SU796881A1 - Random process simulator - Google Patents

Description

one

The invention relates to computing control technology and is intended to reproduce continuous random processes with controllable instantaneous value distribution function and power spectral density. The invention can be used to build simulators and is especially effective when used as a simulator of random effects during vibration and shock tests on electrodynamic stands. A simulator of random processes is known, which contains an adder, amplifiers and filters that cause the complexity of the simulator setup. J

The closest technical solution to this is a random process simulator containing a source of reference voltages, the first and second transducers in series the code-voltage and polarity modulator, the output of which is connected to the output of the simulator, the pulse counter whose input is connected to the output the first pulse generator, serially connected to the second pulse generator, the first frequency divider and the first reversible counter, the output of which is connected to the control input of the second converter code-voltage gate, the first register, the output of which is connected to the control inputs of the first code-voltage converter and the polarity modulator, the second register, the output of which is connected to the control input of the first frequency divider, random number sensor, the input of which is connected with the output of the memory unit, the control unit, the outputs of which are connected to the control inputs of the memory unit, random number generator, pulse counter, first rever, solid counter, first and second registers 2 and 3, respectively.

0

The disadvantage of the generator is that it allows to obtain a limited class of spectral densities of random processes.

The purpose of the invention is the expansion

5 areas of application of the simulator.

To achieve this goal, the simulator contains in series between the input of the simulator and the input of the memory unit a statistical

0 analyzers and a comparison unit, a third pulse generator, a second frequency divider, a second reversible counter and a functional converter, connected in series between the source voltage of the reference voltage source and the input of the first code voltage converter, the third and fourth code voltage converters, control The input input of which is connected to the output of the function converter, the third register, the output of which is connected to the control input of the third code-voltage converter, and in d - to the corresponding output of the control unit the fourth register, the output of which is connected to the control input of the second frequency divider, and input - to the corresponding output of the control unit, the other output of which is connected to the control input of the second reversible counter, the other output of which is connected to the information input of the module polarity torus, the output of the random number sensor is connected to the setup inputs of all registers and a pulse counter ..

Fig. 1 shows a block diagram of a simulator: Fig. 2 shows a spectral power density for a random process that is a sequence of triangular isosceles with different polarities with random parameters; FIG. 3 shows the spectral power density of a typical triangular pulse; 4 is an approximation of an arbitrary spectral density. power by superposition of power spectral densities for a triangular pulse.

The simulator contains a block of 1 memory, a sensor of 2 random numbers, a block of 3 controls, a first generator of 4 pulses, a counter of 5 pulses, a source b of reference voltages, a first -7 and a second 8 code-voltage converters modulator 9 polarity, first register 10, second pulse generator 11, first frequency divider 12, second register 13, first reversing counter 14, statistical analyzer 15, comparison unit 16, third pulse generator 17, second frequency divider 18, second reversing counter 19, functional converter 20, third 21 and quarters There are 22 code-voltage converters, third 23 and fourth 24 registers.

Blocks 15, 16, 1, 2, 3, 5, and 4 are interconnected in series, in addition, memory block 1 and control block 3 are also interconnected. Blocks 6, 21, 22, 7, 8, and 9 are connected in series. The output of the polarity modulator 9 is the output of the simulator. Simulator input

connected to the input of the statistical analyzer 15. The output of the sensor 2 random numbers is connected to the set inputs of registers 10, 13, and 24, the counter 5., and the second reversible counter 19. The outputs of the first register 10 are connected to the control inputs of the first converter 7 code-voltage and modulator 9 polarity. Blocks 11, 12, and 14 are interconnected in series, with the outputs of the second register 13 being connected to the control inputs of the first frequency divider 12. The outputs of the first reversing counter 14 are connected to the control inputs of the second converter 8 code-voltage converter. Blocks 17-20 are connected between themselves in series, while the outputs of the functional converter 20 are connected to the control inputs of the fourth code converter 22, and the outputs of the fourth register 24 are connected to the control inputs of the second frequency divider 18. The inputs of the third voltage / voltage converter 21 are connected to the outputs of the third register 23, the first input of which is connected to the corresponding output of the control unit and the second input to the output of the sensor

2 random numbers. The second output of the second reversible counter 19 is connected to the input of the modulator 9 of polarity. To the control inputs of the registers 10, 13, 23 and 24 and the reversible counters 14 and 19 are connected the corresponding outputs of the control unit 3.

FIG. 2, 3, and 4; the following notation is used: (ly and С (ш) - frequency and spectral power density, respectively; 1 - given function G (ОУ); 2 - reproducible function G (и); 3 - shifted the spectral power density of a typical triangular pulse; WK, the filling frequency of a typical triangular pulse of harmonic oscillations; G () are the values of a given power spectral density at the points and O.

Imagine the time function Q (t), which describes the output signal of the simulator, in the form of a product of five functions, namely

q (t)) U2 (t) u ,, a) U4WU5 (t), (D

where the functions on the right-hand side of (1) are the transfer functions of the corresponding code-voltage converters and modulators 9 in terms of polarity.

Consider the mode of operation of the simulator at UT, (t) -U4 (t) 1. This mode is provided by the appropriate control signals from the unit

3 controls to the third input of the second reversible counter 19 and the first input of the third register 23, so that the second reversible counter 19 and the third register 23 / and therefore are functional. The converter 20, the third 21 and the fourth 22 code-voltage converters are set to a fixed link that determines the transfer of the output potential of the source b of reference voltages to the input of the first code-converter 7 without large-scale conversion. In this mode, the simulator generates a stream of bipolar pulses of a triangular isosceles shape, in which the amplitude A of the apex, the duration t7 of the base and the duration T - of the interval between the pulses are independent random variables distributed according to the corresponding laws F (A), F (tr) and F (t). The polarity sign also varies randomly, with the likelihood of positive and negative polarity pulses being determined by the likelihood of values falling into the corresponding, subdivisions of the function definition. Taking into account the above, the functions U (t),) and Uy (t) can be written as follows U (-t) A. those. and (t) takes a random value for each pulse; O. 2 (i-), i.e. 1) 2 (1) describes a triangular pulse of unit amplitude and random duration; those. ) for each pulse it is random, but also constant over time C value. Taking into account (2) - (4), we have Q (t, (-). Before starting the simulator operation. In memory block 1, sequences of control codes are inserted that ensure the generation of random parameters during the operation of sensor 2 random numbers A, T and T are distributed according to the relevant laws F (A}, FCC) and F (T). Each cycle of the simulator begins with the fact that for the next pulse of the output stream its group of specific values of the scientific research institute A, f and T are generated. This sensor 2. random numbers are sequentially connected by control unit 3 to the corresponding areas memory unit 1, namely, the code storage area for setting A, then T, and finally, the first of this random number group determining the amplitude A of the generated pulse, is entered into the first register 10 by the command from control unit 3 into the first register 10. The second random number setting the duration of the next random interval T between pulses is rewritten by the control unit 3 from the outputs of the random number sensor 2 to the pulse counter 5. The third random number, corresponding to the duration f of the next pulse, is entered by the command from the control unit 3 and in the second register 13. At the output of the source b of the reference voltages, a voltage level equal to the maximum amplitude value of the generated pulses transmitted by the third 21 and fourth 22 code-voltage converters without changing to the input of the first code-voltage converter is set. . In accordance with the specific values of the random codes recorded in the first register 10, voltage levels in the range from zero to the maximum value set at the input of the source 6 of the reference voltages can be generated at the output of the first code-voltage converter 7. In the initial state, in the reverse counter 14, zeros are written. When the reversible counter 14 is filled from zero to the maximum value at the output of the second converter 8 code-voltage, the leading edge of a triangular pulse is formed. Since the reversible counter 14 is filled with a constant (for a particular C value) input pulse frequency, the increments of the output voltage of the second code-voltage converter 8 are constant, and, therefore, the front font of the generated pulse has the form of an inclined straight line. only in the reversible counter 14 all units will be recorded, the counter is included in the countdown mode. At this moment, the formation of the leading edge of the output pulse ends, and the voltage at the output of the second converter 8, the code voltage, is equal to its input voltage, i.e., equal to the specific value of the amplitude A of the pulse generated at the output

the first converter 7 code-voltage. Since the counting goes with the same frequency as the forward one, the length of the falling front is equal to the length of the leading edge, and the output pulse has the form of an isosceles triangle.

The modulator 9 of the polarity of the signal transmits a pulse from the output of the second converter B to the output of the simulator, either keeping its polarity or changing it in accordance with the specific amplitude code, A of the pulse recorded in the first register 10. For example, if a particular coded value amplitude A corresponds to the actual amplitude A entering in the year of the negative range of the function definition area F (A), then the signal polarity modulator 9 recognizes this code and changes the polarity of the signal generated at the output of the second reobrazovatel-8 code voltage. The polarity modulator 9 can be implemented using known digital comparison circuits, threshold circuits, and operational amplifiers of direct current. Since the actual output pulses from the second code-voltage converter 8 have a stepwise triangular isosceles shape due to the discreteness of the pulse amplitude increments, the modulator 9 of the signal polarity smoothes the edges of the pulse using known filtering schemes.

Using a random number written to the second register 13, the corresponding conversion factor of the frequency divider 12 is set. This means that for each valid value of duration: 1 pulse, the reversible counter 14 is filled with the corresponding frequency obtained by dividing the frequency of the pulsed flow received through the frequency divider 12 by the reversing counter 14 from the second pulse generator 11. Such control of the filling rate of the reversible counter 14 is necessary so that at any valid value of the pulse duration C, the reversible counter 14 has time to replicate the double cycle - from zero to the maximum value and vice versa, which ensures the formation of a strictly symmetric (equilateral triangular pulse shape for any combination of specific values of the duration C of the pulse and its amplitude A, i.e., at any possible inclinations of the fronts of the pulses.

Using the first pulse generator 4 and the counter 5, a random code sweep recorded in the counter 5 is carried out in the time interval T between the given and subsequent pulses. At the end of the interval T, the control unit 3 ensures the implementation of a new cycle of operation of the random process simulator in the mode of forming triangular-shaped pulses with random parameters.

It can be shown that the spectral power density of a random process in this mode is determined by the expression

/ SW (t Uil4) 4

(u; b / 4k (Mi,

(6)

h

where D- is the process dispersion;

TQ is the expectation of the duration of the interval between pulses;

fcCt) is the density distribution of the pulse duration; the expression in brackets raised to the power is the square of the Fourier spectrum of a typical triangular pulse.

The shape of the curve 6 (III) is shown in Figure 2 for the real frequency range.

With f-r (t) cG (t-to), i.e. for a process with pulses of constant duration, the spectral power density takes the form

/ 5-1п (ШГо / 4) 4

(7) V OU-CO / 4 /

The shape of the curve G (cf) is shown in FIG.

The second mode of operation of the random process simulator makes it possible to implement the well-known principle of transferring the spectral characteristics of a typical pulse to the high-frequency region by multiplying the typical signal by a harmonic oscillation. Consider this mode with non-random parameters. ,

T-C-Oonst CoU ,, tt) - VK

(eight)

U Ct) - cos N (-1

where Bk is the amplitude of the k-harmonic.

The formation of 114 (1;) according to (8) easily and with any desired accuracy can be carried out using a digital functional converter 20. The positive harmonic half-period (8) is quantized at a value of A) to 1 per m clock cycles and the value), l, 2, ..., (m -1.) is represented at the output of the functional converter by a 20 n-bit code. By choosing the values of m and n, it is possible to provide the representation (8) fc with any desired accuracy. The value m determines the size of the S of the second reversible counter 19, while S 1 oq2ni .. The negative half-period is formed by changing the polarity of the output signal of the simulator by supplying a control signal from the second output and the second reversing counter 19 to the input of the modulator 9 field rnness. The pulses from the third generator 17 are supplied through the second frequency divider 18 to the counting input of the second reversible counter 19, the output of which is converted by the functional converter 20 into codes, controlling the output voltage of the fourth code-voltage converter 22 and ensuring reception of the signal 04 (1). The change in frequency Ul is provided by determining the required coefficient of recalculation of the second frequency divider 18 using a code generated by the sensor 2 random numbers. The set of amplitudes BK and frequencies tsu,, 2, ... g is determined by calculation from a given spectral power density (curve 1 in Fig. 4) and the known spectral power of a typical triangular pulse (Fig. 3), Codes amplitudes Hz and frequencies 4 appear at the output of the sensor 2 random numbers with probabilities q (k) specified according to some distribution law F ((k)) the values of q (k) are recorded in memory block 1 before the simulator starts working Before the formation of the next pulse, the codes of the amplitudes B ,, and 4acTOi 1% are shifted from the output of the face 2 cl tea numbers of numbers, respectively, into the third 23 and fourth 24 registers from the signals from control unit 3, go to the corresponding control inputs of these registers. Thus, the output signal in the second mode is described by a time function (IA., -. , jj. $. Qi ((t) U4 (-t)) - B ,, c05 (2 (2ACl- | -), fe-b- i Set m e / const 0. This signal corresponds to the spectral power density, is described by the formula. (UJ-uV) G, 2 (“)) - 2B qtic: I (i-Schc (cm-fUV)„. 4 0. Expression (10) describes the superposition of the spectral power densities of a typical pulse, triangular in shape (curves 3 in Fig. 4), shifted along the frequency axis by frequencies OU and multiplied by weighting factors proportional to the values of G (tf). Reproducible according to (10) the spectral power density (curve 2 in figure 4) approximates a given function of the spectral power density (curve 1 in figure 4) of an arbitrary type with any desired degree of accuracy. To ensure equal distribution of the random phase of the signal (8), the random number sensor 2 absorbs uniformly random codes recorded before starting the formation of the next triangular pulse to the second reversible counter 19 according to the signal of the control unit 3 fed to the third (control) input of the second reversible counter 19. Thus, in this juvitator, by simple means, it is possible to expand its functionality ayy process with any required spectral power density. This allows the use of a simulator to solve a wider range of technical problems, for example, in vibratory tests. Using a statistical analyzer 15 and a comparison unit 16, a random process simulator automatically maintains the required characteristics, such as power spectral density, at the output of a dynamic object, such as an electrodynamic vibration stand, to the input of which the process is fed from the output of the simulator and from the output of special sensors the sigad supplied to the input of the simulator and the statistical analyzer 15. In this case, the statistical analyzer 15 calculates the estimates of the laws of and dividing the power spectral density of the input process and the comparison unit 16 compares the estimates with the set of parameters and carries out correction, incoming into the expression (10) for detecting deviations from predetermined statistical estimates. The described mode can be Hcnojn designed to automatically set up a simulator of random effects on the reproduction of a random process, adequate to the recorded process under operating conditions, and applied when tuning to the input of the simulator. This mode provides a high degree of automation of shock, vibration and other types of tests. The same quality ensures complete autonomy of this simulator does not require preliminary processing of information on computers to configure it, which opens up the possibility of widespread implementation of the simulator not only at large test centers equipped with computers, but and in the plants conducting the above tests.

Of particular note is the simplicity of the simulator design and the manufacturability of its manufacture on the basis of standard and unified blocks and elements of digital integrated circuitry, which improves the reliability of the simulator in comparison with the known devices of similar purpose and reduces its size and weight and power consumption, which satisfies one of the essential requirements for laboratory autonomous instruments.

The technical and economic efficiency of the implementation of the invention is determined by the following factors: the simplification and cheapening of test control equipment; increasing the reliability of tests and accuracy of reproduction of specified characteristics; ensuring high. stability of long-term tests; simplifying and automating the process of setting up the simulator to reproduce random processes with the required laws of the distribution of instantaneous values and (or) power spectral density.

Claims (3)

1. USSR author's certificate No. 3963631, cl. G 06 M 7/00, 1971. 2. USSR author's certificate in accordance with application 2075321/24, cl. G 06 F 1/02, 1974 (prototype). 3. Authors certificate of the USSR, № 517018, cl. G 06 F 1/02, 1974.