CN1123965C - Production method of phase-based pulse width modulated sinusoidal voltage waveform data - Google Patents

Abstract

The invention relates to a method for generating waveform data of sinusoidal voltage pulse width modulation based on phase. Firstly, the frequency measurement parameter ( _NF ) and the product X× _NF are obtained, and the modulating wave waveform parameter and equal-amplitude ratio parameter are selected to calculate the modulation Wave variation, under the control of the calculation clock pulse signal, calculate the modulation wave waveform data according to the waveform parameters and the modulation wave variation, and at the same time count the pulses of the calculation clock pulse signal to obtain the phase count value, according to the reference wave phase parameters and phase The count value generates reference wave waveform data, compares the modulated wave waveform data with the reference wave data to obtain pulse level data, and the pulse level data and phase count value are combined to form pulse width modulated waveform data. The method of the present invention can realize the generation of sinusoidal voltage pulse width modulation waves under phase synchronous control, and the generated sinusoidal voltage pulse width modulation waves have good performance in terms of harmonics and symmetry of each phase, and can provide digital control means, control It looks intuitive and easy.

Description

A Phase-Based Generation Method of Sinusoidal Voltage Pulse Width Modulation Waveform Data

Technical field:

The invention relates to a method for generating waveform data of sinusoidal voltage pulse width modulation based on phase, and belongs to the technical field of power electronics in power system control.

Background technique:

The sinusoidal voltage pulse width modulated wave is a rectangular pulse signal with a certain amplitude but variable pulse width, which contains a main sinusoidal voltage fundamental component. The sinusoidal voltage pulse width modulation wave is widely used in the inverter device of the power system. Under the control of this signal, the power electronic device can be turned on and off to realize the conversion of electric energy from DC to AC. Referring to the book "Power Electronics Technology" with the international standard number 7111056175, the basic method of generating a sinusoidal voltage pulse width modulation wave is to compare the amplitude of the modulation wave and the reference wave, and output high and low levels according to the size of the result, so as to obtain a sinusoidal voltage pulse wide modulation wave. In the above method, the modulating wave generally adopts a triangular wave signal, while the reference wave can adopt a signal in the form of a sine wave or a trapezoidal wave. The modulating wave and reference wave are generated by their respective waveform generating parts, and the frequency and amplitude of the waveform generation are adjustable. The frequency F of the fundamental wave component of the sinusoidal voltage pulse width modulation wave is equal to the frequency of the reference wave, so the frequency of the fundamental wave can be controlled by adjusting the frequency of the reference wave. The frequency of the modulating wave can be fixed or adjustable, but the ratio of the frequency of the modulating wave to the frequency of the reference wave is an important parameter of sinusoidal voltage pulse width modulation, and this parameter is usually called the carrier ratio P. The carrier ratio P and the waveform relationship between the modulated wave and the reference wave will affect the high-order harmonic characteristics of the output pulse width modulated wave. On the other hand, if the amplitudes of the modulating wave and the reference wave are adjustable, the ratio M of the reference wave amplitude to the modulating wave amplitude can be changed. This ratio M is called the modulation ratio. When M<1, the fundamental sinusoidal component V of the sinusoidal voltage PWM wave is proportional to the modulation ratio M. In addition, constant V/F ratio control is widely used in motor speed control applications, that is, the fundamental component V of the sinusoidal voltage pulse width modulation wave is required to be adjusted in direct proportion to the fundamental frequency F. From this analysis, it can be seen that as long as the modulation ratio M is adjusted proportionally according to the fundamental frequency F, the V/F ratio constant control can be realized. References also give a software implementation method for generating sinusoidal voltage pulse-width modulated waves using microcontrollers, DSP, CPU and other micro-processing chips. This method is based on the basic principle of the aforementioned modulation wave and reference wave comparison method, through natural sampling or regular sampling or direct pulse width modulation and other approximate methods to derive the calculation formula of pulse width. The software implementation method is to obtain the sine wave by calculating these formulas. The pulse width data sequence of the voltage pulse width modulation wave, and control the holding time of the pulse output level according to this set of data, so as to obtain the pulse width modulation waveform.

There are some deficiencies in the above method for generating sinusoidal voltage pulse width modulated waves. First of all, these methods cannot know the phase angle of the transmitted pulse corresponding to the fundamental wave, and therefore cannot directly generate a pulse width modulated wave based on the phase angle. When these methods dynamically adjust parameters such as fundamental frequency F, modulation ratio M, and carrier ratio P, the integrity of a fundamental cycle sinusoidal voltage pulse width modulation wave will be destroyed, and the harmonic content will be increased. However, if the pulse is generated entirely based on the phase Avoid this situation. Furthermore, since the signals and data generated by the above methods do not contain phase angle information, it is not easy to realize pulse generation under phase synchronous control.

The analog circuit design and implementation of the above method is relatively complicated, and often cannot meet the control requirements of multiple performances. These control requirements include the adaptive adjustment of the carrier ratio P according to the fundamental frequency F, the symmetry control of the modulation wave waveform corresponding to the reference wave, the adjustment of the modulation ratio M, and the constant control of the V/F ratio. Moreover, the analog implementation method is cumbersome in operation control, poor intuition, the system is easily disturbed by temperature and noise, and has low accuracy and stability. For this reason, hope to adopt the digital way to produce the sinusoidal voltage pulse width modulation waveform. However, if the calculation and generation of pulse width modulation data is realized by software, the pulse width data needs to be continuously calculated according to the changing parameters such as fundamental frequency F, modulation ratio M, and carrier ratio P, so as to ensure continuous pulse generation. To provide the pulse width data sequence, the workload of program algorithm and execution is very large, and if the number of pulse output channels increases, the amount of calculation will increase exponentially. On the other hand, the software design of a control system will also include other control algorithms, so software programming must consider how to avoid program execution conflicts, and how to avoid operational delays and data loss. In fact, using software design to realize the generation of sinusoidal voltage PWM wave is also very complicated, and if the structure level is not clear enough and the program priority is not properly arranged during design, wrong waveforms will be generated and reliability will be reduced.

Invention content:

The object of the present invention is to propose a method for generating phase-based sinusoidal voltage pulse width modulation waveform data. The pulse width modulation waveform data generated by the method can not only reflect the pulse level change, but also provide the basic information corresponding to the pulse level change. Wave phase angle, based on these data and with the corresponding pulse generation method, the sinusoidal voltage pulse width modulation wave signal can be generated; based on this method, the adaptive adjustment of the carrier ratio P according to the fundamental frequency F can be realized, and the modulation wave waveform corresponding to Based on the symmetry control of the reference wave, the adjustment of the modulation ratio M and the constant control of the V/F ratio can be realized; based on this method, simple digital logic methods such as counting, addition, comparison logic and storage can be used to realize, thereby providing digital control means .

The generation method of the phase-based sinusoidal voltage pulse width modulation waveform data proposed by the present invention comprises the following steps:

1. Use the clock signal clk with a fixed frequency of F _clk to measure the frequency of the square wave pulse signal sig generated by the front-end frequency synthesis part to obtain the frequency measurement parameter _NF , or use the formula $N_f=\frac{f_{\mathrm{clk}}}{K\&times;f},$ Get N _F , and then calculate the product X×N F according to the control range of the frequency F of the fundamental component of the sinusoidal voltage pulse width modulation wave and the controllable parameter _X set by the modulation ratio M. The frequency of the above sig is the fundamental frequency K times of F, K=(0.01～10)×N _S , N _S is the number of phase sampling points selected in each sinusoidal voltage pulse width modulation wave fundamental period, and the frequency F _clk of the clock signal clk satisfies F _clk >K × F, F is the control upper limit of the fundamental frequency F.

2. Select the modulating wave waveform parameter and the equal-amplitude ratio parameter ER according to the parameter _NF or the fundamental frequency F, wherein the modulating wave waveform parameter includes the basic cycle point PL, the cycle point adjustment code RCODE and the length RCLEN of the cycle point adjustment code.

3. According to the controllable parameter Y set by the modulation ratio M, calculate the modulation wave variation GX according to the formula GX=Y×ER, or according to the data X× _NF obtained in the first step above, according to the formula GX=2 ^-k ×X×N _F ×ER, calculate the modulation wave variation GX, where k is the data bit adjustment parameter set by the output range of GX.

4. Under the control of the calculation clock pulse signal, according to the waveform parameters PL, RCLEN and RCODE obtained in the second step above and the modulation wave variation GX obtained in the third step above, the modulation wave is calculated according to the calculation method of the increase and decrease of the modulation wave amplitude Waveform data CD. The operation process is as follows:

① According to the adjustment code length RCLEN, the adjustment bit count pulse is counted circularly from 0 to RCLEN-1, and the adjustment bit number RN is obtained.

② Select the RNth bit data in the adjustment code RCODE according to the number of adjustment bits RN to output as the adjustment signal.

③According to the basic number of cycles PL, when the adjustment signal is invalid, select the number of cycle points L=PL; when the adjustment signal is valid, select the number of cycle points L=PL+1.

④According to the number of cycle points L, count and judge the calculation clock pulse signal in segments, and output a set of increase and decrease control signals with four states of 'increase', 'decrease', 'hold' and 'reset' and an adjustment bit count Pulse signal.

⑤Under the control of the calculation clock pulse signal, calculate the increase or decrease of the modulation wave amplitude according to the above-mentioned increase or decrease control signal and the modulation wave variation GX, and obtain the modulation wave waveform data CD.

5. Perform pulse counting on the calculation clock pulse signal to obtain a phase count value PH.

6. Based on the phase relationship between the sinusoidal voltage pulse width modulation waves of each channel, set the reference wave phase parameters of each phase, and then read the phase parameters of each phase from the reference wave waveform storage space according to the parameters and the above-mentioned phase count value PH. Reference wave waveform data RD ₁ ˜RD _i , where i is the phase number of the sinusoidal voltage pulse width modulation wave to be output.

7. Compare the modulated wave waveform data CD obtained in the above step 4 with the reference wave data RD ₁ ~ RD _i of each phase obtained in the above step 6 in turn, and compare the size with binary code, and combine the binary code values of each phase It is the pulse level data PB.

8. The combination of the pulse level data PB and the phase count value PH obtained in the above step 5 becomes the pulse width modulation waveform data, and finally outputs effective pulse width modulation waveform data, and the pulse width modulation of one phase point is processed each time After the waveform data, a calculation clock pulse signal is output, which is used to control the operation of the modulating wave waveform data CD in the above-mentioned step 4, and to perform pulse counting to generate phase counting in the above-mentioned step 5 Value PH.

The method for generating the phase-based sinusoidal voltage pulse width modulation waveform data proposed by the present invention, the output pulse width modulation waveform data can not only reflect the pulse level change, but also give the fundamental wave phase angle corresponding to the pulse level change, based on these The sinusoidal voltage pulse width modulation wave signal can be generated by combining the data with the corresponding pulse generation method, so that the generation of the sinusoidal voltage pulse width modulation wave under phase synchronous control can be easily realized. The method of the present invention can realize the adaptive adjustment of the carrier ratio P according to the fundamental frequency F, can realize the symmetry control of the modulation wave waveform corresponding to the reference wave, can also realize the adjustment of the modulation ratio M and the constant control of the V/F ratio, so the present invention The control function of the inventive method is complete, and the sinusoidal voltage pulse width modulation wave generated by obtaining data according to the present invention has good performance in terms of harmonics and symmetry of each phase. The method of the invention can be realized by using simple digital logic methods such as counting, addition, comparison logic and storage, so as to provide digital control means, and the control is intuitive and simple.

Description of drawings:

FIG. 1 is a functional block diagram of a method for generating waveform data of a phase-based sinusoidal voltage pulse width modulation according to the present invention.

Fig. 2 is a functional block diagram of the modulation wave amplitude increase and decrease operation method in the modulation wave waveform data generating part of the present invention.

Fig. 3 is a discrete waveform of a modulation cycle generated by the modulation wave amplitude increase and decrease operation method of the present invention when L=8, 9, 10 and 11.

Fig. 4 is a table of modulating wave waveform parameters according to an embodiment of the present invention.

Detailed ways:

Fig. 1 has provided the functional block diagram of the generation method of the sinusoidal voltage pulse width modulation waveform data based on the phase of the present invention, further in conjunction with this functional block diagram each step of the inventive method is described in detail as follows:

1. The signal sig input to the frequency measurement part is a square wave pulse signal whose frequency is K times the frequency F of the fundamental wave component of the sinusoidal voltage pulse width modulation wave. Take K=(0.01～10)×N _S , where N _S is the number of phase sampling points in each sinusoidal voltage PWM wave fundamental cycle, or N _S is the PWM wave data in each fundamental wave cycle Calculate the number of judgments, N _S is a positive integer, and the greater the N _{S is} , the higher the accuracy of the pulse data. There is a fixed clock signal clk in the frequency measurement part, its frequency is F _clk , and it should satisfy F _clk >K×F. The frequency measurement part counts the number of pulse counts of the clock signal clk in one sig signal period, and takes this count value as the output parameter _NF . Further, according to the controllable parameter X set by the control range of the fundamental frequency F and the modulation ratio M, count the number of pulses of the clock signal clk1 in X sig signal periods, so this count value is the output parameter X× _NF . The method of _NF and X× _NF thus obtained is equivalent to the following calculation formula: $N_f=\frac{f_{\mathrm{clk}}}{K\&times;f};$ $x\&times;N_f=\frac{x\&times;f_{\mathrm{clk}}}{K\&times;f}$ .

If the value of the fundamental frequency F is directly input into the system, the parameters _NF and X× _NF can also be obtained by calculation according to the above formula.

2. Select the modulating wave waveform parameter and the equal-amplitude ratio parameter ER according to the parameter _NF or the fundamental frequency F. The modulating wave waveform parameter is a set of data including the basic number of cycles PL, the adjustment code RCODE for the number of cycles, and the length RCLEN of the adjustment code for the number of cycles. . The basic point number PL of the cycle wave is the integer quotient of the carrier ratio P of the modulated wave divided by the phase sampling point _NS , and the carrier ratio P should be an integer satisfying 5<P<0.1× _NS . The cycle point number adjustment code RCODE is a binary data sequence, and its data bit length is RCLEN, and RCLEN takes a positive integer value satisfying RCLEN≤P. ER is an equal-amplitude ratio parameter, ER=q×P, where 0≤q≤1. PL, RCLEN and RCODE will be sent together as the waveform parameters of the modulating wave to the modulating wave waveform data generation part; and the equal amplitude ratio parameter ER will be sent to the modulating wave variation calculation part.

3. There are two optional working modes for the calculation part of the modulation wave variation: direct control mode and constant V/F ratio mode. In the direct control mode, the product of the controllable parameter Y set by the modulation ratio M and the equal-amplitude ratio parameter ER from the modulating wave waveform parameter part will be used as the modulating wave variation GX, namely:

GX=Y×ER;

In the mode of constant V/F ratio, the modulation wave variation GX is calculated according to the data X×N _F obtained in the first step and the constant amplitude ratio parameter ER according to the following formula:

GX＝2 ^-k ×X×N _F ×ER

The parameter k is a data bit adjustment parameter determined by the output value range of GX, and k is an integer.

4. The generation part of the adjustment wave waveform data is under the control of the calculation clock pulse signal. The basic point number PL of the cycle wave, the adjustment code RCODE, the adjustment code length RCLEN and the modulation wave variation GX are obtained by the modulation wave amplitude increase and decrease operation method to obtain the modulation wave waveform data CD , Figure 2 further shows the specific principle structure of the method.

As shown in Figure 2, the adjustment bit counting part counts the adjustment bit counting pulses output by the subsection counting and judging part. Counting starts from 0. The count value RN is input to the adjustment bit selection part, and the RNth bit data of the adjustment code RCODE is selected, and the adjustment bit selection part outputs an adjustment signal according to this data. Under the control of this adjustment signal, the cycle points selection part will select the cycle points L according to the parameter PL. When the adjustment signal is invalid, select the number of cycle points L=PL; when the adjustment signal is valid, select the number of cycle points L=PL+1.

Then, the subsection counting and judging part counts the calculation clock pulses and judges the subsection position of the current calculated phase point in one cycle of the modulation wave composed of L discrete points, thereby determining ', 'hold' and 'reset' four states of a group of increase and decrease control signal output, and generate an adjustment bit count pulse signal. The specific segmentation counting and judging methods are as follows: a) read in the parameter L, and divide L into two parts, the lower two digits LL and the remaining high digits LH; b) count the first calculation clock pulse, increase and decrease the control signal The output state is 'reset'; c) then count LH calculation clock pulses, the output state of the increase or decrease control signal is 'increase'; d) if LL is equal to 2 or 3, count 1 calculation clock pulse, and the output state of the increase or decrease control signal Otherwise, skip this step; e) Count LH counting clock pulses again, and the output state of the increase/decrease control signal is "decrease"; f) If LL is equal to 1 or 3, count 1 counting clock pulses, increase The output state of the decrease control signal is 'hold'; otherwise, skip this step; g) count LH counting clock pulses again, and the output state of the increase or decrease control signal is 'decrease'; h) if LL is equal to 2 or 3, count 1 Calculate the clock pulse, the output state of the increase or decrease control signal is 'hold'; otherwise, skip this step; i) count LH-1 calculation clock pulses at the end, and the output state of the increase or decrease control signal is 'increase'; j) in the aforementioned In the i-th step, when the last LH-1th pulse is counted, an effective adjustment bit counting pulse is sent out, and a new round of segmentation counting and judgment is started again in the a-th step.

The above segmented counting and judging method is equivalent to generating the output state of the increase and decrease control signal corresponding to each point on the discrete waveform of the modulation wave at point L of a cycle according to the remainder when L is divided by 4. An equivalent is given below describe:

(1) If L is an integer multiple of 4, corresponding to the first point of the modulation waveform, the increase and decrease control signal is 'reset'; $2~(\frac{L}{4}+1)$ point, the increase and decrease control signal is 'increase'; $(\frac{L}{4}+2)~(\frac{3L}{4}+1)$ point, the increase and decrease control signal is 'decrease'; $(\frac{3L}{4}+2)~L$ point, the increase and decrease control signal is 'increase';

点，增减控制信号为‘保持’；第 $\left(\frac{L+5}{2}\right)~\left(\frac{3L+5}{4}\right)$ 点，增减控制信号为‘减’；第 $\left(\frac{3L+9}{4}\right)~L$ 点，增减控制信号为‘增’；(2) If L is divided by 4 and the remainder is 1, then corresponding to the first point of the modulation waveform, the increase and decrease control signal is 'reset'; $2~\left(\frac{L+3}{4}\right)$ point, the increase and decrease control signal is 'increase'; $\left(\frac{L+7}{4}\right)~\left(\frac{L+1}{2}\right)$ point, the increase and decrease control signal is 'decrease';

point, the increase and decrease control signal is 'hold'; $\left(\frac{L+5}{2}\right)~\left(\frac{3L+5}{4}\right)$ point, the increase and decrease control signal is 'decrease'; $\left(\frac{3L+9}{4}\right)~L$ point, the increase and decrease control signal is 'increase';

点，增减控制信号为‘保持’；第 $\left(\frac{L+10}{4}\right)~\left(\frac{3L+2}{4}\right)$ 点，增减控制信号为‘减’；第 点，增减控制信号为‘保持’；第 $\left(\frac{3L+10}{4}\right)~L$ 点，增减控制信号为‘增’；(3) If L is divided by 4 and the remainder is 2, then corresponding to the first point of the modulation waveform, the increase and decrease control signal is 'reset'; $2~\left(\frac{L+2}{4}\right)$ point, the increase and decrease control signal is 'increase';

point, the increase and decrease control signal is 'hold'; $\left(\frac{L+10}{4}\right)~\left(\frac{3L+2}{4}\right)$ point, the increase and decrease control signal is 'decrease'; point, the increase and decrease control signal is 'hold'; $\left(\frac{3L+10}{4}\right)~L$ point, the increase and decrease control signal is 'increase';

点，增减控制信号为‘保持’；第 $\left(\frac{L+5}{2}\right)~\left(\frac{3L+3}{4}\right)$ 点，增减控制信号为‘减’；第

点，增减控制信号为‘保持’；第 $\left(\frac{3L+11}{4}\right)~L$ 点，增减控制信号为‘增’。(4) If L is divided by 4 and the remainder is 3, then corresponding to the first point of the modulation waveform, the increase and decrease control signal is 'reset'; $2~\left(\frac{L+1}{4}\right)$ point, the increase and decrease control signal is 'increase'; point, the increase and decrease control signal is 'hold'; $\left(\frac{L+9}{4}\right)~\left(\frac{L+1}{2}\right)$ point, the increase and decrease control signal is 'decrease';

point, the increase and decrease control signal is 'hold'; $\left(\frac{L+5}{2}\right)~\left(\frac{3L+3}{4}\right)$ point, the increase and decrease control signal is 'decrease';

point, the increase and decrease control signal is 'hold'; $\left(\frac{3L+11}{4}\right)~L$ point, the increase and decrease control signal is 'increase'.

And every time the last point of one cycle of the modulation wave is counted, an adjustment bit counting pulse is generated to the adjustment bit counting part to adjust the number of discrete points of the next modulation wave cycle.

Finally, under the control of the calculation clock pulse signal, the modulation wave amplitude increase and decrease part calculates the modulation wave waveform data CD according to the input increase and decrease control signal and the modulation wave variation parameter GX. When each calculation clock pulse arrives, the calculation rule of the modulation wave waveform is: when the increase or decrease control signal is 'reset', the modulation wave amplitude value CD is set to 0; when the increase or decrease control signal is 'hold', the modulation wave The wave amplitude value CD remains unchanged; when the increase/decrease control signal is 'increase', the modulation wave amplitude value CD increases by GX; when the increase/decrease control signal is 'decrease', the modulation wave amplitude value CD decreases by GX. Finally, the modulated wave waveform data CD obtained by the modulating wave waveform data generating part will be input to the data comparing part, as shown in FIG. 1 .

5. The phase angle counting part counts the calculation clock pulses to obtain the phase count value PH, and the counting range is from 0 to N _S -1, where N _S is the number of phase sampling points in each fundamental wave cycle. When the pulse width modulation waveform data generation part completes the calculation, judgment and storage of the pulse width modulation waveform data of a phase sampling point, it sends a calculation clock pulse to the phase angle counting part, so the phase count value PH increases by 1; when the PH Increase to N _S -1, and after the pulse width modulation waveform data generation part completes the processing of the last point of the pulse width modulation waveform data in a fundamental cycle, the phase count value PH is reset to 0.

6. Based on the phase relationship between the sinusoidal voltage pulse width modulation waves of each channel, set the reference wave phase parameters of each phase, and sum these parameters with the current phase count value PH, and the result is used as the reference wave phase before this phase , and then read out the reference waveform data RD ₁ ˜RD _i of each phase from the reference waveform storage space at the corresponding address, where i is the phase number of the reference waveforms with different phase relationships, and i is a fixed positive integer. The sampled waveform data of the reference wave is stored in the reference wave waveform storage space, and the phase angle corresponding to the reference wave waveform data of each sampling point is in a linear correspondence with the storage address of the data.

7. The data comparison part compares the reference wave waveform data RD ₁ ~ RD _i of each phase with the modulation wave waveform data CD, and performs binary coding according to the comparison result. If the coding results of each phase comparison are b ₁ , b ₂ , ..., _bi , then they are combined into pulse level data PB=b ₁ b ₂ ... _bi .

8. The pulse width modulation waveform data generation part combines the pulse level data PB and the current phase count value PH into pulse width modulation waveform data, and outputs valid pulse width modulation waveform data. In order to avoid data redundancy, it is only necessary to output valid PWM waveform data indicating that the pulse level changes, so the PWM waveform data is generally always output when the phase count value PH is equal to 0, when PH When it is not equal to 0, the pulse width modulation waveform data generation part will compare whether the pulse level data of the current point is the same as the pulse level data of the previous point. If not, it means that the pulse level output has changed. Point calculated PWM waveform data output. After processing the pulse width modulation waveform data of a phase point each time, the pulse width modulation waveform data generation part also outputs a calculation clock pulse signal to control the modulation wave waveform data generation part and the phase angle counting part.

Based on the sinusoidal voltage pulse width adjustment waveform data obtained by the method of the present invention, the sinusoidal voltage pulse width adjustment wave can be generated if matched with a corresponding pulse generation method.

The main feature of the method of the present invention is to start with the generation of the discrete waveform of the modulated wave, and provide a method for generating sinusoidal voltage pulse width modulation waveform data based on the phase angle. The generated pulse width modulation waveform data not only includes the change of the pulse level, The corresponding phase angles are also included.

The method of the invention can select the carrier ratio P according to the fundamental frequency F, so as to control the pulse frequency of the pulse width modulation wave. Because the inventive method selects different cycle basic points PL according to the size of the fundamental frequency F or the parameter _NF that is inversely proportional to F, and PL is equal to the quotient of the carrier ratio P divided by the phase sampling points _NS , so the value of PL directly reflects The size of the carrier ratio P in the sinusoidal voltage pulse width modulation determines the frequency of the modulating wave. The pulse frequency of the final pulse width modulation wave can be controlled by the selection of PL, so as to meet the requirements of different types of power electronic devices for driving pulse frequency. However, the number of phase sampling points N _S in a fundamental cycle may not be divisible by P, so the present invention further introduces two parameters of the cycle point adjustment code RCODE and the cycle point adjustment code length RCLEN to compensate the remainder, and through these two The selection of parameters can also satisfy the symmetry requirement of the modulating wave waveform relative to the fundamental wave cycle, which will be explained in the subsequent discussion.

First of all, it should be pointed out that the modulating wave amplitude increase and decrease calculation method adopted in the aforementioned modulating wave waveform data generation part is an algorithm rule for selecting the output state of the increasing or decreasing control signal according to the number of cycle points L of the modulating wave. This method makes the modulating wave A signal of one cycle can always satisfy the odd symmetry. Figure 3 shows the discrete waveform of one cycle of the modulation wave when L is 8, 9, 10, and 11 respectively. Regardless of the remainder of L divided by 4, the modulation waveform is odd and symmetrical. Further, if both the number of phase sampling points N _S and the carrier ratio P are integer multiples of 3, two possible situations can be discussed. The first case is when N _S can divide P exactly, and at this time N _S =P×PL, the modulation wave waveform in one fundamental wave cycle composed of P modulation cycles is oddly symmetrical. Furthermore, if the discrete modulation wave is shifted by 120° according to the phase angle of the fundamental wave, that is, it is shifted on the time axis according to the phase sampling point Point, the waveform obtained in this way will be exactly the same as the original modulation waveform, so when N _S can divide P, the modulation waveform also has three-phase symmetry. The second case is when N _S cannot divide P exactly, assuming NS=P×PL+R (0<R<P), in this way there will be PR modulation cycles containing PL sampling points in a fundamental wave cycle, and In addition, the R modulation cycles include PL+1 sampling points. If the two modulation cycles are evenly arranged according to certain rules, the odd symmetry and three-phase symmetry of the modulation wave of one fundamental cycle can be guaranteed. For this reason, the present invention introduces two parameters, RCLEN and RCODE, to control the arrangement of cycles with different numbers of points.

个PL+1点的周波在这 个周波组成的变化周期中两边对称的分布，即可满足整个基波周期内调制波的奇对称性。这样，周波的排布方式也就组成了调整码RCODE，而RCODE的长度为RCLEN。比如，当选择N_S＝3600，P＝21时，3600＝21×171+9，在一个基波周期中有21个调制周波，其中，12个周期有171个点，9个周期有172个点。要首先保证三相对称，则只要考虑在基波120°内的7个调制周波中4个171点周波(用0表示)和3个172点周波(用1表示)如何排列。采用两边对称的排法，比如排列为0011100、0101010或1001001，都可以实现调制波的奇对称性。如果选择0101010的排法，整个基波周期中调制波周波点数排列为：171，172，171，172，171，172，171；171，172，171，172，171，172，171；171，172，171，172，171，172，171。所以对于载波比N_S＝3600，P＝21的情况，周波点数调整码RCODE＝“0101010”，其长度RCLEN＝7。Since N _S =P×PL+R, if N _S and P are multiples of 3, R must also be a multiple of 3. Therefore, the design can guarantee that the modulated wave has at least A cycle is a change cycle, thus satisfying the three-phase symmetry, if

The cycle of a PL+1 point is here The symmetrical distribution on both sides of the change cycle consisting of two cycles can satisfy the odd symmetry of the modulation wave in the entire fundamental wave cycle. In this way, the arrangement of cycles constitutes the adjustment code RCODE, and the length of RCODE is RCLEN. For example, when N _S =3600, P=21, 3600=21×171+9, there are 21 modulation cycles in one fundamental cycle, of which, there are 171 points in 12 cycles, and 172 points in 9 cycles point. To first ensure the three-phase symmetry, it is only necessary to consider how to arrange four 171-point cycles (indicated by 0) and three 172-point cycles (indicated by 1) among the seven modulation cycles within 120° of the fundamental wave. The odd symmetry of the modulated wave can be achieved by adopting the symmetrical arrangement on both sides, such as 0011100, 0101010 or 1001001. If you choose the arrangement method of 0101010, the number of modulation wave points in the entire fundamental wave cycle is arranged as follows: 171, 172, 171, 172, 171, 172, 171; 171, 172, 171, 172, 171, 172, 171; 171, 172 , 171, 172, 171, 172, 171. Therefore, for the case where the carrier ratio N _S =3600 and P=21, the cycle point number adjustment code RCODE=“0101010” and its length RCLEN=7.

If the modulation wave has odd symmetry, it can be basically guaranteed that the pulse width modulation wave does not contain a DC component; if the modulation wave of a fundamental cycle is symmetrical in three phases, when generating a three-phase sinusoidal voltage pulse width modulation wave with a mutual difference of 120°, it can be Ensure the three-phase symmetry of the output voltage. Therefore, the method of the present invention can satisfy the symmetrical performance requirement of the sinusoidal voltage pulse width modulation wave by controlling the parameters RCLEN and RCODE.

When the carrier ratio P is changed according to the fundamental frequency F, the amplitude A _Δ of the modulating wave should not change accordingly. For this reason, corresponding to different carrier ratios P, the modulating wave waveform parameter part also outputs an equal-amplitude ratio control parameter ER, and ER is proportional to P, that is:

ER=q×P;

where q is a scaling factor. When the modulation wave variation calculation part adopts the direct control method to generate the modulation wave variation GX, GX=Y×ER, after approximate derivation, the maximum amplitude A _Δ of the modulation wave can be expressed as: $A_{\mathrm{\&Delta;}}=\frac{N_{\mathrm{the\; s}}\&times;q}{4}\&times;Y;$

Therefore, when N _S and q are fixed, the maximum amplitude A _Δ of the modulated wave is only controlled by the parameter Y, but not affected by the carrier ratio P.

If the reference amplitude is a constant A _ref , the formula for the modulation ratio M can be derived: $m=\frac{A_{\mathrm{ref}}}{A_{\mathrm{\&Delta;}}}=\frac{4\&times;A_{\mathrm{ref}}}{N_{\mathrm{the\; s}}\&times;q\&times;Y}$ .

The above formula shows that the method of the present invention can realize the adjustment of the modulation ratio M in the pulse width modulation process through the control of the input parameter Y.

When the V/F ratio constant control method is used to generate GX in the calculation part of the modulation wave variation, GX=2 ^-k ×X×N _F ×ER, then the maximum amplitude A _Δ of the modulation wave is:

A _Δ ＝2 ^-k-2 ×N _S ×q×X×N _F ;

This formula also shows that after the parameter ER is introduced, the maximum amplitude A _Δ of the modulated wave is not affected by the carrier ratio P. also because $N_f=\frac{f_{\mathrm{clk}}}{K\&times;f},$ The modulation ratio M under constant V/F ratio control mode is: $m=\frac{2^{k+2}\&times;K\&times;A_{\mathrm{ref}}}{N_{\mathrm{the\; s}}\&times;q\&times;x\&times;f_{\mathrm{clk}}}\&times;f;$

From the above formula, when the parameters k, K, A _ref , N, q, X and F _clk are fixed, the modulation ratio M and the fundamental frequency F are proportional. And because when M<1, the fundamental component V of the sinusoidal voltage pulse width modulation wave is proportional to the modulation ratio M, so the fundamental component V is also proportional to the fundamental frequency F. Therefore, this control method can realize the control with a constant ratio of V and F, thereby providing sinusoidal voltage pulse width modulation waveform data with a constant V/F ratio for applications such as motor control.

The method for generating sinusoidal voltage pulse width modulation waveform data provided by the invention can not only be realized by program software, but also provides a feasible solution for the realization of digital logic. As previously mentioned, the multiplication operation of calculating X×N _F can be converted into counting the number of pulses of the clock signal clk in X sig signal periods; It is equal to Y in the direct control mode, and equal to 2 ^-k ×XN×N _F in the constant V/F ratio control mode), which can be realized by directly adopting multiplication logic, or converting this multiplication into addition. If the ER data is expressed as e _j-1 e _j-2 …e ₀ in binary, then the calculation of GX can be decomposed into the following formula:

D×ER＝e _j-1 ×D×2 ^j-1 +e _j-2 ×D×2 ^j-2 +…+e ₀ ×D×2 ⁰

In this way, GX can be obtained by shifting and j-1 addition operations. Therefore, based on the method of the present invention, digital logic such as counting, addition, comparison and storage space can be used to realize the generation of sinusoidal voltage pulse width modulation waveform data.

Finally, the method of the invention only needs to control the frequency of the input signal sig, adjust the parameters X, Y and the phase parameters of reference waves of each phase to realize the calculation and generation of sinusoidal voltage pulse width modulation waveform data, with few control parameters and clear meaning. Therefore, in the power electronic control system, the method of the invention can be used to realize the functional sub-module that generates sinusoidal voltage pulse width modulation waveform data based on the phase. The module has independent and complete functions, easy to adjust and control, and has good stability and reliability.

Implement the present invention according to the principle structure as shown in Figure 1, the range of the fundamental wave frequency F corresponding to the generated sinusoidal voltage pulse width modulation waveform data is 10Hz～138Hz, and the number of phase sampling points in one fundamental wave period is N _S =3600 .

1. The pulse interval time of the signal sig input to the frequency measurement part corresponds to the 0.1 degree phase time of the fundamental wave, so the frequency of sig is 3600 times of the fundamental wave frequency, that is, K=3600. The frequency of the frequency measurement clock signal clk is F _clk =8×10 ⁶ Hz, count the number of pulses of the clock signal clk in the pulse interval of one signal sig, and take this number as the parameter _NF . The parameter X can be set through a storage space, and the number of pulses of the clock signal clk in the X pulse intervals of the signal sig is counted, and this number is regarded as X× _NF .

2. According to the parameter table shown in Fig. 4, select modulating wave waveform parameters PL, RCODE, RCLEN and equal amplitude ratio parameter ER according to the input _NF . In the table of FIG. 4 , the first column is the input parameter _NF , and it is segmented according to the value of _NF . According to the fundamental frequency range, it can be known that: _16≤NF ≤255. The 4th, 5th, 6th, and 7th columns give the data corresponding to the output parameters PL, RCODE, RCLEN and ER selected by each numerical value segment _{of NF} , and the 2nd and 3rd columns in the table give the data corresponding to each numerical value _{of NF} The carrier ratio P and P of the segment are divided by the number of phase sampling points N _S and the remainder value R is the size. P and R are not used as input and output data, and are given here for reference. It can be seen from the table that the relationship between the equal amplitude ratio ER and P is $\mathrm{ER}=\frac{1}{3}\&times;P,$ the coefficient of proportionality $q=\frac{1}{3}$ .

3. The calculation of the modulation wave variation GX is controlled by a state control bit data vfmode. When vfmode=0, select the direct control mode, that is, GX=Y×ER; when vfmode=1, select V/F ratio constant control way, that is, GX=2 ^-k ×X× _NF ×ER, where k=8. Since ER is a positive integer with a maximum value of 56, ER is expressed as ER=e ₅ e ₄ …e ₀ by using a 6-bit binary number, then:

D×ER＝e ₅ ×D×2 ⁵ +e ₄ ×D×2 ⁴ +…+e ₀ ×D×2 ⁰ ;

GX is calculated by 5 additions, and only the integer part of the GX data is kept at the end.

4. The modulating wave waveform data generation part is implemented according to the principle structure shown in Fig. 2, and the modulating wave waveform data CD is generated under the control of the calculation clock pulse signal. The amplitude A _Δ of the modulated wave obtained is:

5. Every time the pulse width modulation waveform data generation part calculates and stores the pulse width modulation waveform data of one phase point, it sends a calculation clock pulse to the phase angle counting part, so the phase count value PH increases by 1 from 0 to 3599; If the phase count value PH increases to 3599, and after another calculation clock pulse is received, the phase count value PH is reset to 0.

6. The reference wave stored in the reference wave waveform storage space is a sine wave, and its maximum amplitude is A _ref =2 ¹⁰ , which is sampled once every 0.1 degrees. The storage address ranges from 0 to 3599, and the address is j (j=1 , ..., 3599) stored data corresponds to reference wave waveform data whose phase angle is (0.1×j) degrees. The reference wave waveform data generation part outputs three-phase reference wave waveform data with a mutual phase difference of 120°, and the reference wave phase parameters corresponding to the three phases are 0, 1200 and 2400 respectively. If ST is used to represent the reference wave phase parameter of a certain phase, that is, ST is equal to 0 or 1200 or 2400. When the pulse width modulation wave level of the sampling point whose phase count value is PH is calculated, the phase reference wave outputs the reference wave waveform data whose phase angle is 0.1×(ST+PH) degrees, so if ST+PH<3600, Then read the phase reference waveform data from the storage address (ST+PH); if ST+PH≥3600, then read the reference waveform data from the storage address (ST+PHASE-3600) . In this way, at each phase sampling point, three reference waveform data RD ₁ -RD ₃ respectively corresponding to the three phases are read out.

8. The data comparison part compares the magnitudes of the modulation wave waveform data CD and the three-phase reference wave waveform data RD ₁ -RD ₃ . If CD>RD _i (i=1,2,3), then the pulse level data output by this phase is b _i =0 (i=1,2,3); if CD≤RD _i (i=1,2 , 3), then the pulse level data output by this phase is b _i =1. Finally, b ₁ , b ₂ and b ₃ are combined into pulse level data PB=b ₁ b ₂ b ₃ .

9. In the pulse width modulation waveform data generation part, the 3-bit binary data PB and the 13-bit binary data PH are combined into 16-bit pulse width modulation waveform data. When the phase PH is equal to 0, the data is stored in the storage space. When PH is not equal to 0, the PWM waveform data generation part will compare whether the current PB data is the same as the PB data calculated at the previous point, and if not, store the PWM waveform data at this point into the waveform data storage space.

Thus, the modulation ratio M of the sinusoidal voltage PWM wave represented by the waveform data generated in this embodiment is:

If vfmode=0 is set, that is, the direct control mode is selected, and the modulation ratio M can be directly controlled by modifying the parameter Y. If you set vfmode=1, that is, choose the constant V/F ratio control mode, you can adjust the proportional coefficient of fundamental frequency F and modulation ratio M by setting parameter X. When X=20 and the fundamental frequencies F are 10Hz, 48Hz and 60Hz respectively, the modulation ratio M calculated according to the above formula is correspondingly:

When F=10Hz, M=0.197;

When F=48Hz, M=0.944;

When F=60Hz, M=1.180.

Based on this embodiment, the generation of sinusoidal voltage pulse width modulation waveform data can be correctly implemented.

Claims (1)

1. A method for generating phase-based sinusoidal voltage pulse width modulation waveform data, characterized in that the method may further comprise the steps: (1) Use the clock signal (clk) with a fixed frequency of F _clk to measure the frequency of the square wave pulse signal (sig) generated by the front-end frequency synthesis part, and obtain the frequency measurement parameter (N _F ), or use the formula $N_f=\frac{f_{\mathrm{clk}}}{K\&times;f},$ Get the frequency measurement parameter (N _F ), and then calculate the product X× according to the control range of the frequency (F) of the fundamental component of the sinusoidal voltage pulse width modulation wave and the controllable parameter (X) set by the modulation ratio (M) N _F , the frequency of the above-mentioned square wave signal (sig) is K times the fundamental frequency (F), K=(0.01～10)×N _S , N _S is selected from the fundamental wave cycle of each sinusoidal voltage pulse width modulation wave Fixed phase sampling points, the frequency (F _clk ) of the clock signal (clk) satisfies F _clk > K × F, and F is the control upper limit of the fundamental frequency F; (2) Select the modulating wave waveform parameters and equal amplitude ratio parameters (ER) according to the frequency measurement parameters ( _NF ) or the fundamental frequency (F), where the modulating wave waveform parameters include the basic points of the cycle (PL), the adjustment code of the points of the cycle (RCODE) and the length (RCLEN) of cycle points adjustment code; (3) According to the controllable parameter (Y) set by the modulation ratio (M), according to the formula GX=Y×ER, calculate the modulation wave variation (GX), or according to the data X obtained in the above step (1) ×N _F , according to the formula GX=2 ^-k ×X×N _F ×ER, calculate the modulating wave variation (GX), where k is the data bit adjustment set by the output range of the modulating wave variation (GX) parameter; (4) Under the control of the calculation clock pulse signal, according to the waveform parameters (PL, RCLEN and RCODE) obtained in the above step (2) and the modulation wave variation (GX) obtained in the above step (3), the modulation wave is obtained by calculation Waveform data (CD), the operation process is as follows: ①According to the adjustment code length (RCLEN), the adjustment bit count pulse is counted circularly from 0 to RCLEN-1 to obtain the adjustment number of bits (RN), ② According to the number of adjustment bits (RN), select the data of the RNth bit in the adjustment code (RCODE) as the adjustment signal output, ③Based on the basic points of the cycle (PL), when the adjustment signal is invalid, select the cycle point L=PL; when the adjustment signal is valid, select the cycle point L=PL+1, ④ According to the number of cycle points (L), count and judge the calculation clock pulse signal in sections, and output a set of increase and decrease control signals with four states of 'increase', 'decrease', 'hold' and 'reset' and an adjustment Bit counting pulse signal; ⑤Under the control of the calculation clock pulse signal, calculate the increase or decrease of the modulation wave amplitude according to the above-mentioned increase or decrease control signal and the modulation wave variation (GX), and obtain the modulation wave waveform data (CD); (5) Carry out pulse counting to calculation clock pulse signal, obtain phase count value (PH); (6) According to the phase relationship between the sinusoidal voltage pulse width modulation waves of each channel, set the reference wave phase parameters of each phase, and then read from the reference wave waveform storage space according to the parameters and the above-mentioned phase count value (PH) Reference waveform data of each phase (RD ₁ ~RD _i ); (7) Compare the modulated wave waveform data (CD) obtained in the above step (4) with the reference wave data (RD ₁ ~ RD _i ) of each phase obtained in the above step (6) in sequence, and the comparison size is binary coded, And combine the binary coded value of each phase into pulse level data (PB); (8) Combining the pulse level data (PB) and the phase count value (PH) obtained in the above step (5) to become pulse width modulation waveform data, and finally output effective pulse width modulation waveform data. After the pulse width modulation waveform data of a phase point, a calculation clock pulse signal is output, and the calculation clock pulse signal is used to control the operation of the modulated wave waveform data (CD) in the above-mentioned (4) step, and to be used in the above-mentioned first step (5) In the step, count the pulses to generate the phase count value (PH).

Images