used.
  • the pulse width ⁇ T 0 of the short-time voltage pulse is less than the sampling time ⁇ T, and the time-varying current sampling signal outputted by the current-sampling resistor R SNS of the non-inductive brushless motor is subjected to high-frequency filtering after power-on.
  • the six accumulations and SUM1 ⁇ SUM6 are respectively extracted, and a total of six extraction results are obtained; the extraction specifically refers to one of the following three to three methods: 1. Accumulate and SUM1 to SUM6 respectively. The lgN/lg4 bits are shifted as the extraction results S1 to S6, where N is a positive integer power of 4; 2. The average value obtained by dividing the summation and SUM1 to SUM6 by M (retaining sufficient accuracy) is used as the extraction result S1. ⁇ S6; 3, the cumulative sum and SUM1 ⁇ SUM6 directly as the extraction results S1 ⁇ S6;
  • step 1.5 judging whether the position discrimination mode MOD is 1, if yes, the jump proceeds to step 1.6); if not, then jumps to step 1.11);
  • step 1.14 Judging whether S1 is the largest in the extraction results S1, S2, and S3. If the extraction result S1 is the largest, it is determined whether the extraction result S1 is greater than S4. If S1 is greater than S4, the value of the rotor recognition position pos is 1 and jumps. To step 1.14), if S1 is greater than S4, the value of the rotor identification position pos is 4 and jumps to step 1.14); if the extraction result S1 is not the maximum, step 1.12);
  • test number CNT is decremented by 1, and it is judged whether the new test number CNT is 0. If the new test number CNT is not 0, the jump proceeds to step 1.2); if the new test number CNT is 0, the jump execution is performed. Step 1.17);
  • the time varying signal sampling method for identifying the initial position of the non-inductive brushless motor of the present invention has the following advantages: First, the present invention uses a short-time voltage pulse of a specified amplitude, and a pulse width of a short-time voltage pulse when the sample-and-hold circuit is not used.
  • the time-varying current sampling signal outputted by the current sampling resistor R SNS of the non-inductive brushless motor after power-on includes the interference signal of power interference, CPU core interference, I/O port and other approximate white noise, and the low pass is adjusted.
  • the time-varying signal sampling method for identifying the initial position of the non-inductive brushless motor of the present invention does not increase the hardware circuit and does not use the higher-resolution AD converter.
  • the resolution and anti-interference performance of the AD conversion can be improved, and even if a relatively short voltage pulse is applied, the size of the six current sampling signals can be accurately distinguished, thereby accurately identifying the initial position of the rotor;
  • a relatively short voltage pulse can be applied when identifying the initial position of the rotor, so that neither significant motor noise nor motor rotor rotation or jitter is generated during the identification process, and there is no need to add hardware circuitry to the AD converter.
  • FIG. 1 is a schematic diagram of a basic process of a method according to an embodiment of the present invention.
  • FIG. 2 is a schematic diagram of waveforms of time-varying signal sampling in a method according to an embodiment of the present invention.
  • Figure 3 shows the first six sets of measured data for the initial position of the rotor of a brushless DC motor when N is 1.
  • Figure 4 shows the last six sets of measured data for the initial position of the rotor of a brushless DC motor when N is 1.
  • Figure 5 shows the first six sets of measured data for the initial position of the rotor of a brushless DC motor when N is 2.
  • Figure 6 shows the last six sets of measured data for the initial position of the rotor of a brushless DC motor when N is 2.
  • Figure 7 shows the first six sets of measured data for the initial position of the rotor of a brushless DC motor when N is 4.
  • Figure 8 shows the last six sets of measured data for the initial position of the rotor of a brushless DC motor when N is 4.
  • Figure 9 shows the first six sets of measured data for the initial position of the rotor of a brushless DC motor when N is 20.
  • Figure 10 shows the last six sets of measured data for the initial position of the rotor of a brushless DC motor when N is 20.
  • Figure 11 shows the first six sets of measured data for the initial position of the rotor of a brushless DC motor by the inductive method when N is 40.
  • Figure 12 shows the last six sets of measured data for the initial position of the rotor of a brushless DC motor by the inductive method when N is 40.
  • Figure 13 is the first six sets of measured data for the initial position of the rotor of a permanent magnet synchronous motor by the inductive method when N is 1.
  • Figure 14 shows the last six sets of measured data for the initial position of the rotor of a permanent magnet synchronous motor by the inductive method when N is 1.
  • Figure 15 shows the first six sets of measured data for the initial position of the rotor of a permanent magnet synchronous motor by the inductive method when N is 2.
  • Figure 16 shows the last six sets of measured data of the initial position of the rotor of a permanent magnet synchronous motor by the inductive method when N is 2.
  • Fig. 17 is the first six sets of measured data of the initial position of the rotor of a permanent magnet synchronous motor by the inductive method when N is 4.
  • Figure 18 is the last six sets of measured data for the initial position of the rotor of a permanent magnet synchronous motor by the inductive method when N is 4.
  • Fig. 19 is the first six sets of measured data of the initial position of the rotor of a permanent magnet synchronous motor by the inductive method when N is 20.
  • Figure 20 is the last six sets of measured data for the initial position of the rotor of a permanent magnet synchronous motor by the inductive method when N is 20.
  • Figure 21 shows the first six sets of measured data for the initial position of the rotor of a permanent magnet synchronous motor by the inductive method when N is 40.
  • Fig. 22 shows the last six sets of measured data of the initial position of the rotor of a permanent magnet synchronous motor by the inductive method when N is 40.
  • Fig. 23 is a result obtained when the measured data of the same measurement as Fig. 17 is in the second recognition mode.
  • Fig. 24 is a result obtained when the measured data identical to Fig. 18 is in the second recognition mode.
  • FIG. 25 is a schematic flowchart of the calibration cycle number N in the first embodiment of the present invention.
  • the steps of the time varying signal sampling method for identifying the initial position of the non-inductive brushless motor in this embodiment include:
  • Condition 1 The input signal is a constant DC voltage
  • Condition 2 The noise superimposed on the input signal is white noise or approximately white noise, and the white noise has sufficient amplitude.
  • the oversampling condition is met, the in-band noise will be reduced by 3 dB for each doubling of the sampling frequency, and the measurement resolution will be increased by 1/2 bit.
  • the voltage pulse width ⁇ T is small, the current sampling signal is a time-varying signal of approximately triangular wave, and the time-varying signal obviously cannot satisfy the first condition required by the oversampling method, because the signal is always in an approximate straight line during multiple sampling. Significant changes in the rise.
  • the time-varying signal sampling method of this embodiment is essentially a discrete over-sampling method, and the principle is: repeating the same current sampling signal (such as the current sampling signal generated when the voltage pulse U is positive and negative V) is repeated N times. If the influence of various interferences is not counted, the amplitude of the current sampling signal at the same sampling time ⁇ T is stable every time, as shown in the equation (1), so that the condition 1 of oversampling can be satisfied.
  • the embodiment considers that the actual input signal of the low-pass filter includes interference signals such as power supply interference, CPU core interference, and I/O port, and low-pass filtering can be performed by performing high-frequency filtering on the input signal of the low-pass filter.
  • the output signal of the filter is superimposed with a certain amplitude of approximate white noise interference, so that the condition 2 of oversampling can be satisfied. Therefore, in this embodiment, the selected power-on mode is repeated N times (N is large enough), and each current sampling signal is subjected to high-frequency filtering and amplification, and AD sampling is performed at time ⁇ T. For the selected power-on mode, the amplitude of the N sampling signals is substantially constant at ⁇ T, and the signal is superimposed.
  • the interference with approximate white noise is added, so the oversampling condition is met, and the N samples are accumulated and extracted, so that discrete oversampling values with good anti-interference and high resolution can be obtained, which is shorter in this embodiment.
  • the voltage pulse can accurately identify the initial position of the rotor.
  • V SNS R SNS ⁇ (V MOTOR /R)(1-e - ⁇ T ⁇ R/L ) (1)
  • V SNS represents the amplitude of the current sampling signal at the sampling time ⁇ T
  • R SNS represents the resistance value of the current sampling resistor of the non-inductive brushless motor
  • V MOTOR represents the amplitude of the short-time voltage pulse
  • R represents no.
  • the current sampling resistor of the brushless motor is connected in series with the total resistance of the two switching tubes after the on-resistance
  • L represents the current power-on mode of the non-inductive brushless motor corresponding to the inductance of the two-phase winding coil.
  • V SNS is the amount of change, and the longer the ⁇ T time, the current at the sampling time ⁇ T The amplitude V SNS of the sampled signal is larger.
  • the allowable range of the amplitude V SNS of the current sampling signal at the sampling time ⁇ T needs to consider two factors: First, the increase of the amplitude V SNS of the current sampling signal at the sampling time ⁇ T causes the gate-source voltage V of the field effect switching transistor GS is reduced, but the on-resistance of the switching tube is not significantly increased; secondly, the amplitude of the current sampling signal at the sampling time ⁇ T is V SNS for high-frequency filtering, and the voltage input to the input terminal of the AD converter is not exceeded.
  • the longer the ⁇ T time the more obvious the difference in the amplitude of the current generated by the various power-on modes, but the longer the sampling time ⁇ T, the longer the generation
  • the motor noise is greater, and even the rotor of the motor rotates.
  • T 1 , T 2 , T 3 , ..., T N on the t-axis respectively indicate the start times of the first , second , third , ..., N voltage pulses, and the time T1 can be 0;
  • ⁇ T is per The pulse width of a short-time voltage pulse;
  • the curve L is a curve obtained by smoothly connecting the amplitude points of the current sampling signals at the respective ⁇ T times of the first, second, third, ..., N voltage pulses, and T 1 to T N, N times the current sense signal amplitude is smoothly connected to the respective point of times ⁇ T time together to form a curve L in FIG.
  • the high frequency filtering in step 2) specifically refers to performing high frequency filtering by using a low pass filter, and adjusting the cutoff frequency f 0 of the low pass filter so that the signal filtered by the low pass filter is included.
  • the white noise amplitude is greater than the 1 least significant bit LSB of the AD sampler.
  • the extraction in step 4) specifically refers to one of the following three to three methods: 1.
  • the cumulative and SUM i are right shifted by lgN/lg4 bits as the extraction result, where N is 4 Positive integer power; 2.
  • Divide the sum and SUM i by the sum of the number of sampling results contained in the sum and S i and retain the specified precision as the extraction result; 3.
  • the above three methods of 1-3 are not exhaustive for the data collection.
  • those skilled in the art can also adopt other types of extraction methods as needed, which can also be completed.
  • the extraction of all sampling results can also achieve the purpose of increasing the resolution of AD conversion and improving anti-interference performance.
  • the extraction in step 4) specifically refers to method 3, and all the sampling results to be obtained are accumulated, and the accumulated sum is added as the extraction result, and the advantages of method 3 are accumulated to make the finally obtained discrete
  • the difference between the extracted results (accumulated sum) is greater, ensuring that the up and down perturbations between the discrete extraction results are greater, ensuring that condition 2 of the oversampling method is met.
  • the method before the step of extracting the summation and SUM i in step 4), the method further includes the step of performing data filtering on all the obtained sampling results, where the data filtering specifically refers to: first recording the sampling result in step 2). Then, for the number of loops N times, the number of loops obtained by step 2) is N times, the maximum value is selected and the maximum value is removed from the summation SUM i , or the minimum value is selected and the sum SUM i is extracted from The minimum value is removed, or both the maximum and minimum values are selected and the maximum and minimum values are removed from the accumulation sum SUM i .
  • the noise in the sampling result can be removed, and the accuracy of the sampling result can be improved.
  • S1 to S6 represent six extraction results (accumulated sum) extracted by the time varying signal sampling method of the present embodiment
  • POS represents the actual position of the rotor
  • pos represents the rotor identification position identified by the inductive method.
  • i denotes an intermediate variable obtained by inductively identifying the initial position of the rotor, and the conversion relationship between the intermediate variable i and the rotor identification position pos is shown in Table 1.
  • Figure 3 and Figure 4 show 12 rounds of measured data using the inductive method to identify the initial position of the rotor when the number of cycles N is 1, for six rotors numbered 1 to 6 Position, the correct rate is 50%, and the correct rate of each round of measured data is 0%.
  • Figure 5 and Figure 6 show 12 rounds of measured data using the inductive method to identify the initial position of the rotor when the number of cycles N is 2. For the actual positions of the six rotors numbered 1 to 6, the correct rate is 64%. The correct rate of all rounds of measured data is 0%.
  • Figure 7 and Figure 8 show 12 rounds of measured data using the inductive method to identify the initial position of the rotor when the number of cycles N is 4.
  • Figure 13 and Figure 14 show 12 rounds of measured data using the inductive method to identify the initial position of the rotor when the number of cycles N is 1, for six rotors numbered 1 to 6. Position, the correct rate is 36%, and the correct rate of each round of measured data is 0%.
  • Figure 15 and Figure 16 show 12 rounds of measured data using the inductive method to identify the initial position of the rotor when the number of cycles N is 2. For the actual positions of the six rotors numbered 1 to 6, the correct rate is 46%. The correct rate of all rounds of measured data is 0%.
  • Figure 17 and Figure 18 show 12 rounds of measured data using the inductive method to identify the initial position of the rotor when the number of cycles N is 4.
  • the accuracy of the position recognition increases with the increase of the number of cycles N.
  • the accuracy of the initial position recognition of the rotor is obtained. To an acceptable level. Therefore, the setting of the number of cycles N has a very important effect on how to accurately and quickly complete the time-varying signal sampling method in the present embodiment.
  • a time-varying signal sampling method for a non-inductive brushless motor is faster and more accurate, and the number N of cycles of the non-inductive brushless motor needs to be calibrated, so that the cycle number N satisfies the accuracy of the initial position recognition of the rotor.
  • the acceptable degree is as small as possible, which saves time for time-varying signal sampling, making the initial position recognition of the rotor of the non-inductive brushless motor faster and faster.
  • step 1) in the embodiment further includes the step of calibrating the number of loops N before setting the number of loops N, and the detailed steps include:
  • the short-time voltage pulse of the specified amplitude is sequentially applied to the non-inductive brushless motor according to the six power-on modes, and the pulse width ⁇ T 0 of the short-time voltage pulse is greater than or equal to the sampling time ⁇ T when the sample-and-hold circuit is not used.
  • the pulse width ⁇ T 0 of the short-time voltage pulse is less than the sampling time ⁇ T, and the time-varying current sampling signal outputted by the current-sampling resistor R SNS of the non-inductive brushless motor is subjected to high-frequency filtering after power-on.
  • the six accumulations and SUM1 ⁇ SUM6 are respectively extracted, and a total of six extraction results are obtained; the extraction specifically refers to one of the following three to three methods: 1. Accumulate and SUM1 to SUM6 respectively. The lgN/lg4 bits are shifted as the extraction results S1 to S6, where N is a positive integer power of 4; 2. The average value obtained by dividing the summation and SUM1 to SUM6 by M (retaining sufficient accuracy) is used as the extraction result S1. ⁇ S6; 3, the cumulative sum and SUM1 ⁇ SUM6 directly as the extraction results S1 ⁇ S6;
  • step 1.5 judging whether the position discrimination mode MOD is 1, if yes, the jump proceeds to step 1.6); if not, then jumps to step 1.11);
  • step 1.14 Judging whether S1 is the largest in the extraction results S1, S2, and S3. If the extraction result S1 is the largest, it is determined whether the extraction result S1 is greater than S4. If S1 is greater than S4, the value of the rotor recognition position pos is 1 and jumps. To step 1.14), if S1 is greater than S4, the value of the rotor identification position pos is 4 and jumps to step 1.14); if the extraction result S1 is not the maximum, step 1.12);
  • test number CNT is decremented by 1, and it is judged whether the new test number CNT is 0. If the new test number CNT is not 0, the jump proceeds to step 1.2); if the new test number CNT is 0, the jump execution is performed. Step 1.17);
  • the rotor identification position pos is 5; when the intermediate variable i is 1, the rotor recognition position pos is 6; when the intermediate variable i is 0, The rotor identification position pos is 5; when the intermediate variable i is 3, the rotor identification position pos is 1; and so on, as shown in Table 1.
  • Table 1 Location lookup table.
  • the non-inductive brushless motor in this embodiment includes a UVW three-phase, and the number of pairs of rotor poles is four, so that the stator of the non-inductive brushless motor is marked with four sets of marks, and each set of marks includes 6 positions of 1 to 6 Therefore, one circumference is divided into positions. There will be four positions on the stator which are all identified as 1.
  • the actual position of the rotor of the non-inductive brushless motor is POS 1 in the above step 1.2)
  • the rotor of the non-inductive brushless motor is rotated, so that the rotor The positioning pointer is rotated to any position on the stator that is identified as 1.
  • the first position determination mode is steps 1.6) to 1.10), and the first position discrimination mode is an accurate discrimination method;
  • the second position discrimination mode is step 1.11) to 1.13), and the second position discrimination mode is a fuzzy discrimination method.
  • both position discrimination modes enable accurate calibration of the number of cycles N.
  • the second position discriminating mode has a smaller requirement on the number of loops N.
  • the correct rate of position discrimination is only 57%, and the correctness rate of 6 positions is 0% at the same time; the test data is identical to those of Figs. 17 and 18, and the second position discrimination mode is used to obtain Figs. 23 and 24, and the position discrimination is performed at this time. Positive The accuracy rate is 100%, and the accuracy of the 6 positions is 100%.
  • the second position discrimination mode can reduce the number of cycles, that is, shorten the recognition time, but which recognition mode is selected depends on the specific motor characteristics.

Landscapes

  • Engineering & Computer Science (AREA)
  • Power Engineering (AREA)
  • Control Of Ac Motors In General (AREA)
  • Control Of Motors That Do Not Use Commutators (AREA)

Abstract

一种用于识别无感无刷电机初始位置的时变信号采样方法，步骤包括：1）选择加电方式；设定循环次数和采样时间，设置累加和并清0；2）将指定幅度的短时电压脉冲按照选择的加电方式对无感无刷电机加电，在加电后对输出的时变电流取样信号进行高频滤波和放大，在采样时间进行AD采样，并将采样结果累加至累加和；3）将循环次数减１，判断循环次数是否为0，如为0则跳转执行步骤4）；如不为0则跳转执行步骤2）；4）对累加和进行抽取，将抽取结果输出。该方法不需增加硬件电路，不需提高AD转换器的分辨率，即可得到抗干扰性好，分辨率位数高的采样结果，不会产生明显电机噪声，识别时不会使电机转子转动或者抖动。

Description

用于识别无感无刷电机初始位置的时变信号采样方法 【技术领域】
本发明涉及电机驱动技术领域，具体涉及一种用于识别无感无刷电机初始位置的时变信号采样方法，其中无感无刷电机包括无刷直流电机和永磁同步电机。
【背景技术】
随着永磁材料、功率MOS管和控制芯片功能的不断改进和完善，包括无刷直流电机和永磁同步电机的无感无刷电机在工业控制领域、家电领域和新能源领域得到了越来越广泛的应用。相对有位置传感器的无刷电机而言，无位置传感器的无感无刷电机具有体积小、成本低、控制系统不易受干扰和可靠性高等显著优点。
众所周知，要使电机获得有效的启动转矩顺利启动，需要准确知道转子初始位置，如果获得的转子初始位置与实际值偏差较大，电机启动时将会出现带负载能力下降、甚至反转等问题。通常有位置传感器的电机转子初始位置可通过传感器获得，而无位置传感器无感无刷电机的转子初始位置通常可以通过预定位方法或位置识别来确定。
目前确定无感无刷电机转子初始位置的方法主要是预定位法。但是，预定位法对驱动器功率容量要求比较高，存在启动慢、预定位期间电机反转概率大、震荡等缺点，这在某些场合是不允许的。
转子初始位置识别目前一般是基于电感法展开的理论研究。电感法的原理是：电机定子绕组的电感量与转子位置有关，通过向电机任意两相施加短时电压脉冲(即针对无感无刷电机的UVW三相依次采用U正V负、U正W负、V正W负、V正U负、W正U负、W正V负共六种加电方式)，测量并比较由此产生的6个电流取样信号的大小来识别转子初始位置。相对于预定位法，电感法不需要预先定位，而是在识别转子的初始位置后直接启动。在能够准确识别转子初始位置的前提下，电机启动时具有启动转矩大、启动快、启动时不反转、启动时不震荡等显著优点。
但是在实际应用过程中，基于电感法的转子初始位置识别存在下述不利因素：A)所加电压脉冲时间短，因而电流取样信号小；B)存在电源干扰、CPU内核干扰和I/O口等干扰；C)不同位置电机电感量差别微小；D)ADC分辨率有限。正由于上述不利因素的客观存在，使得在技术上采用传统电感法进行时变信号AD采样时很难得到准确的采样值，以致不能准确识别转子初始位置，导致电感法识别转子初始位置的方法实际上并未得到普 遍应用。
【发明内容】
本发明要解决的技术问题是：针对现有技术的上述技术问题，提供一种不需要增加硬件电路，不需要提高AD转换器的分辨率，抗干扰性好，分辨率位数高，不会产生明显电机噪声，识别时不会使电机转子转动或者抖动的用于识别无感无刷电机初始位置的时变信号采样方法。
为了解决上述技术问题，本发明采用的技术方案为：
一种用于识别无感无刷电机初始位置的时变信号采样方法，步骤包括：
1)从无感无刷电机的UVW三相的六种加电方式U正V负、U正W负、V正W负、V正U负、W正U负、W正V负中选择一种加电方式；设定循环次数N和采样时间ΔT，设置累加和SUMi并清0；
2)将指定幅度的短时电压脉冲按照选择的加电方式对无感无刷电机加电，在不使用采样保持电路时所述短时电压脉冲的脉冲宽度ΔT0大于或等于采样时间ΔT，在使用采样保持电路时所述短时电压脉冲的脉冲宽度ΔT0小于采样时间ΔT，在加电后对无感无刷电机的电流取样电阻RSNS输出的时变电流取样信号进行高频滤波和放大，并在采样时间ΔT时刻对高频滤波和放大后的时变电流取样信号进行AD采样，并将采样结果累加至累加和SUMi；
3)将循环次数N减1，判断循环次数N是否为0，如果为0则跳转执行步骤4)；如果不为0则跳转执行步骤2)；
4)对累加和SUMi进行抽取，将抽取结果输出。
优选地，所述步骤2)中的高频滤波具体是指用低通滤波器进行高频滤波，通过调整低通滤波器的截止频率f0，使得经低通滤波器滤波后的信号所包含的白噪声幅度大于AD采样器的1个最低有效位LSB。
优选地，所述步骤4)中的抽取具体是指下述①～③三种方法中的一种：①、将累加和SUMi右移lgN/lg4位后作为抽取结果，其中N是4的正整数次幂；②、将累加和SUMi除以累加和Si中所包含的采样结果的数量得到的平均值并保留指定的精度后作为抽取结果；③、将累加和SUMi直接作为抽取结果。
优选地，所述步骤4)中对累加和SUMi进行抽取之前还包括对得到的所有采样结果进行数据筛选的步骤，所述数据筛选具体是指：首先通过在步骤2)中记录采样结果，然后针对循环次数N次执行步骤2)得到的循环次数N个采样结果，选择出其中的最大值并从累加和SUMi中去除该最大值，或者选择出其中的最小值并从累加和SUMi中去除该最 小值，或者同时选择出其中的最大值和最小值并从累加和SUMi中去除该最大值和最小值。
优选地，所述步骤1)设定循环次数N之前还包括标定循环次数N的步骤，详细步骤包括：
1.1)设定识别转子初始位置时所能接受的最小正确率P；设定采样时间ΔT；设定循环变量M的初始值，该初始值大于或等于4；设定用于统计正确率的测试次数CNT，设定位置判别模式MOD，设置累加和SUM1～SUM6并清0，且每一个累加和对应六种加电方式中的一种加电方式；
1.2)将无感无刷电机的转子实际位置POS定位到1；
1.3)将指定幅度的短时电压脉冲依次按照六种加电方式对无感无刷电机加电，在不使用采样保持电路时所述短时电压脉冲的脉冲宽度ΔT0大于或等于采样时间ΔT，在使用采样保持电路时所述短时电压脉冲的脉冲宽度ΔT0小于采样时间ΔT，在加电后对无感无刷电机的电流取样电阻RSNS输出的时变电流取样信号进行高频滤波和放大，并在采样时间ΔT时刻对经高频滤波和放大后的时变电流取样信号进行AD采样，并将采样结果按加电方式分别累加至对应的累加和，且对每一种加电方式重复步骤1.3)共M次；
1.4)将六个累加和SUM1～SUM6分别进行抽取，共得到六个抽取结果；所述抽取具体是指下述①～③三种方法中的一种：①、将累加和SUM1～SUM6分别右移lgN/lg4位后作为抽取结果S1～S6，其中N是4的正整数次幂；②、将累加和SUM1～SUM6分别除以M得到的平均值(保留足够的精度)后作为抽取结果S1～S6；③、将累加和SUM1～SUM6直接作为抽取结果S1～S6；
1.5)判断位置判别模式MOD是否为1，如果是，则跳转执行步骤1.6)；如果不是，则跳转至步骤1.11)；
1.6)设置用于位置识别的中间变量i为0；
1.7)判断抽取结果S1大于抽取结果S4是否成立，如果成立则将中间变量i加1；
1.8)判断抽取结果S2大于抽取结果S5是否成立，如果成立则将中间变量i加2；
1.9)判断抽取结果S3大于抽取结果S6是否成立，如果成立则将中间变量i加4；
1.10)根据中间变量i的值查找预设的位置查找表，得到转子识别位置pos，跳转至1.14)；
1.11)判断抽取结果S1、S2、S3中S1是否最大，如果抽取结果S1是最大，再判断抽取结果S1大于S4是否成立，如果S1大于S4成立则令转子识别位置pos的值为1并跳转至步骤1.14)，如果S1大于S4不成立则令转子识别位置pos的值为4并跳转至步骤1.14)；如果抽取结果S1不是最大，执行步骤1.12)；
1.12)判断抽取结果S1、S2、S3中S2是否最大，如果抽取结果S2是最大，再判断抽取结果S2大于S5是否成立，如果抽取结果S2大于S5成立则令转子识别位置pos的值为2并跳转至步骤1.14)，如果抽取结果S2大于S5不成立则令转子识别位置pos的值为5并跳转至步骤1.14)；如果抽取结果S2不是最大，执行步骤1.13)；
1.13)判断抽取结果S1、S2、S3中S3是否最大，如果抽取结果S3是最大，再判断抽取结果S3大于S6是否成立，如果成立则令转子识别位置pos的值为3并跳转至步骤1.14)，如果不成立则令转子识别位置pos的值为6并跳转至步骤1.14)；如果抽取结果S3不是最大，则增加循环变量M的值，重新设定用于统计正确率的测试次数CNT，跳转执行步骤1.2)；
1.14)记录转子识别位置pos和无感无刷电机当前的转子实际位置POS；
1.15)判断无感无刷电机的转子当前实际位置是否为6，如果不为6，预先将无感无刷电机的转子实际位置POS定位到下一位置，并跳转执行步骤1.3)；如果为6，则执行下一步；
1.16)将测试次数CNT减1，判断新的测试次数CNT是否为0，如果新的测试次数CNT不为0，则跳转执行步骤1.2)；如果新的测试次数CNT为0，则跳转执行步骤1.17)；
1.17)将每一条记录中转子识别位置pos和转子实际位置POS进行比较，在CNT次的测试中，统计出每次的六个位置识别全部正确的正确率C；判断正确率C小于识别转子初始位置时所能接受的最小正确率P是否成立，如果成立，则增加循环变量M的值，重新设定用于统计正确率的测试次数CNT，跳转执行步骤1.2)，如果不成立，则将当前的循环变量M的值作为六种加电方式共同的循环次数N的标定结果。
本发明用于识别无感无刷电机初始位置的时变信号采样方法具有下述优点：首先，本发明使用指定幅度的短时电压脉冲，在不使用采样保持电路时短时电压脉冲的脉冲宽度ΔT0大于或等于采样时间ΔT，在使用采样保持电路时短时电压脉冲的脉冲宽度ΔT0小于采样时间ΔT，在加电后采样时间ΔT时刻对高频滤波和放大后的时变电流取样信号进行AD采样得到采样结果，将上述操作重复循环次数N次，使得N次采集的时变电流取样信号的幅值基本不变，满足过采样方法中输入信号是一个恒定的直流电压的条件①；其次，在加电后无感无刷电机的电流取样电阻RSNS输出的时变电流取样信号中包含了电源干扰、CPU内核干扰、I/O口等近似白噪声的干扰信号，通过调整低通滤波器的截止频率f0，使得通过低通滤波器后信号所包含的白噪声幅度大于AD采样器的1个最低有效位LSB，从而能够满足过采样方法中输入信号上叠加的噪声为白噪声或近似白噪声，且白噪声有足够的幅度的条件②。基于上述两个条件①、条件②的满足，使得本发明用于识别无感无刷电机初 始位置的时变信号采样方法一方面在不增加硬件电路、不改用更高分辨率的AD转换器的前提下，能够提高AD转换的分辨率和抗干扰性能，即使施加相对较短的电压脉冲，也能准确分辨六种电流采样信号的大小，进而准确识别转子的初始位置；另一方面由于因识别转子的初始位置时可以施加相对较短的电压脉冲，因此在识别过程中既不会产生明显的电机噪声，也不会使电机转子转动或抖动，具有不需要增加硬件电路，对AD转换器要求低，抗干扰性好，分辨率高，不会产生明显电机噪声，识别时不会使电机转子转动或者抖动的优点。
【附图说明】
图1为本发明实施例方法的基本流程示意图。
图2为本发明实施例方法中时变信号采样的波形示意图。
图3为N为1时电感法识别某无刷直流电机的转子初始位置的前6组实测数据。
图4为N为1时电感法识别某无刷直流电机的转子初始位置的后6组实测数据。
图5为N为2时电感法识别某无刷直流电机的转子初始位置的前6组实测数据。
图6为N为2时电感法识别某无刷直流电机的转子初始位置的后6组实测数据。
图7为N为4时电感法识别某无刷直流电机的转子初始位置的前6组实测数据。
图8为N为4时电感法识别某无刷直流电机的转子初始位置的后6组实测数据。
图9为N为20时电感法识别某无刷直流电机的转子初始位置的前6组实测数据。
图10为N为20时电感法识别某无刷直流电机的转子初始位置的后6组实测数据。
图11为N为40时电感法识别某无刷直流电机的转子初始位置的前6组实测数据。
图12为N为40时电感法识别某无刷直流电机的转子初始位置的后6组实测数据。
图13为N为1时电感法识别某永磁同步电机的转子初始位置的前6组实测数据。
图14为N为1时电感法识别某永磁同步电机的转子初始位置的后6组实测数据。
图15为N为2时电感法识别某永磁同步电机的转子初始位置的前6组实测数据。
图16为N为2时电感法识别某永磁同步电机的转子初始位置的后6组实测数据。
图17为N为4时电感法识别某永磁同步电机的转子初始位置的前6组实测数据。
图18为N为4时电感法识别某永磁同步电机的转子初始位置的后6组实测数据。
图19为N为20时电感法识别某永磁同步电机的转子初始位置的前6组实测数据。
图20为N为20时电感法识别某永磁同步电机的转子初始位置的后6组实测数据。
图21为N为40时电感法识别某永磁同步电机的转子初始位置的前6组实测数据。
图22为N为40时电感法识别某永磁同步电机的转子初始位置的后6组实测数据。
图23为与图17完全相同的实测数据用第二种识别模式时得到的结果。
图24为与图18完全相同的实测数据用第二种识别模式时得到的结果。
图25为本发明实施例一中标定循环次数N的流程示意图。
【具体实施方式】
如图1所示，本实施例用于识别无感无刷电机初始位置的时变信号采样方法的步骤包括：
1)从无感无刷电机的UVW三相的六种加电方式U正V负、U正W负、V正W负、V正U负、W正U负、W正V负中选择一种加电方式；设定循环次数N和采样时间ΔT，设置累加和SUMi并清0；
2)将指定幅度的短时电压脉冲按照选择的加电方式对无感无刷电机加电，在不使用采样保持电路时所述短时电压脉冲的脉冲宽度ΔT0大于或等于采样时间ΔT，在使用采样保持电路时所述短时电压脉冲的脉冲宽度ΔT0小于采样时间ΔT，在加电后对无感无刷电机的电流取样电阻RSNS输出的时变电流取样信号进行高频滤波和放大，并在采样时间ΔT时刻对高频滤波和放大后的时变电流取样信号进行AD采样，并将采样结果累加至累加和SUMi；
3)将循环次数N减1，判断循环次数N是否为0，如果为0则跳转执行步骤4)；如果不为0则跳转执行步骤2)；
4)对累加和SUMi进行抽取，将抽取结果输出。
众所周知，采用过采样方法需满足如下两个条件，条件①：输入信号是一个恒定的直流电压；条件②：输入信号上叠加的噪声为白噪声或近似白噪声，且白噪声有足够的幅度。在满足过采样条件下，采样频率每增加一倍，带内噪声将减小3dB，而测量分辨率将增加1/2位。当电压脉冲宽度ΔT较小时，电流取样信号是近似三角波的时变信号，时变信号显然不能满足过采样方法所要求的第一个条件，因为在多次采样的过程中信号一直在呈近似直线上升的显著变化。本实施例的时变信号采样方法实质上是一种离散的过采样方法，其原理为：将同一个电流采样信号(比如加电压脉冲U正V负时产生的电流采样信号)重复产生N次，如果不计各种干扰的影响，每次在相同的采样时间ΔT时刻的电流采样信号的幅值是平稳的，如式(1)所示，从而能够满足过采样的条件①。同时，本实施例考虑到低通滤波器的实际输入信号包含有电源干扰、CPU内核干扰、I/O口等干扰信号，通过对低通滤波器的输入信号进行高频滤波，可以使低通滤波器的输出信号叠加了一定幅度的近似白噪声的干扰，从而能够满足过采样的条件②。因此，本实施例对选定的加电方式重复N次(N足够大)，对每次电流取样信号进行高频滤波和放大，并均在ΔT时刻进行AD采样。对选定的加电方式，这N次取样信号的幅值在ΔT时刻是基本不变的，且信号上叠 加了近似白噪声的干扰，因此符合过采样条件，对这N次采样值进行累加并抽取，就能获得抗干扰性好、分辨率较高的离散过采样值，采用本实施例用较短的电压脉冲即可准确识别转子的初始位置。
VSNS＝RSNS·(VMOTOR/R)(1-e-ΔT·R/L)        (1)
式(1)中，VSNS表示采样时间ΔT时刻的电流采样信号的幅值，RSNS表示无感无刷电机的电流取样电阻的电阻值，VMOTOR表示短时电压脉冲的幅度，R表示无感无刷电机的电流取样电阻串联两个开关管导通电阻后的总电阻，L表示无感无刷电机当前的加电方式对应两相绕组线圈的电感量。对特定的电机、特定转子位置、特定加电方式和特定母线电压而言，L、R、VMOTOR、RSNS都是常数，VSNS是变化量，ΔT时间越长，采样时间ΔT时刻的电流采样信号的幅值VSNS就越大。采样时间ΔT时刻的电流采样信号的幅值VSNS的允许范围需考虑两个因素：一是采样时间ΔT时刻的电流采样信号的幅值VSNS的增加会使场效应开关管的栅源电压VGS降低，但以不明显增大开关管的导通电阻为准；二是采样时间ΔT时刻的电流采样信号的幅值VSNS进行高频滤波、放大后进入AD转换器输入端的电压不要超过其电气特性承受的范围。在采样时间ΔT时刻的电流采样信号的幅值VSNS的允许范围内，ΔT时间越长，各种加电方式产生的电流幅值的差异性就越明显，但是采样时间ΔT时间越长，产生的电机噪声就越大，甚至电机转子发生转动。采用本实施例的方法，用较小的采样时间ΔT即可准确分辨各种加电方式产生的电流幅值的差异性。
参见图2，t轴上的T1、T2、T3、…、TN分别表示第1、2、3、…、N个电压脉冲的起始时刻，T1时刻可以为0；ΔT是每个短时电压脉冲的脉冲宽度；曲线L是由第1、2、3、…、N个电压脉冲在各自的ΔT时刻电流采样信号的幅值点平滑连接而成的曲线，将T1～TN的N次电流采样信号在各次的ΔT时刻的幅值点平滑地连接起来，就形成了图2中的曲线L，那么对同一个电流采样信号人为地重复产生N次并进行采样和抽取的过程，就等效于对曲线L的N次过采样，参见图2可知每个取样信号在ΔT时刻的幅值是基本不变的，而且受白噪声干扰信号的影响还会发生一定的上下波动，从而能够满足过采样方法的条件①和条件②，而根据过采样原理可知，N次过采样既增加了AD转换的分辨率，又提高了抗干扰性能。因此即使施加相对较小的短时电压脉冲，也能准确分辨六种电流采样信号的大小，进而准确识别转子的初始位置。
本实施例中，步骤2)中的高频滤波具体是指用低通滤波器进行高频滤波，通过调整低通滤波器的截止频率f0，使得经低通滤波器滤波后的信号所包含的白噪声幅度大于AD采样器的1个最低有效位LSB。
本实施例中，步骤4)中的抽取具体是指下述①～③三种方法中的一种：①、将累加 和SUMi右移lgN/lg4位后作为抽取结果，其中N是4的正整数次幂；②、将累加和SUMi除以累加和Si中所包含的采样结果的数量得到的平均值并保留指定的精度后作为抽取结果；③、将累加和SUMi直接作为抽取结果。需要说明的是，上述①～③三种方法并非为数据采集是所采用的抽取的穷举，毫无疑问，本领域的技术人员也可以根据需要采用其他类型的抽取方法，其同样也可以完成所有采样结果的抽取，同样也能够达到增加AD转换的分辨率、提高抗干扰性能的目的。本实施例中，步骤4)中的抽取具体是指方法③，即将得到的所有采样结果进行累加，将累加得到的累加和作为抽取结果，方法③的优点在进行累加后使得最终得到的离散的抽取结果(累加和)之间差异更大，确保离散的抽取结果之间上下扰动更大，能够确保满足过采样方法的条件②。
本实施例中，步骤4)中对累加和SUMi进行抽取之前还包括对得到的所有采样结果进行数据筛选的步骤，所述数据筛选具体是指：首先通过在步骤2)中记录采样结果，然后针对循环次数N次执行步骤2)得到的循环次数N个采样结果，选择出其中的最大值并从累加和SUMi中去除该最大值，或者选择出其中的最小值并从累加和SUMi中去除该最小值，或者同时选择出其中的最大值和最小值并从累加和SUMi中去除该最大值和最小值。毫无疑问，通过上述数据筛选，能够去除采样结果中的噪声，提升采样结果的精确度。
参见图3～图12所示的针对某无刷直流电机进行的实测数据、图13～图22所示的针对某永磁同步电机进行的实测数据。图3～图12中，S1～S6表示采用本实施例时变信号采样方法抽取得到的六个抽取结果(累加和)，POS表示转子实际位置，pos表示采用电感法识别得到的转子识别位置，i表示采用电感法识别转子初始位置得到的中间变量，中间变量i和转子识别位置pos之间的转换关系详见表1。
针对某无刷直流电机进行的实测数据中：图3和图4为循环次数N取值为1时采用电感法识别转子初始位置的12轮实测数据，针对编号为1～6的六个转子实际位置，其正确率总计为50％，而每一轮实测数据的全部正确率为0％。图5和图6为循环次数N取值为2时采用电感法识别转子初始位置的12轮实测数据，针对编号为1～6的六个转子实际位置，其正确率总计为64％，而每一轮实测数据的全部正确率为0％。图7和图8为循环次数N取值为4时采用电感法识别转子初始位置的12轮实测数据，针对编号为1～6的六个转子实际位置，其正确率总计为89％，而每一轮实测数据的全部正确率为33％。图9和图10为循环次数N取值为8时采用电感法识别转子初始位置的12轮实测数据，针对编号为1～6的六个转子实际位置，其正确率总计为93％，而每一轮实测数据的全部正确率为58％。图11和图12为循环次数N取值为40时采用电感法识别转子初始位置的12轮实测数据，针对编号为1～6的六个转子实际位置，其正确率总计为100％，而每一轮实测数据的全部 正确率为100％。
针对某永磁同步电机进行的实测数据中：图13和图14为循环次数N取值为1时采用电感法识别转子初始位置的12轮实测数据，针对编号为1～6的六个转子实际位置，其正确率总计为36％，而每一轮实测数据的全部正确率为0％。图15和图16为循环次数N取值为2时采用电感法识别转子初始位置的12轮实测数据，针对编号为1～6的六个转子实际位置，其正确率总计为46％，而每一轮实测数据的全部正确率为0％。图17和图18为循环次数N取值为4时采用电感法识别转子初始位置的12轮实测数据，针对编号为1～6的六个转子实际位置，其正确率总计为57％，而每一轮实测数据的全部正确率为0％。图19和图20为循环次数N取值为20时采用电感法识别转子初始位置的12轮实测数据，针对编号为1～6的六个转子实际位置，其正确率总计为93％，而每一轮实测数据的全部正确率为58％。图21和图22为循环次数N取值为40时采用电感法识别转子初始位置的12轮实测数据，针对编号为1～6的六个转子实际位置，其正确率总计为100％，而每一轮实测数据的全部正确率为100％。
经过上述实测数据表明,给定合适的脉冲宽度ΔT以后，随着循环次数N的增加，位置识别的准确率也随着增加，当循环次数N增加到一定值以后，转子初始位置识别的准确率达到可接受的程度。因此，循环次数N的设定对于本实施例如何精确、快速地完成时变信号采样方法具有非常重要的作用。毫无疑问，针对无感无刷电机可以设定一个通用的足够大的循环次数N(例如N＝40)即可使得转子初始位置识别的准确率达到可接受的程度；但是如要使得针对某一种无感无刷电机的时变信号采样方法更加快捷和精准，则需要针对该无感无刷电机的循环次数N进行标定，从而使得循环次数N在满足使得转子初始位置识别的准确率达到可接受的程度的前提下尽可能地小，从而节约时变信号采样的时间，使得无感无刷电机的转子初始位置识别更加快捷迅速。
需要说明的是，对于U正V负、U正W负、V正W负、V正U负、W正U负、W正V负六种加电方式，既可以采用统一的循环次数N，也可以根据不同加电方式采用不同的循环次数N。本实施例中，对于U正V负、U正W负、V正W负、V正U负、W正U负、W正V负六种加电方式采用统一的循环次数N。如图25所示，本实施例中步骤1)设定循环次数N之前还包括标定循环次数N的步骤，详细步骤包括：
1.1)设定识别转子初始位置时所能接受的最小正确率P；设定采样时间ΔT；设定循环变量M的初始值，该初始值大于或等于4；设定用于统计正确率的测试次数CNT，设定位置判别模式MOD，设置累加和SUM1～SUM6并清0，且每一个累加和对应六种加电方式中的一种加电方式；
1.2)将无感无刷电机的转子实际位置POS定位到1；
1.3)将指定幅度的短时电压脉冲依次按照六种加电方式对无感无刷电机加电，在不使用采样保持电路时所述短时电压脉冲的脉冲宽度ΔT0大于或等于采样时间ΔT，在使用采样保持电路时所述短时电压脉冲的脉冲宽度ΔT0小于采样时间ΔT，在加电后对无感无刷电机的电流取样电阻RSNS输出的时变电流取样信号进行高频滤波和放大，并在采样时间ΔT时刻对经高频滤波和放大后的时变电流取样信号进行AD采样，并将采样结果按加电方式分别累加至对应的累加和，且对每一种加电方式重复步骤A1.3)共M次；
1.4)将六个累加和SUM1～SUM6分别进行抽取，共得到六个抽取结果；所述抽取具体是指下述①～③三种方法中的一种：①、将累加和SUM1～SUM6分别右移lgN/lg4位后作为抽取结果S1～S6，其中N是4的正整数次幂；②、将累加和SUM1～SUM6分别除以M得到的平均值(保留足够的精度)后作为抽取结果S1～S6；③、将累加和SUM1～SUM6直接作为抽取结果S1～S6；
1.5)判断位置判别模式MOD是否为1，如果是，则跳转执行步骤1.6)；如果不是，则跳转至步骤1.11)；
1.6)设置用于位置识别的中间变量i为0；
1.7)判断抽取结果S1大于抽取结果S4是否成立，如果成立则将中间变量i加1；
1.8)判断抽取结果S2大于抽取结果S5是否成立，如果成立则将中间变量i加2；
1.9)判断抽取结果S3大于抽取结果S6是否成立，如果成立则将中间变量i加4；
1.10)根据中间变量i的值查找预设的位置查找表，得到转子识别位置pos，跳转至1.14)；
1.11)判断抽取结果S1、S2、S3中S1是否最大，如果抽取结果S1是最大，再判断抽取结果S1大于S4是否成立，如果S1大于S4成立则令转子识别位置pos的值为1并跳转至步骤1.14)，如果S1大于S4不成立则令转子识别位置pos的值为4并跳转至步骤1.14)；如果抽取结果S1不是最大，执行步骤1.12)；
1.12)判断抽取结果S1、S2、S3中S2是否最大，如果抽取结果S2是最大，再判断抽取结果S2大于S5是否成立，如果抽取结果S2大于S5成立则令转子识别位置pos的值为2并跳转至步骤1.14)，如果抽取结果S2大于S5不成立则令转子识别位置pos的值为5并跳转至步骤1.14)；如果抽取结果S2不是最大，执行步骤1.13)；
1.13)判断抽取结果S1、S2、S3中S3是否最大，如果抽取结果S3是最大，再判断抽取结果S3大于S6是否成立，如果成立则令转子识别位置pos的值为3并跳转至步骤1.14)，如果不成立则令转子识别位置pos的值为6并跳转至步骤1.14)；如果抽取结果S3 不是最大，则增加循环变量M的值，重新设定用于统计正确率的测试次数CNT，跳转执行步骤1.2)；
1.14)记录转子识别位置pos和无感无刷电机当前的转子实际位置POS；
1.15)判断无感无刷电机的转子当前实际位置是否为6，如果不为6，预先将无感无刷电机的转子实际位置POS定位到下一位置，并跳转执行步骤1.3)；如果为6，则执行下一步；
1.16)将测试次数CNT减1，判断新的测试次数CNT是否为0，如果新的测试次数CNT不为0，则跳转执行步骤1.2)；如果新的测试次数CNT为0，则跳转执行步骤1.17)；
1.17)将每一条记录中转子识别位置pos和转子实际位置POS进行比较，在CNT次的测试中，统计出每次的六个位置识别全部正确的正确率C；判断正确率C小于识别转子初始位置时所能接受的最小正确率P是否成立，如果成立，则增加循环变量M的值，重新设定用于统计正确率的测试次数CNT，跳转执行步骤1.2)，如果不成立，则将当前的循环变量M的值作为六种加电方式共同的循环次数N的标定结果。
本实施例中预设的位置查找表中，当中间变量i为0时，转子识别位置pos为5；当中间变量i为1时，转子识别位置pos为6；当中间变量i为0时，转子识别位置pos为5；当中间变量i为3时，转子识别位置pos为1；以此类推，详见表1。
表1：位置查找表。
i 0 1 2 3 4 5 6 7
pos 5 6 无效 1 4 无效 3 2
本实施例中的无感无刷电机包含UVW三相、转子极对数是4，因此无感无刷电机的定子上标记有4组标记，每一组标记中包含1～6共6个位置，因此一个圆周划分为个位置。定子上会存在四个均被标识为1的位置，前述步骤1.2)中将无感无刷电机的转子实际位置POS定位到1时，具体是指转动无感无刷电机的转子，使得转子的定位指针转动至定子上任意一个被标识为1的位置。
本实施例中，根据位置判别模式MOD的初始值设定采用两种可选的位置判别模式，第一种位置判别模式为步骤1.6)～1.10)，第一种位置判别模式为精确判别法；第二种位置判别模式为步骤1.11)～1.13)，第二种位置判别模式为模糊判别法。但是两种位置判别模式都能够实现对循环次数N的准确标定。和第一种位置判别模式相比，第二种位置判别模式对循环次数N的要求更小一些，以图17、18为例来说明如下：循环次数N＝4，采用第一种位置判别模式时，位置判别的正确率仅为57％，6个位置同时正确率为0％；与图17、18完全相同测试数据，采用第二种位置判别模式得到图23、24，此时位置判别的正 确率为100％，6个位置同时正确率为100％。很明显采用第二种位置判别模式能减少循环次数，也即缩短了识别时间，但是具体选用哪种识别模式，需要根据具体电机特性来决定。
以上所述仅是本发明的优选实施方式，本发明的保护范围并不仅局限于上述实施例，凡属于本发明思路下的技术方案均属于本发明的保护范围。应当指出，对于本技术领域的普通技术人员来说，在不脱离本发明原理前提下的若干改进和润饰，这些改进和润饰也应视为本发明的保护范围。

Claims (5)

  1. 一种用于识别无感无刷电机初始位置的时变信号采样方法，其特征在于步骤包括：
    1)从无感无刷电机的UVW三相的六种加电方式U正V负、U正W负、V正W负、V正U负、W正U负、W正V负中选择一种加电方式；设定循环次数N和采样时间ΔT，设置累加和SUMi并清0；
    2)将指定幅度的短时电压脉冲按照选择的加电方式对无感无刷电机加电，在不使用采样保持电路时所述短时电压脉冲的脉冲宽度ΔT0大于或等于采样时间ΔT，在使用采样保持电路时所述短时电压脉冲的脉冲宽度ΔT0小于采样时间ΔT，在加电后对无感无刷电机的电流取样电阻RSNS输出的时变电流取样信号进行高频滤波和放大，并在采样时间ΔT时刻对高频滤波和放大后的时变电流取样信号进行AD采样，并将采样结果累加至累加和SUMi；
    3)将循环次数N减1，判断循环次数N是否为0，如果为0则跳转执行步骤4)；如果不为0则跳转执行步骤2)；
    4)对累加和SUMi进行抽取，将抽取结果输出。
  2. 根据权利要求1所述的用于识别无感无刷电机初始位置的时变信号采样方法，其特征在于：所述步骤2)中的高频滤波具体是指用低通滤波器进行高频滤波，通过调整低通滤波器的截止频率f0，使得经低通滤波器滤波后的信号所包含的白噪声幅度大于AD采样器的1个最低有效位LSB。
  3. 根据权利要求2所述的用于识别无感无刷电机初始位置的时变信号采样方法，其特征在于，所述步骤4)中的抽取具体是指下述①～③三种方法中的一种：①、将累加和SUMi右移lgN/lg4位后作为抽取结果，其中N是4的正整数次幂；②、将累加和SUMi除以累加和Si中所包含的采样结果的数量得到的平均值并保留指定的精度后作为抽取结果；③、将累加和SUMi直接作为抽取结果。
  4. 根据权利要求3所述的用于识别无感无刷电机初始位置的时变信号采样方法，其特征在于：所述步骤4)中对累加和SUMi进行抽取之前还包括对得到的所有采样结果进行数据筛选的步骤，所述数据筛选具体是指：首先通过在步骤2)中记录采样结果，然后针对循环次数N次执行步骤2)得到的循环次数N个采样结果，选择出其中的最大值并从累加和SUMi中去除该最大值，或者选择出其中的最小值并从累加和SUMi中去除该最小值，或者同时选择出其中的最大值和最小值并从累加和SUMi中去除该最大值和最小值。
  5. 根据权利要求1～4中任意一项所述的用于识别无感无刷电机初始位置的时变信号 采样方法，其特征在于：所述步骤1)设定循环次数N之前还包括标定循环次数N的步骤，详细步骤包括：
    1.1)设定识别转子初始位置时所能接受的最小正确率P；设定采样时间ΔT；设定循环变量M的初始值，该初始值大于或等于4；设定用于统计正确率的测试次数CNT，设定位置判别模式MOD，设置累加和SUM1～SUM6并清0，且每一个累加和对应六种加电方式中的一种加电方式；
    1.2)将无感无刷电机的转子实际位置POS定位到1；
    1.3)将指定幅度的短时电压脉冲依次按照六种加电方式对无感无刷电机加电，在不使用采样保持电路时所述短时电压脉冲的脉冲宽度ΔT0大于或等于采样时间ΔT，在使用采样保持电路时所述短时电压脉冲的脉冲宽度ΔT0小于采样时间ΔT，在加电后对无感无刷电机的电流取样电阻RSNS输出的时变电流取样信号进行高频滤波和放大，并在采样时间ΔT时刻对经高频滤波和放大后的时变电流取样信号进行AD采样，并将采样结果按加电方式分别累加至对应的累加和，且对每一种加电方式重复步骤1.3)共M次；
    1.4)将六个累加和SUM1～SUM6分别进行抽取，共得到六个抽取结果；所述抽取具体是指下述①～③三种方法中的一种：①、将累加和SUM1～SUM6分别右移lgN/lg4位后作为抽取结果S1～S6，其中N是4的正整数次幂；②、将累加和SUM1～SUM6分别除以M得到的平均值(保留足够的精度)后作为抽取结果S1～S6；③、将累加和SUM1～SUM6直接作为抽取结果S1～S6；
    1.5)判断位置判别模式MOD是否为1，如果是，则跳转执行步骤1.6)；如果不是，则跳转至步骤1.11)；
    1.6)设置用于位置识别的中间变量i为0；
    1.7)判断抽取结果S1大于抽取结果S4是否成立，如果成立则将中间变量i加1；
    1.8)判断抽取结果S2大于抽取结果S5是否成立，如果成立则将中间变量i加2；
    1.9)判断抽取结果S3大于抽取结果S6是否成立，如果成立则将中间变量i加4；
    1.10)根据中间变量i的值查找预设的位置查找表，得到转子识别位置pos，跳转至1.14)；
    1.11)判断抽取结果S1、S2、S3中S1是否最大，如果抽取结果S1是最大，再判断抽取结果S1大于S4是否成立，如果S1大于S4成立则令转子识别位置pos的值为1并跳转至步骤1.14)，如果S1大于S4不成立则令转子识别位置pos的值为4并跳转至步骤1.14)；如果抽取结果S1不是最大，执行步骤1.12)；
    1.12)判断抽取结果S1、S2、S3中S2是否最大，如果抽取结果S2是最大，再判断 抽取结果S2大于S5是否成立，如果抽取结果S2大于S5成立则令转子识别位置pos的值为2并跳转至步骤1.14)，如果抽取结果S2大于S5不成立则令转子识别位置pos的值为5并跳转至步骤1.14)；如果抽取结果S2不是最大，执行步骤1.13)；
    1.13)判断抽取结果S1、S2、S3中S3是否最大，如果抽取结果S3是最大，再判断抽取结果S3大于S6是否成立，如果成立则令转子识别位置pos的值为3并跳转至步骤1.14)，如果不成立则令转子识别位置pos的值为6并跳转至步骤1.14)；如果抽取结果S3不是最大，则增加循环变量M的值，重新设定用于统计正确率的测试次数CNT，跳转执行步骤1.2)；
    1.14)记录转子识别位置pos和无感无刷电机当前的转子实际位置POS；
    1.15)判断无感无刷电机的转子当前实际位置是否为6，如果不为6，预先将无感无刷电机的转子实际位置POS定位到下一位置，并跳转执行步骤1.3)；如果为6，则执行下一步；
    1.16)将测试次数CNT减1，判断新的测试次数CNT是否为0，如果新的测试次数CNT不为0，则跳转执行步骤1.2)；如果新的测试次数CNT为0，则跳转执行步骤1.17)；
    1.17)将每一条记录中转子识别位置pos和转子实际位置POS进行比较，在CNT次的测试中，统计出每次的六个位置识别全部正确的正确率C；判断正确率C小于识别转子初始位置时所能接受的最小正确率P是否成立，如果成立，则增加循环变量M的值，重新设定用于统计正确率的测试次数CNT，跳转执行步骤1.2)，如果不成立，则将当前的循环变量M的值作为六种加电方式共同的循环次数N的标定结果。
PCT/CN2016/079011 2015-04-16 2016-04-12 用于识别无感无刷电机初始位置的时变信号采样方法 WO2016165600A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201510180505.5A CN104767433B (zh) 2015-04-16 2015-04-16 用于识别无感无刷电机初始位置的时变信号采样方法
CN201510180505.5 2015-04-16

Publications (1)

Publication Number Publication Date
WO2016165600A1 true WO2016165600A1 (zh) 2016-10-20

Family

ID=53649099

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2016/079011 WO2016165600A1 (zh) 2015-04-16 2016-04-12 用于识别无感无刷电机初始位置的时变信号采样方法

Country Status (2)

Country Link
CN (1) CN104767433B (zh)
WO (1) WO2016165600A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110720174A (zh) * 2017-06-14 2020-01-21 罗伯特·博世有限公司 用于确定借助pwm控制的逆变器来供给的旋转的多相电机的相电流的方法
CN114089038A (zh) * 2021-11-16 2022-02-25 许昌许继软件技术有限公司 同步相量测量装置动态数据的时标秒位跳变处理方法及系统

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104767433B (zh) * 2015-04-16 2016-04-27 曾菊阳 用于识别无感无刷电机初始位置的时变信号采样方法
CN111134620A (zh) * 2019-12-30 2020-05-12 北京美丽年华文化有限公司 一种基于大数据技术的肤质检测系统
CN112290840A (zh) * 2020-11-04 2021-01-29 无锡同方聚能控制科技有限公司 一种直流无刷电机的转子初始位置检测系统

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2006094623A (ja) * 2004-09-24 2006-04-06 Hitachi Home & Life Solutions Inc 冷蔵庫用ｄｃインバータ駆動装置
CN103078586A (zh) * 2012-12-25 2013-05-01 南京航空航天大学 一种基于电感法的三相电励磁双凸极电机起动升速无位置技术
CN103618485A (zh) * 2013-11-14 2014-03-05 南京航空航天大学 一种无位置传感器无刷直流电机初始位置检测方法
CN104767433A (zh) * 2015-04-16 2015-07-08 曾菊阳 用于识别无感无刷电机初始位置的时变信号采样方法

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7592761B2 (en) * 2005-09-29 2009-09-22 Agile Systems Inc. System and method for starting and operating a motor
CN1333518C (zh) * 2005-12-30 2007-08-22 上海新时达电气有限公司 测量永磁同步电机的初始转子位置方法
CN102931903B (zh) * 2012-10-18 2015-09-16 青岛斑科变频技术有限公司 一种永磁同步电动机转子初始位置检测方法

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2006094623A (ja) * 2004-09-24 2006-04-06 Hitachi Home & Life Solutions Inc 冷蔵庫用ｄｃインバータ駆動装置
CN103078586A (zh) * 2012-12-25 2013-05-01 南京航空航天大学 一种基于电感法的三相电励磁双凸极电机起动升速无位置技术
CN103618485A (zh) * 2013-11-14 2014-03-05 南京航空航天大学 一种无位置传感器无刷直流电机初始位置检测方法
CN104767433A (zh) * 2015-04-16 2015-07-08 曾菊阳 用于识别无感无刷电机初始位置的时变信号采样方法

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110720174A (zh) * 2017-06-14 2020-01-21 罗伯特·博世有限公司 用于确定借助pwm控制的逆变器来供给的旋转的多相电机的相电流的方法
CN110720174B (zh) * 2017-06-14 2023-03-07 罗伯特·博世有限公司 用于确定借助pwm控制的逆变器来供给的旋转的多相电机的相电流的方法
CN114089038A (zh) * 2021-11-16 2022-02-25 许昌许继软件技术有限公司 同步相量测量装置动态数据的时标秒位跳变处理方法及系统
CN114089038B (zh) * 2021-11-16 2024-04-16 许昌许继软件技术有限公司 同步相量测量装置动态数据的时标秒位跳变处理方法及系统

Also Published As

Publication number Publication date
CN104767433A (zh) 2015-07-08
CN104767433B (zh) 2016-04-27

Similar Documents

Publication Publication Date Title
WO2016165600A1 (zh) 用于识别无感无刷电机初始位置的时变信号采样方法
EP3537675B1 (en) Method for demodulating an fsk signal of a pma standard wireless charging device
US11255696B2 (en) Bridge sensor biasing and readout
US7176677B2 (en) Stationary position detection circuit and motor drive circuit
CN103185837B (zh) 一种电力系统频率测量的方法
CN107425781B (zh) 一种基于线性磁链模型和线性回归分析的srm位置预估方法
CN211043427U (zh) 直流电机转速检测装置
CN109001629A (zh) 一种电机的反电动势测试方法
US11728755B2 (en) Method for determining a rotor position of an electric rotating machine, and an electric rotating machine for carrying out a method of this kind
US9966887B2 (en) BLDC zero crossing with BEMF, gating, and tri-state detect circuitry
CN105974256B (zh) 正余弦编码器断线检测方法
CN103718454B (zh) 永磁同步电动机的电动机常数计算方法以及电动机常数计算装置
EP3534528A1 (en) Motor control device and current ripple detection method for dc motor
Lee et al. Automatic power frequency rejection instrumentation for nonintrusive frequency signature tracking
CN104597276B (zh) 从信号中去除偏移的设备及其方法，磁性传感器单元
CN109782013B (zh) 电机测速方法、装置、可读存储介质及电动设备
US11398787B2 (en) Optimizing comparator operation for a varying DC component input
CN111989581B (zh) 电池监视系统
US7605569B2 (en) Acquisition circuit and controller circuit for an alternator
CN107482786B (zh) 一种无线充电系统负载估计方法
WO2022179265A1 (zh) 一种快速计算电机电参数的频率检测电路及其方法
JP3814153B2 (ja) 電力変換装置
CN209435138U (zh) 永磁同步电机的反电动势采样电路
CN106597067B (zh) 一种任意波形任意点的电压或电流测量装置及方法
JP5631170B2 (ja) 回転数計測装置

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: 16779575

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 16779575

Country of ref document: EP

Kind code of ref document: A1