Summary of the invention
The technical problem to be solved in the present invention is: at the existing problem of present petrolic control mode, provide a kind of and can be in the course of the work change and the motor service condition changes and self adaption generates the strategy of dynamic arteries and veins spectrum parameter, and then the arteries and veins that provides a kind of and original stand to demarcate is composed combined pulse spectrum parameter that parameters combination the controls method to engine control according to the motor correlation properties.
The technical solution adopted for the present invention to solve the technical problems is: this combined pulse spectrum is to the method for engine ignition timing control, it is characterized in that: comprise basic arteries and veins spectrum parameter and dynamic arteries and veins spectrum parameter, substantially arteries and veins spectrum parameter is the arteries and veins spectrum parameter through stand demarcation or process stand and road parameters optimization demarcation, dynamically ignition timing arteries and veins spectrum parameter is that the control system self study is online from the arteries and veins spectrum parameter of demarcating and self-optimizing generates, basic ignition timing arteries and veins spectrum parameter and dynamic ignition timing arteries and veins spectrum parameter constitute the combined pulse spectrum parameter, by the ignition timing control strategy self adaptive control are carried out in the gasoline engine ignition timing by control system.
The composition of arteries and veins spectrum parameter is the experimental process arteries and veins spectrum parameter region sum of different operating mode subregions, and each zone all is divided into closed loop control target and open loop control area by this regional control target value.
Control system comprises that microprocessor, cerebellum joint control CMAC, ferroelectric memory, sensor signal, signal conditioning circuit, power supply detect, power driving circuit, sensor signal links to each other with microprocessor by signal conditioning circuit, ferroelectric memory and cerebellum joint control CMAC are interconnected, cerebellum joint control CMAC and microprocessor are interconnected, microprocessor links to each other with power driving circuit, and power driving circuit links to each other with ignition module.
Dynamically the generation method of ignition timing arteries and veins spectrum parameter is, variation and motor oneself factor according to working condition and service condition change a series of igniting timing auto-adaptive parameters that study generates, this ignition timing auto-adaptive parameter changes ignition timing adaptive learning and experience cluster according to condition by operating mode in the course of the work, uses repeatedly and revises in real time and constantly refresh;
Dynamically the generation method of ignition timing arteries and veins spectrum parameter is produced by following step:
A, ignition timing arteries and veins spectrum generate determining of zone: with the basic correction arteries and veins spectrum of the control ignition timing under a certain working condition, and relevant each the characteristic signal value that characterizes working condition this moment is a back end, basic correction arteries and veins spectrum parameter y with this node is a central value, determine Δ y, to consider following several aspect to the definite of Δ y, the first is with desired ignition timing and actual ignition timing deviation, its two be detonation signal gating probability through the frequency-selecting detection in 5%, it three is that the flare up fire angle is less than a given threshold value; To above three aspects determine to get little after, basic reference radius is determined in the variance ratio and the match of crank angle acceleration dual factors of pressing suction pressure, finds out dynamic arteries and veins spectrum and generates zone (y-Δ y, y+ Δ y);
B, dynamically ignition timing arteries and veins spectrum generates determining of optimizing zone: carry out dynamic ignition timing arteries and veins spectrum in the variance ratio size of utilizing each characteristic signal value that this operating mode of sign is relevant with ignition timing in this back end with the dimension space zone and generate tendency judgement, thereby judge that littler zone is at (y-Δ y) or at (y+ Δ y) on one side, determine that back intermediate value with (y-Δ y) or (y+ Δ y) zone is the new node after the target approaches, and with this target is the center, determine that the dynamic ignition timing arteries and veins spectrum after new the approaching generates the zone, so repeatedly, constantly approach, up to the regional min of minimum (y-Δ y, y+ Δ y) occur, this zone is the optimizing zone;
C, the dynamically generation of ignition timing arteries and veins spectrum: when relevant each the characteristic signal value that characterizes this operating mode levels off to one when being similar to zero constant ε, and the probability distribution of carrying out the correlated characteristic signal that probability statistics handle in allowed limits, determine the median point y among the min (y-Δ y, y+ Δ y)
m, this point is the dynamic ignition timing arteries and veins spectrum parameter of generation;
D, determine dynamic ignition timing arteries and veins spectrum: repeat above process a-c, and CMAC carries out adaptive learning and tracking to the ignition timing control target at overall process cerebellum joint control, and deviation approached adjust and carry out the experience cluster, when the probability distribution of the variance ratio ε of relevant each characteristic signal value and correlated characteristic signal is stabilized in the excursion of a permission, determine this dynamic ignition timing arteries and veins spectrum parameter, deposit ferroelectric memory in, at this moment, relevant each characteristic signal value with pairing ignition timing of dynamic ignition timing arteries and veins spectrum parameter of determining is one group of back end, this node is dynamic ignition timing arteries and veins spectrum parameter, and the set of this dynamic ignition timing arteries and veins spectrum parameter constitutes dynamic arteries and veins spectrum;
E, to refreshing of dynamic ignition timing arteries and veins spectrum: the dynamic ignition timing arteries and veins spectrum of generation is in control procedure, because motor self-characteristic and Environmental Conditions change, its ignition timing control target is also changed to some extent, its back end of forming is when carrying out the process of a-d, when the probability distribution of the variance ratio ε of relevant each characteristic signal value and correlated characteristic signal is stabilized in the unallowed excursion, regenerate new dynamic ignition timing arteries and veins spectrum parameter, CMAC carries out adaptive learning and tracking to the ignition timing control target through the cerebellum joint control, and to deviation approach adjust and the experience cluster definite, original back end address location is refreshed;
Utilize the probability of happening of detonation sensor signal to distribute and the experience cluster of signal values such as the flare up fire angle time probability statistics of flame ionization sensor device signal reflection and crank position acceleration, the dynamic arteries and veins that adaptive learning generates under different operating modes and the condition is composed parameter.
Control strategy comprises ignition timing combination control tactics and Correction and Control strategy or other control strategies.
The combination control tactics of ignition timing and Correction and Control strategy:
A, cooperative action object: act on combination ignition timing arteries and veins spectrum, corresponding to the identical or very close ignition timing target condition of tabling look-up, existing its this ignition timing arteries and veins spectrum, when the dynamic ignition timing arteries and veins spectrum of generation was arranged again, promptly action condition was that the pairing control target of this operating mode has dynamic ignition timing arteries and veins spectrum;
B, combination principle; To same operating mode, have very close condition with operating mode with conditioned disjunction, it is the existing basic ignition timing arteries and veins spectrum that is present in of back end, also have and be present in dynamic ignition timing arteries and veins spectrum, in the element of forming back end, identical and the target component of relevant each characteristic signal value selects dynamic ignition timing arteries and veins to compose parameter not simultaneously; When incomplete same the but target component of relevant each the characteristic signal value of ignition timing is identical, to this characteristic signal value inequality respectively by last cycle values with calculate variance ratio when time cycle values, relatively this variance ratio is got little arbitration, determines that the ignition timing arteries and veins composes parameter; When identical the and target component of relevant each characteristic signal value differs big, get two target intermediate values and approach by dynamic ignition timing arteries and veins spectrum generation strategy and generate new dynamic ignition timing arteries and veins spectrum parameter and insert during dynamic ignition timing arteries and veins composes;
C, combined method: after from dynamic ignition timing arteries and veins spectrum, selecting dynamic ignition timing arteries and veins spectrum parameter, former same operating mode, have basic ignition timing arteries and veins spectrum parameter conductively-closed under the very close condition with operating mode with conditioned disjunction; Dynamically ignition timing arteries and veins spectrum parameter is controlled control target, when the dynamic ignition timing arteries and veins that is determined use is composed parameter to the ignition timing target control time, when the variance ratio of relevant each characteristic signal value can't be stabilized in the allowed band, abandon this dynamic ignition timing arteries and veins spectrum parameter, get back to the basic ignition timing arteries and veins spectrum under this operating mode, this condition, use dynamic ignition timing arteries and veins spectrum generation strategy and relearn generation;
Under d, the above cooperative action, use by dynamic ignition timing arteries and veins spectrum parameter to the part control target, to same operating mode, or a part of basic ignition timing arteries and veins spectrum parameter that replaces, or the whole basic ignition timing arteries and veins that replace under this operating mode are composed;
In correction approached, the Correction and Control strategy of use was:
This correction strategy is made up of with the correction strategy that approaches of cerebellum joint control CMAC conventional correction strategy: conventional correction strategy is the correction of the basic ignition timing arteries and veins being composed from the characteristic signal value of relevant each sensor of reflection engine operating condition, and this part is exported the basic ignition timing control final controlling element of revising the ignition timing target is controlled under conventional control mode; Cerebellum joint control CMAC approaches in the correction strategy, the one, the characteristic signal processing mode that adopts new relevant each sensor is carried out flexible measurement method to the amount of recording not directly and is inferred, and infers and obtain soft measurement features signal value basic ignition timing arteries and veins spectrum is revised; The 2nd, utilize cerebellum joint control CMAC realistic objective to be rectified a deviation by expectation target, and in the correction process, carry out weights couplings and the characteristic signal value of relevant each sensor of adaptive learning is revised basic ignition timing arteries and veins spectrum, the 3rd, the signature variations rate that provides by each sensor, and the signature variations rate that soft measurement is inferred determines the scope of approaching, and constantly approaches the minimum deviation scope by variance ratio basic ignition timing arteries and veins spectrum is revised
This correction strategy is made up of with the correction strategy that approaches of cerebellum joint control CMAC conventional correction strategy: conventional correction strategy is the correction of the basic ignition timing arteries and veins being composed from the characteristic signal value of relevant each sensor of reflection engine operating condition, and this part is exported the basic ignition timing control final controlling element of revising the ignition timing target is controlled under conventional control mode; Cerebellum joint control CMAC approaches in the correction strategy, and the one, the characteristic signal processing mode that adopts new relevant each sensor is carried out flexible measurement method deduction and fuzzy inference to the amount of recording not directly and is obtained soft measurement features signal value basic ignition timing arteries and veins spectrum is revised; The 2nd, utilize cerebellum joint control CMAC to carry out carrying out in the correction process weights coupling and the characteristic signal value of relevant each sensor of adaptive learning is composed the basic ignition timing arteries and veins and revised by expectation target and realistic objective, the 3rd, the signature variations rate that provides by each sensor, and the signature variations rate that soft measurement is inferred determines the scope of approaching, and constantly approaches the minimum deviation scope by variance ratio basic ignition timing arteries and veins spectrum is revised.
Each relevant sensor signals comprises accelerator pedal signal, the crank position of motor and tach signal, top dead center signal, dtc signal, fuel injection pulsewidth signal, throttle position signal, fuel oil temperature signal, current supply circuit voltage signal, water temperature sensor signal, suction pressure signal, air fuel ratio signal, EGR lead signal, and detonation signal sensor and the multifactor relevant match of flame ionization sensor device between the probability distribution of several cycles.
Control system is carried out given ignition timing expected value predictive control according to the signature variations rate judgement working conditions change trend of motor and relevant each sensor of ignition timing to the excessive control target of part time lag deviation, be back end with the predictive control desired value simultaneously, utilize the self adaption of cerebellum joint control CMAC to approach adjustment capability and learning ability, the error that reduction or elimination each side signal lag are brought;
Control system is also carried out economic model, dynamic mode, normal mode judgement according to the signature variations rate judgement working conditions change trend of motor and relevant each sensor of ignition timing, and the selected different ignition timing closed loop control target of self adaption is controlled under different control modes; In control procedure, hocket by control and study, the pattern target is optimized, and in control from now on according to the change of condition, constantly revise and optimised;
Same operating mode down-firing advance angle has a plurality of closed loop control targets, when condition changes, control mode also changes, target as economic model during with dynamic mode is different, control system is according to control strategy, utilize each working condition parameter to carry out pattern analysis and judge that one of closed loop control target by selected pattern under selected automatically this operating mode is carried out closed loop control.Operating mode is meant medium and small load condition, high load working condition, state of starting operating, acceleration and deceleration operating mode and idling operation.
Compared with prior art, combined pulse spectrum of the present invention to the beneficial effect that method had of engine ignition timing control is: owing to adopted with the synthetic combined pulse spectrum parameter control mode of adaptive learning method, make controlled system change and unknown variations has obtained correction to the influence of motor, thus control accuracy and speed when having improved open loop control.Also utilize the planning and the generation of dynamic arteries and veins spectrum parameter, the closed loop control target is revised and selected, feedback signal determines that the single engine control system that causes can't respond when having improved the variation of motor self-condition, the strategy that produces dynamic arteries and veins spectrum parameter by adaptive learning control shifts to an earlier date predictive control, the correction of maximum possible the control hysteresises brought of various time-lag effects, improved the real-time of control.
Embodiment
Implementation process:
The combined pulse spectrum that the present invention proposes is embodied in conventional controller (motor central control unit ECU) and cerebellum joint control CMAC according in the realization of control strategy to the control function to the method for engine ignition timing control; 1-5 illustrates implementation process below in conjunction with accompanying drawing.
As shown in Figure 1: control system comprises that microprocessor, cerebellum joint control CMAC, ferroelectric memory, sensor signal, signal conditioning circuit, power supply detect, power driving circuit, sensor signal links to each other with microprocessor by signal conditioning circuit, ferroelectric memory and cerebellum joint control CMAC are interconnected, cerebellum joint control CMAC and microprocessor are interconnected, microprocessor links to each other with power driving circuit, and power driving circuit links to each other with ignition module.
Relevant sensor signals mainly comprises: suction pressure signal, throttle position signal, cooling water temperature signal, crankshaft-position signal, fuel injection pulsewidth signal, detonation sensor signal, voltage detection signal etc.
Microprocessor is by 32 CPU core, built-in conventional controller control strategy and algorithm, all kinds of arteries and veins spectrum and other relevant control target data and communication bus processor etc.
Cerebellum joint control CMAC is a kernel by another sheet 32-bit microprocessor, constitutes with external circuit; Its built-in adaptive learning algorithm and control strategy, with the common forming control system core of master microprocessor, accept external signal and change, in time make decisions according to strategy, carry out the adaptive learning cluster and refresh dynamic arteries and veins spectrum parameter, send command control outside actuator action and operation.
Ferroelectric memory backs up system basic ignition timing arteries and veins spectrum parameter, after participating in operating mode control behind the adaptive learning, be judged as make system on request that part of dynamic arteries and veins spectrum parameter of stable operation also can deposit in wherein as posterior infromation.Can automatically solid-state substantially arteries and veins spectrum parameter be write the microprocessor from ferroelectric memory when the microprocessor decision-making system is out of control.
Cell voltage signal detects by power supply and handles back access microprocessor, and power supply detects and mu balanced circuit is made up of DC/DC transducer, over-current overvoltage protector, voltage change signal transmitter and anti-jamming circuit.
The power drive control circuit adopts special-purpose controlling and driving chip and peripheral circuit, drive ignition module etc.
What here specify is, for ease of distinguishing new controlling method, the present invention is with conventional process mode and method, as use of PID control strategy etc., all be defined as conventional controller, conventional controller is as a control system part, and another part of control system is referred to as cerebellum joint control CMAC.
As shown in Figure 2: control system is handled the throttle position signal of intention and the multifactor coherence of these signals according to the crankshaft-position signal of reflection rotating speed and top dead center position, the suction pressure signal of reflection payload, fuel injection pulsewidth signal, the reflection of reflection air fuel ratio situation, provides basic ignition timing arteries and veins parameter spectrum by the ignition timing control strategy.
Because the duty parameter relevant at ignition timing control is a multifactor correlated process in control system, ignition timing and engine speed, load, air fuel ratio, cooling water temperature, compression ratio,, suction pressure, octane number of fuel oil, mixed gas turbulent extent, EGR leads and the shape of firing chamber all has relation; When implementing control, control system is gathered the engine operating condition parameter relevant with ignition timing, and the detonation signal during closed loop control, in the time of if necessary, flare up fire angle (flame ionization sensor device signal) also carried out probability statistics to be handled, better effect will be arranged, this is because this sensor can characterize combustion process, promptly measure the flame front due in, utilize turbulent flow, the soft measurement processing capacity of the statistical theory of Combustion and cerebellum joint control CMAC, extrapolate the laminar flame velocity of propagation, thereby extrapolate burning to continue the angle and self adaption correction ignition timing arteries and veins spectrum parameter; What specify is that basic ignition timing arteries and veins spectrum parameter will be a primary condition with above correlative factor, be divided into Economy, power character, emission performance and the comprehensive basic ignition timing arteries and veins spectrum of distinguishing by operating mode and condition by the IGNITION CONTROL strategy, adaptively selected in control procedure.
Conventional controller is revised output by gathering the duty parameter relevant with ignition timing to basic ignition timing arteries and veins spectrum, it is revised by relevant selected switch influence, relevant selected switch is provided with at octane number of fuel oil and compression ratio, because octane value difference, velocity of combustion is also different, the compression ratio difference, flame travel is different with the propagation time; Wherein, octane value selected is to determine according to probability density or flame ionization sensor device and the crankshaft-position signal acceleration of detonation signal in the unit time; Compression ratio is selected to be compression ratio difference by different motors, determines that by stand test influences a constant.What specify is the effect of combined pulse spectrum parameter control strategy to the ignition timing arteries and veins spectrum of input conventional controller, under the effect of combined strategy, when suitable dynamic arteries and veins is composed parameter, differentiation according to control, determined dynamic arteries and veins spectrum parameter when making stable condition of motor, will be when this all conditions possesses, the basic ignition timing arteries and veins spectrum parameter that the former stand of all or part of replacement of possibility is demarcated, ignition timing this moment control strategy passes through combined strategy directly with this operating mode in the ferroelectric memory, dynamic arteries and veins spectrum parameter under this condition is sent into conventional controller, through conventional controller correcting process output ignition timing correction arteries and veins spectrum parameter, this arteries and veins spectrum driving parameter ignition control module (device) control plug ignition.
Cerebellum joint control CMAC carries out adaptive learning and tracking to the ignition module control target of revising (arteries and veins spectrum parameter), and according to the variance ratio of the relevant duty parameter of motor, the probability density distribution of detonation sensor signal, the probability density distribution of flame ionization sensor device, the suction pressure variance ratio, the fuel injection pulsewidth variance ratio, the throttle position variance ratio, the crank angle acceleration, delay angle and burning continue the deviation at angle and the method that deviation variation rate provides by Fig. 2 and Fig. 3, generate the dynamic arteries and veins spectrum of ignition timing parameter, this arteries and veins spectrum parameter writes ferroelectric memory after the optimizing condition is determined; When the ignition timing control strategy is declared when determine using combined pulse spectrum parameter strategy, the dynamic arteries and veins spectrum of ignition timing parameter is carried out the ignition timing self adaptive control with the synthetic combined pulse spectrum parameter of basic fuel injection arteries and veins spectrum parameter to engine ignition control module (device) under the effect of combined strategy.
Self adaptive control for ignition timing, the effect of comprehensive flame ionization sensor device, having provided detonation sensor signal probability by n circulation pinking generation after the effect of frequency-selecting cymoscope here is that optimum igniting is adjusted targets threshold with at 2%-5%, and the probability that pinking takes place surpasses this threshold value with the post ignition angle; The signal of related sensor is differentiated the optimum condition that is confirmed as this igniting arteries and veins spectrum parameter through the experience cluster in this scope, when partial condition changes, cerebellum joint control CMAC will be constantly study and control of expectation target with this threshold value, thereby replace traditional pinking safe angle distance.
As shown in Figure 3: according to the ADAPTIVE CONTROL of combined pulse spectrum of the present invention to the method for engine ignition timing control, once the motor that comes into operation, except that beginning it is by the basic ignition timing arteries and veins spectrum parameter work of optimizing through stand test, because the effect of adaptive strategy, constantly self adaption produces the dynamic ignition timing arteries and veins spectrum of new optimization parameter, thereby the motor of work a period of time, its basic ignition timing arteries and veins spectrum parameter more or less changes, promptly be come into operation simultaneously with through the motor after the identical operating time, its with basic ignition timing arteries and veins spectrum parameter of control target also change no longer identical.
Control unit is implemented control to ignition timing system will use basic ignition timing arteries and veins spectrum parameter, to provide different desired values corresponding to different operating modes, this desired value produces deviation owing to various Environmental Conditions, condition, transmission time lag, mechanism driving time lag, property difference etc. with realistic objective; To the signal feed back of realistic objective, owing to the characteristic of sensor, the transmission time lag of signal, the time lag of signal operation processing procedure etc.; Also comprise at random device property sudden change that the interference that produces and interference cause etc.; More than or the like the existence of factor, have influence on the real-time and the accuracy of control, add can't " because of ground system easy " stand adjusted demarcates arteries and veins spectrum parameter, makes control system can't accurately determine control target.
Use cerebellum joint control CMAC, ignition timing ADAPTIVE CONTROL in conjunction with traditional PID control is implemented control from two aspects, carry out self adaptive control at the deviation of ignition advance angle on the one hand, as asking deviation by crankshaft-position signal and control system ignition advance angle control signal and asking deviation variation rate, ask delay angle and burning to continue the deviation at angle, utilize the ignition timing of exhaust gas temperature change rate feedback to combustion effects, and infer air fuel ratio by rotating speed and throttle position, and then definite air fuel ratio is to the influence of ignition timing, and the effect of gas handling system ejector makes the flare up fire angle change release the gas compressor rotation speed change to the ignition timing influence etc., promptly uses the self-correcting self adaptive control (on-line identification system) of guaranteeing system stability and guaranteeing ignition timing target capabilities optimum.Be the adaptive learning control of the relevant sensor signals of a certain control target being carried out variance ratio and probability distribution tracking on the other hand, as the crank angle acceleration, the probability distribution at pinking and flare up fire angle etc., be that pick-up transducers feedback quantity probability statistics and variance ratio trend criterion and system stability sex experience cluster are approached, and the self study after the stable objects generate dynamic ignition timing arteries and veins spectrum target prediction is controlled in real time.
Among Fig. 3,
(1) establishes controlled output quantity y
j(j=1,2,3 ...), the time t of a continuous measurement motor n work cycle with this understanding
iVariance ratio with tach signal, throttle position signal, fuel injection pulsewidth signal, suction pressure signal, detonation signal, flare up fire angle signal, atmospheric pressure signal, water temperature signal, fuel oil temperature signal, EGR valve opening signal, exhaust temperature signal, battery tension signal and above each signal, and handle flame ionization sensor device and detonation sensor signal especially, be optimum igniting adjustment threshold value 2%5% with in as press probability that n circulation pinking take place after the effect of frequency-selecting cymoscope according to the detonation sensor signal, and replaces traditional pinking safe angle distance.
Above signal is through control system and with following formula
Carry out The Fitting Calculation and handle, obtain basic controlling target y
i
In the formula, a
0Be basic arteries and veins spectrum value or sensor signal value, a
1Be control target variance ratio or sensor signal variance ratio; t
iI cycle time; y
iI circuit control target mean change amount.
(2) data recycle by match concerns
Match, in the formula, d1/dt, d2/dt ..., dn/dt is respectively relevant measured value variance ratio; Controlled output quantity is changed into y
j=y
J-1+ Δ y, (i=1,2,3 ...);
(3) the design control law logic determines that controlled output quantity is by y
J-1Change y into
jAfterwards, each measured variance ratio of motor levels off to zero, and this levels off to zero value ε and is regarded as optimum condition, and optimised the selecting of the desired value y under this condition becomes new control target, and the corresponding condition changing of tabling look-up.
y
j-y
J-1, b is arranged at>0 o'clock
j-b
J-1>ε, then y
J+1=y
j+ Δ y (Δ y>0).
y
j-y
J-1, b is arranged at<0 o'clock
j-b
J-1<ε, then y
J+1=y
j+ Δ y (Δ y>0).
y
j-y
J-1, b is arranged at>0 o'clock
j-b
J-1<ε, then y
J+1=y
j-Δ y (Δ y>0).
y
j-y
J-1, b is arranged at<0 o'clock
j-b
J-1<ε, then y
J+1=y
j-Δ y (Δ y>0).
(4) be to be the process that y is carried out self adaptive control of feedback parameter with each related sensor measured value of motor and diverse vector thereof more than.(each condition flag value parameter changes little this process if make the motor smooth working in the circulation of n task, promptly have and level off to zero ε, probability distribution in allowed limits), then control target y is defined within (y-Δ y, y+ Δ y) in the solid zone, draw the control target area of space, and in the zone constantly interpolation force into the little space zone and control; Each parameter of motor that show this moment is the engine optimum conditional parameter, and the desired value under this condition is the optimised control target of selecting.This moment, this control target was write dynamic arteries and veins spectrum parameter region by the experience cluster in accordance with regulations, and this address is refreshed if be substituted when the data of the condition of not conforming to are arranged.
(5) when variation tendency stable or optimum condition in n circulation occurs, control strategy will be predicted elimination by the pattern of study to each time-lag effect of control unit and motor, association's control when similar events as takes place once more; Working conditions change or when changing with condition under the operating mode again by mentioned above principle, so repeatedly;
(6) in the above process, n is that the boundary determines with stable robustness, and the ε value is a multifactor relevant small quantity constant.These two parameters are verified repeatedly when the stand data and are added their confirmation.
As shown in Figure 4: when motor enters work, control system is judged the current operating mode classification of motor according to the state signal of different operational conditions and each sensor, the state that is basic operation condition and current related sensor constitutes the current base regime condition that control system is selected operating mode, control system is determined current working according to above-mentioned condition, calculates the basic ignition timing arteries and veins spectrum parameter under this operating mode of output.
If this process has the dynamic ignition timing arteries and veins spectrum parameter that generates through adaptive learning to exist, control system is declared ratio through optimizing stability, if this dynamic ignition timing arteries and veins spectrum more is better than basic ignition timing arteries and veins spectrum, then output is newly-generated dynamic ignition timing arteries and veins spectrum parameter.This dynamic ignition timing arteries and veins spectrum is carried out current correction by the feedback signal of sensor, and revised dynamic ignition timing arteries and veins spectrum branchs parameter is controlled each final controlling element by drive circuit, as the control ignition module to determine ignition timing; When this process is carried out, the one, by the last circuit actuator target localization situation of each location status signal measurement feedback, control system is carried out the adaptive weight correction with the revise goal value calculation deviation and the deviation variation rate input cerebellum joint control CMAC of realistic objective value and output, the 2nd, obtain the controlled object inversion model by feed-forward mode training and tracking, if with x (k) expression system mode, during u (k) expression dominant vector, control to actuator is described as x (k+1)=g[x (k), x (k)]; The 3rd, control system is by calculating the variance ratio of its semaphore in the regulation cycle period to the sensor signal relevant with control target, determine variation tendency by variance ratio, with determine controlling party to, utilize this variation tendency prediction to provide the desired output target by control strategy, calculate by deviation and deviation variation rate with the actual measurement target, the variance ratio of each relevant sensor signals calculates, and constantly revises weights, level off to zero stable trend by each variance ratio, approach control target.
When the experience cluster signal of system self-adaption study and the deviation variation rate minimum that reaches stable threshold based on variance ratio and deviation hour, this predictive control target is confirmed as the dynamic ignition timing arteries and veins spectrum value that will select.This arteries and veins spectrum value is sent the output that is used for control target in erasable storage, repeat aforesaid process, constantly calculates each associated change rate of last circuit, control and study in current circulation, prediction output in next circulation.Study hockets with control.
When stability threshold occurs, the dynamic ignition timing arteries and veins spectrum that the dynamic arteries and veins spectrum parameter value of this predictive control target promptly generates, deposited in the ferroelectric memory, each sensor signal value when stability threshold occurs also is confirmed as determining the working condition signal of this arteries and veins spectrum output simultaneously.In like manner, when condition changes, repeat above process, constantly generate corresponding dynamic ignition timing arteries and veins spectrum.
Hocket in the process in study and control, the signal value of the dynamic ignition timing arteries and veins spectrum of generation and each related sensor when generating this dynamic ignition timing arteries and veins spectrum is stored by cluster optimization by the data processing principle in the control strategy; The principle of optimizing is divided two aspects, the one, constantly the operational condition of base regime condition and memory is found out the optimizing zone to dynamic ignition timing arteries and veins spectrum value by trend constantly approaches control, determine the back end when optimal conditions ε occurs, reduce the space hold rate like this, also shortened the generation cycle of dynamic ignition timing arteries and veins spectrum simultaneously.The 2nd, adopt compact type address space storage policy, avoid the deallocation of unnecessary unit, promptly adopt unified address complementation computing to obtain training the space of depositing weights, realize requirement to satisfy hardware.
As above cerebellum joint control CMAC flow process, to the working procedure of ignition timing control.
In current study and control stage, cerebellum joint control CMAC is according to the ignition advance angle position of last circuit in the control system, the crank angle acceleration, the suction pressure variance ratio, the air fuel ratio variance ratio, water temperature and intake temperature variance ratio, force in the ignition advance angle arteries and veins spectrum parameter of into optimizing with operating mode by smallest interval and be back end with the arteries and veins spectrum parameter value of condition, the variance ratio scope of associated sensor signal excursion (as rotating speed) and signal is determined working condition input space Ug=[a, b] * [c, d], determine Um=[e between arteries and veins spectrum parameter tracking modified spatial according to target of prediction and realistic objective deviation range and deviation variation rate scope, f] * [g, h], as the crank angle acceleration 1 to 10, its variance ratio is 0 to 3, then the standard product space is Ug=[1,10] * [0,3]; And choose suitable quantification progression, provide initial weight coefficient matrix, change and the variance ratio of signal and current actuator position signal and change rate signal are back end with current associated sensor signal, choose suitable parameters and space hypergeometry body radius, find out the space hypergeometry body that comprises this point according to given sample, determine selection matrix S, this moment, the output of cerebellum joint control CMAC was defined in to activate the linear combination of basic function on the hypergeometry body that node is the center, promptly
Wherein: B (x
t)=diag[b
1(x
t), b
2(x
t) ..., b
m(x
t)], q=[q
1, q
2,, q
n]
TBe the weight coefficient vector, S
m=[S
T, m]
N * mFor weight coefficient is selected vector,, only need the local weight coefficient of adjusting to get final product like this for each sample.Like this through constantly study and control, constantly repeat above process, study hockets with control, generate satisfactory dynamic arteries and veins spectrum parameter, ignition advance angle in next circulation is carried out predictive control, distribute according to pinking and flame conduction angle threshold probability again, determine the input space, and definite final controlling element relevant with ejector adjusted the input space, selection matrix S, the learning experience cluster of (a plurality of cyclic process) through has after a while reached the realistic objective value by repeatedly approaching, the elimination of maximum capacity the control deviation brought of time lag, thereby make ignition timing reach accurate control.
As shown in Figure 5: 31,32 pin of microprocessor U1 link to each other with 29,24 pin of storage U16 respectively, and 40 pin connect the VCC high level by resistance R 1, by capacitor C 1 ground connection, by switch S 1 ground connection; Be connected to crystal oscillator Y1 between 73,74 pin of microprocessor U1, and by capacitor C 2, C3 ground connection;
After the signal of air inlet pressure sensor entered phaselocked loop U3 and carries out the V/F conversion treatment through buffer storage U2, A/D mouth P50, P51 pin by photo coupler OP1 is input to microprocessor U1 carried out analytical calculation for microprocessor U1 and handle.
Throttle position signal is input to the A/D mouth P46 pin of microprocessor U1 after being input to operational amplifier U5 amplification processing after the step-down, carry out analytical calculation for microprocessor U1 and handle.
After the signal of pinking amplifies processing by the signal selective frequency amplifier circuit of being made up of operational amplifier U6 and peripheral circuit thereof, be input to the detecting circuit of forming by operational amplifier U6E, the output signal of cymoscope is carried out analytical calculation for microprocessor U1 and is handled through the P16 pin of a not gate buffering back input microprocessor U1.
Phase inverter U7 and gate circuit U8 form fuel injection signal pulse width discriminating circuit; Fuel injection signal is input to the INTP0 mouth P01 pin of microprocessor U1, carries out analytical calculation for microprocessor U1 and handles.
CRANK SENSOR is input to after magnetic transformer U9 carries out conversion treatment, is input to the A/D mouth P57 pin of microprocessor U1, carries out analytical calculation for microprocessor U1 and handles.
After power supply was handled by the power sense circuit of being made up of phaselocked loop U10, the P26 pin by photo coupler OP3 input microprocessor U1 detected the battery voltage amount in real time, for system provides the reliability constant voltage dc source.
Microprocessor U11, latch U12, storage dynamic U13 constitute cerebellum joint control CMAC, under the control of microprocessor U1, according to built-in control strategy adaptive learning, and approach regulated by the air fuel ratio desired value; Storage dynamic U19 is a flash memories, and it is regulated parameter to birdsing of the same feather flock together and refreshes storage, in the controller control that participates under the control of microprocessor U17 under the new operating mode.
Constitute preparation increased flash memory device by expansion mouthful U15 and storage U14, storage system arteries and veins spectrum MAP data.
Microprocessor U1 utilizes its I/O port P30-P37, by switching value driver U16, U17 fire signal is gathered with feedback analysis and is declared than after handling, and BT1-BT4 controls in real time to the igniting of motor by the power drive pipe.