CN105631108A - Dsp Builder based satellite flywheel simulator achievement method - Google Patents

Abstract

The invention provides a Dsp Builder based satellite flywheel simulator achievement method and relates to the field of satellite testing. The Dsp Builder based satellite flywheel simulator achievement method aims at solving the problem that a traditional satellite flywheel model calculating method is not high in calculating accuracy and low in calculating speed. The Dsp Builder based satellite flywheel simulator achievement method comprises the following steps that 1, a graphical modeling method is adopted to establish a flywheel Runge-Kutta mdl file model in a Simulink environment, and flywheel Runge-Kutta modeling and simulation are performed; 2, the flywheel Runge-Kutta model established in the Simulink environment and provided with the suffix of mdl is converted into a common hardware description language VHDL file; 3, a top-layer module is compiled in a QuartusII environment by using verilog language, input/output ports are allocated, and a converted VHDL hardware module is called to achieve calculation of the flywheel Runge-Kutta model. The Dsp Builder based satellite flywheel simulator achievement method is suitable for the field of satellite testing.

Description

Satellite flywheel simulator based on Dsp Builder realizes method

Technical field

The present invention relates to satellite test field, particularly relate to a kind of satellite flywheel simulator based on DspBuilder and realize method.

Background technology

Satellite flywheel, namely counteraction flyback is a kind of important actuator on satellite, plays a part to adjust the attitude of satellite in satellite attitude orbit control. Flywheel simulator, namely counteraction flyback simulator is a part for satellite control system ground checkout equipment, the attitude of satellite and control effect that actuator momentum flywheel determines can be simulated, replace actual part to carry out open loop and the closed test of the attitude of satellite and orbits controlling Computer Sub-System. The input of flywheel simulator main analog flywheel and the electrical characteristics of output, receive the control command (control voltage) that host computer sends, simulate the work process of actual flywheel, output Speed of Reaction Wheels pulse, rotary speed direction and flywheel shaft gentleness motor current signal. Simulator needs the signal gathered to be 4 road voltage signals, through flywheel model calculation (floating-point operation), generates Speed of Reaction Wheels and direction signal. Flywheel simulator has analog input/output port, by isolating device and satellite upper-part electrical isolation, and by communication interface, the direct fault location instruction from dynamics computer can be carried out fault simulation. In flywheel simulator, by solving flywheel mathematical model, obtain the rotary speed data of flywheel, simulate the rotating speed output of true flywheel.

The mathematical model of satellite flywheel is as follows:

$\frac{d\mathrm{\ω}}{dt}=\frac{K_t\frac{U}{R}-(K_v\mathrm{\ω}+K_t\frac{K_e\mathrm{\ω}}{R})-T_{d}sign\left(\mathrm{\ω}\right)}{J},$

Constraints is: | �� |�ܦ�_max,Wherein �� is Speed of Reaction Wheels, K_tFor fly-wheel motor moment coefficient, K_vFor flywheel viscosity friction coefficient, K_eFor flywheel back emf coefficient, U is that flywheel controls voltage, and R is flywheel internal resistance, T_dFor flywheel drag torque, J is Rotary Inertia of Flywheel, and sign is sign function operator;

Flywheel simulator fault parameter injects as shown in table 1:

The flywheel simulator fault parameter that table 1 injects

Solving numerical solution conventional in this flywheel differential equation engineering is Fourth order Runge-Kutta. Fourth order Runge-Kutta can equation shown in formula 1��formula 8 describe: wherein parameter h is integration step, in calculating process, only it is to be understood that the approximation w of previous moment w_nWith integration step h, it is possible to calculate K₁��K₄Value, thus calculating the approximation w of subsequent time w_n+1, by w_n+1And t_n+1As the input of iteration next time, so iterate, until being met the result of calculation of requirement. The mathematical model of counteraction flyback realizes according to the solution procedure of Fourth order Runge-Kutta.

${\mathrm{\ω}}_{n+1}={\mathrm{\ω}}_n+\frac{h}{6}(K_1+2K_2+2K_3+K_4)---\left(1\right)$

t_n+1=t_n+h(2)

K₁=f (t_n,��_n)(3)

$K_2=f(t_n+\frac{h}{2},{\mathrm{\ω}}_n+\frac{h}{2}{K}_1)---\left(4\right)$

$K_3=f(t_n+\frac{h}{2},{\mathrm{\ω}}_n+\frac{h}{2}{K}_2)---\left(5\right)$

K₄=f (t_n+ h, ��_n+hK₃)(6)

$\frac{d\mathrm{\ω}}{dt}=f(t,\mathrm{\ω})---\left(7\right)$

$f(t,\mathrm{\ω})=\frac{K_t\frac{U}{R}-(K_v\mathrm{\ω}+K_t\frac{K_e\mathrm{\ω}}{R})-T_{d}sign\left(\mathrm{\ω}\right)}{J}---\left(8\right)$

Traditional flywheel simulator adopts FPGA (field programmable gate array) to carry out computing and the process of counteraction flyback model as processor, it is utilize hardware description language verilog or vhdl directly to write flywheel Models computed logic, the direct solution differential equation in FPGA processor in QuartusII software environment. Concrete methods of realizing is by designing the soft core of floating-point operation IP in FPGA, realizing floating-point operation. But owing to lacking the support of FPGA tool kit, when using floating-point FPGA operator, it is necessary to a large amount of logics and interconnection resource, cause that floating-point operation performance is too poor. Hardware description language is used directly to describe the advantage place that floating-point operation is not FPGA in summary in FPGA; Traditional flywheel simulator adopts fixed point Dsp to carry out computing and the process of counteraction flyback model as processor, is subject to the restriction of this body structure of fixed-point processor, and model solution precision is not high, and solving speed is relatively slow; In like manner, traditional flywheel simulator adopts single-chip microcomputer to carry out computing and the process of flywheel model as processor, although single-chip microcomputer cost is low, but performance is also low, causes that flywheel model solution precision is not high, and solving speed is slow.

Summary of the invention

The present invention solves that traditional satellite flywheel simulator exists solving precision not high, the problem that solving speed is slow, and propose to realize method based on the satellite flywheel simulator of DspBuilder.

Below this instrument of DspBuilder is done one simply to introduce.

The approach that EDA (electric design automation) technology completes hardware designs is utilized to have a variety of, traditional simulation method utilizes hardware description language Verilog or vhdl directly to write flywheel Models computed logic in QuartusII software environment, the most typical design cycle of the direct solution differential equation in FPGA processor, including design object editor (with hardware description language), comprehensive, emulation, adaptive, programming. But the system for relating to class of algorithms aspect designs, and such as flywheel Models computed, this flow process will seem very inconvenient, and programming is got up very complicated. DspBuilder can help designer to complete the different types of applied system design based on FPGA. Except patterned system modelling, DspBuilder can also be automatically performed most design process and emulation. DspBuilder is system-level (algorithm level) design tool, its framework is on multiple software tools, and the design tool of system-level (algorithm simulating modeling) and RTL (hardware realization) two design fields is coupled together, all it is placed on Matlab/Simulink design platform, and QuartusII is placed in backstage as bottom-layer design instrument, farthest play the advantage of this instrument. DspBuilder depends on the tool of mathematical analysis Matlab/Simulink of MathWorks company, occurs with the form of the Blockset of Simulink. Design and emulation can be patterned in Simulink, further through SignalCompiler, the modelling file (suffix is mdl) of Matlab/Simulink is changed into corresponding Hardware Description Language VHDL design document (suffix is vhd) simultaneously, and for controlling comprehensive and compiling tcl script. Comprehensively and hereafter process is all completed by QuartusII.

Satellite flywheel simulator based on DspBuilder realizes method, do not use conventional methods direct hardware description language and describe floating-point operation, avoid directly describing with hardware description language the inferior position place of floating-point operation in FPGA, then the flywheel model modeling that employing high-level language is patterned on Matlab/Simulink design platform, finally the flywheel model file of Simulink being converted into general Hardware Description Language VHDL, patterned Differential Equation Modeling complexity is far smaller than employing hardware description language and directly describes. The final VHDL language converted adopts 32 floating-point operations that flywheel model is solved, accuracy that very accurate solving result, the operational precision of flywheel model and arithmetic speed transmit for whole satellite simulator data can be obtained in a short period of time and transmission speed has vital impact.

Satellite flywheel simulator based on DspBuilder realizes method, sequentially includes the following steps:

One, adopt patterned modeling method to build flywheel fourth order Runge-Kutta mdl file model under Simulink environment, carry out the flywheel fourth order Runge-Kutta model modeling under Simulink environment and emulation;

Two, it is mdl flywheel fourth order Runge-Kutta model the suffix built under Simulink environment, changes into general Hardware Description Language VHDL file, carry out hardware description language conversion;

Three, under QuartusII environment, write top-level module with verilog language, distribute input/output port, call the VHDL hardware module changed into by flywheel Runge Kutta mdl model file, it is achieved flywheel Runge Kutta solution to model is calculated.

The present invention includes following beneficial effect:

1, the satellite flywheel simulator based on DspBuilder realizes method, system for the class of algorithms designs, hardware description language is not adopted directly to describe floating-point operation, then the flywheel model modeling that employing high-level language is patterned on Matlab/Simulink design platform, patterned Differential Equation Modeling complexity is far smaller than employing hardware description language and directly describes;

2, the VHDL language that the present invention finally converts adopts 32 floating-point operations that flywheel model is solved, it is possible to obtain accurate flywheel Models computed in a short period of time.

Accompanying drawing explanation

Fig. 1 is relation block diagram between each computing module;

Fig. 2 is flywheel Runge Kutta solution process schematic diagram.

Detailed description of the invention

Understandable for enabling the above-mentioned purpose of the present invention, feature and advantage to become apparent from, below in conjunction with Fig. 1, Fig. 2 and detailed description of the invention, the present invention is further detailed explanation.

The satellite flywheel simulator based on DspBuilder described in detailed description of the invention one, present embodiment realizes method, sequentially includes the following steps:

One, adopt patterned modeling method to build flywheel fourth order Runge-Kutta mdl file model under Simulink environment, carry out the flywheel fourth order Runge-Kutta model modeling under Simulink environment and emulation;

Two, it is mdl flywheel fourth order Runge-Kutta model the suffix built under Simulink environment, changes into general Hardware Description Language VHDL file, carry out hardware description language conversion;

Three, under QuartusII environment, write top-level module with verilog language, distribute input/output port, call the VHDL hardware module changed into by flywheel Runge Kutta mdl model file, it is achieved flywheel Runge Kutta solution to model is calculated.

Present embodiment includes following beneficial effect:

1, the satellite flywheel simulator based on DspBuilder realizes method, system for the class of algorithms designs, hardware description language is not adopted directly to describe floating-point operation, then the flywheel model modeling that employing high-level language is patterned on Matlab/Simulink design platform, patterned Differential Equation Modeling complexity is far smaller than employing hardware description language and directly describes;

2, the VHDL language that present embodiment finally converts adopts 32 floating-point operations that flywheel model is solved, it is possible to obtain accurate flywheel Models computed in a short period of time.

Detailed description of the invention two, present embodiment are that the satellite flywheel simulator based on DspBuilder described in detailed description of the invention one is realized further illustrating of method, and the fourth order Runge-Kutta model described in step one is made up of following eight formula:

${\mathrm{\ω}}_{n+1}={\mathrm{\ω}}_n+\frac{h}{6}(K_1+2K_2+2K_3+K_4)---\left(1\right)$

t_n+1=t_n+h(2)

K₁=f (t_n,��_n)(3)

$K_2=f(t_n+\frac{h}{2},{\mathrm{\ω}}_n+\frac{h}{2}{K}_1)---\left(4\right)$

$K_3=f(t_n+\frac{h}{2},{\mathrm{\ω}}_n+\frac{h}{2}{K}_2)---\left(5\right)$

K₄=f (t_n+ h, ��_n+hK₃)(6)

$\frac{d\mathrm{\ω}}{dt}=f(t,\mathrm{\ω})---\left(7\right)$

$f(t,\mathrm{\ω})=\frac{K_t\frac{U}{R}-(K_v\mathrm{\ω}+K_t\frac{K_e\mathrm{\ω}}{R})-T_{d}sign\left(\mathrm{\ω}\right)}{J}---\left(8\right)$

Wherein, h is integration step, and �� is Speed of Reaction Wheels, K_tFor fly-wheel motor moment coefficient, K_vFor flywheel viscosity friction coefficient, K_eFor flywheel back emf coefficient, U is that flywheel controls voltage, and R is flywheel internal resistance, T_dFor flywheel drag torque, J is Rotary Inertia of Flywheel, and sign is sign function fortune symbol, K₁��K₄For intermediate variable, t_nAnd t_n+1Respectively n moment and n+1 moment, ��_nAnd ��_n+1Respectively n moment Speed of Reaction Wheels and n+1 moment Speed of Reaction Wheels.

Detailed description of the invention three, present embodiment are that the satellite flywheel simulator based on DspBuilder described in detailed description of the invention one or two is realized further illustrating of method, the detailed process of the flywheel fourth order Runge-Kutta model modeling under the Simulink environment described in step one:

According to the flywheel mathematical model differential equation, obtain equation below:

$f(t,\mathrm{\ω})=\frac{K_t\frac{U}{R}-(K_v\mathrm{\ω}+K_t\frac{K_e\mathrm{\ω}}{R})-T_{d}sign\left(\mathrm{\ω}\right)}{J}---\left(8\right)$

By formula 3 it can be seen that K₁Value be flywheel initial value ��_n��t_nSubstitute into the value that formula 8 is tried to achieve, so K₁Computing module is namely initial value ��_n��t_nInput to the operational equation of function f (t, ��) gained; By formula 4 it can be seen that K₂Value be handlet_nSubstitute into the value that formula 8 is tried to achieve, so K₂For K₁Value deductWithProduct; In like manner, K₃Value be handlet_nSubstitute into the value that formula 8 is tried to achieve, i.e. K₃For K₁Value deductWithProduct; K₄Value be (a ��_n+hK₃)��t_nSubstitute into the value that formula 8 is tried to achieve, i.e. K₄For K₁Value deductWith hK₃Product;

Under Simulink environment, according to the calculation step of Fourth order Runge-Kutta, graphically call the figure module in DspBuilder and other Simulink storehouses, calculate Kn (n=1,2,3 or 4) computing module and �� according to formula 1��formula 8_n+1Five parts of computing module, complete the once resolving of flywheel model; Relation between each computing module such as Fig. 1 shows; So far, the flywheel fourth order Runge-Kutta model modeling under Simulink environment is completed.

Detailed description of the invention four, present embodiment are that the satellite flywheel simulator based on DspBuilder one of detailed description of the invention one to three Suo Shu is realized further illustrating of method, and the flywheel fourth order Runge-Kutta model emulation process under the Simulink environment described in step one is:

Utilizing the graphical simulation of Simulink, the input of given fourth order Runge-Kutta model, checking output result is correct, completes model emulation.

Detailed description of the invention five, present embodiment are that the satellite flywheel simulator based on DspBuilder one of detailed description of the invention one to four Suo Shu is realized further illustrating of method, and the detailed process of step 2 is as follows:

After the modeling of the flywheel Runge Kutta model under completing Simulink environment, by the flywheel Runge Kutta model file of the SignalCompiler Simulink that the first step is realized, suffix is mdl, changes into general Hardware Description Language VHDL file and for controlling comprehensive and compiling tcl script; The hdl file that conversion obtains is based on RTL, and VHDL that can be comprehensive describes.

Detailed description of the invention six, present embodiment are that the satellite flywheel simulator based on DspBuilder one of detailed description of the invention one to four Suo Shu is realized further illustrating of method, and the detailed process of step 3 is as follows:

Top-level module is write with verilog hardware description language under QuartusII environment, distribution input/output end port, call the VHDL hardware module changed into by flywheel Runge Kutta mdl model file, realize the iterative process of flywheel Runge Kutta, namely flywheel is once resolved in step one �� of generation_n+1��t_n+1It is assigned to initial value ��_n��t_n, down it being iterated computing successively, finally realize flywheel Runge Kutta solution to model and calculate, flywheel Runge Kutta solution process schematic diagram is as shown in Figure 2.

Claims (6)

1. the satellite flywheel simulator based on DspBuilder realizes method, it is characterised in that it sequentially includes the following steps:

One, adopt patterned modeling method to build flywheel fourth order Runge-Kutta mdl file model under Simulink environment, carry out the flywheel fourth order Runge-Kutta model modeling under Simulink environment and emulation;

Two, it is mdl flywheel fourth order Runge-Kutta model the suffix built under Simulink environment, changes into general Hardware Description Language VHDL file, carry out hardware description language conversion;

Three, under QuartusII environment, write top-level module with verilog language, distribute input/output port, call the VHDL hardware module changed into by flywheel Runge Kutta mdl model file, it is achieved flywheel Runge Kutta solution to model is calculated.

2. the satellite flywheel simulator based on DspBuilder as claimed in claim 1 realizes method, it is characterised in that the fourth order Runge-Kutta model described in step one is made up of following eight formula:

${\mathrm{\ω}}_{n+1}={\mathrm{\ω}}_n+\frac{h}{6}(K_1+2K_2+2K_3+K_4)---\left(1\right)$

t_n+1=t_n+h(2)

K₁=f (t_n,��_n)(3)

$K_2=f(t_n+\frac{h}{2},{\mathrm{\ω}}_n+\frac{h}{2}{K}_1)---\left(4\right)$

$K_3=f(t_n+\frac{h}{2},{\mathrm{\ω}}_n+\frac{h}{2}{K}_2)---\left(5\right)$

K₄=f (t_n+ h, ��_n+hK₃)(6)

$\frac{d\mathrm{\ω}}{\mathrm{\ω}}=f(t,\mathrm{\ω})---\left(7\right)$

$f(t,\mathrm{\ω})=\frac{K_t\frac{U}{R}-(K_v\mathrm{\ω}+K_t\frac{K_e\mathrm{\ω}}{R})-T_{d}sign\left(\mathrm{\ω}\right)}{J}---\left(8\right)$

Wherein, h is integration step, and �� is Speed of Reaction Wheels, K_tFor fly-wheel motor moment coefficient, K_vFor flywheel viscosity friction coefficient, K_eFor flywheel back emf coefficient, U is that flywheel controls voltage, and R is flywheel internal resistance, T_dFor flywheel drag torque, J is Rotary Inertia of Flywheel, and sign is sign function fortune symbol, K₁��K₄For intermediate variable, t_nAnd t_n+1Respectively n moment and n+1 moment, ��_nAnd ��_n+1Respectively n moment Speed of Reaction Wheels and n+1 moment Speed of Reaction Wheels.

3. the satellite flywheel simulator based on DspBuilder as claimed in claim 2 realizes method, it is characterised in that the detailed process of the flywheel fourth order Runge-Kutta model modeling under the Simulink environment described in step one:

According to the flywheel mathematical model differential equation, obtain equation below:

$f(t,\mathrm{\ω})=\frac{K_t\frac{U}{R}-(K_v\mathrm{\ω}+K_t\frac{K_e\mathrm{\ω}}{R})-T_{d}sign\left(\mathrm{\ω}\right)}{J}---\left(8\right)$

By formula 3 it can be seen that K₁Value be flywheel initial value ��_n��t_nSubstitute into the value that formula 8 is tried to achieve, so K₁Computing module is namely initial value ��_n��t_nInput to the operational equation of function f (t, ��) gained; By formula 4 it can be seen that K₂Value be handlet_nSubstitute into the value that formula 8 is tried to achieve, so K₂For K₁Value deductWithProduct; In like manner, K₃Value be handlet_nSubstitute into the value that formula 8 is tried to achieve, i.e. K₃For K₁Value deductWithProduct; K₄Value be (a ��_n+hK₃)��t_nSubstitute into the value that formula 8 is tried to achieve, i.e. K₄For K₁Value deductWith hK₃Product;

Under Simulink environment, according to the calculation step of Fourth order Runge-Kutta, graphically call the figure module in DspBuilder and other Simulink storehouses, calculate Kn (n=1,2,3 or 4) computing module and �� according to formula 1��formula 8_n+1Five parts of computing module, complete the once resolving of flywheel model; Namely the flywheel fourth order Runge-Kutta model modeling under Simulink environment is completed.

4. the satellite flywheel simulator based on DspBuilder as claimed in claim 3 realizes method, it is characterised in that the flywheel fourth order Runge-Kutta model emulation process under the Simulink environment described in step one is:

Utilizing the graphical simulation of Simulink, the input of given fourth order Runge-Kutta model, checking output result is correct, completes model emulation.

5. the satellite flywheel simulator based on DspBuilder as claimed in claim 4 realizes method, it is characterised in that the detailed process of step 2 is as follows:

After the modeling of the flywheel Runge Kutta model under completing Simulink environment, by the flywheel Runge Kutta model file of the SignalCompiler Simulink that the first step is realized, suffix is mdl, changes into general Hardware Description Language VHDL file and for controlling comprehensive and compiling tcl script; The hdl file that conversion obtains is based on RTL, and VHDL that can be comprehensive describes.

6. the satellite flywheel simulator based on DspBuilder as claimed in claim 5 realizes method, it is characterised in that the detailed process of step 3 is as follows:

Top-level module is write with verilog hardware description language under QuartusII environment, distribution input/output end port, call the VHDL hardware module changed into by flywheel Runge Kutta mdl model file, realize the iterative process of flywheel Runge Kutta, namely flywheel is once resolved in step one �� of generation_n+1��t_n+1It is assigned to initial value ��_n��t_n, down it is iterated computing successively, finally realizes flywheel Runge Kutta solution to model and calculate.