• 专利附录
  • 专利说明
  • 权利要求
  • 类似技术
  • 同族专利
  • 引用文献
  • 法律状态
  • 优先权
    • 专利摘要:
    For the determination of a non-linear controller for a nonlinear system, it is proposed that a parameter set (KPID (k)) of the controller (1) be determined by means of an optimization with a multi-criteria evolutionary algorithm, in which in each evolution step a plurality of parameter sets (KPID (KPID) k)), each representing a possible solution to the optimization, are determined and at least two quality values (fi) are determined for each parameter set (KPID (k)) and the quality values (fi) are optimized by the multi-criteria evolutionary algorithm.
    公开号:AT512251A2
    申请号:T50129/2013
    申请日:2013-02-28
    公开日:2013-06-15
    发明作者:
    申请人:Avl List Gmbh；
    IPC主号:
    专利说明:

    AV-3524 AT
    Method of designing a nonlinear controller for non-linear processes
    The present invention relates to a method for designing a nonlinear controller for nonlinear processes modeled in the form of a local model network
    Linear regulators, e.g. PID (proportional integral derivative controller) controllers are widely used to control linear systems. However, conventional linear regulators only work unsatisfactory for non-linear systems. However, most real processes are non-linear. Therefore, it has been proposed to approximate nonlinear systems by local linearizations, whereby linear regulators can be designed for the local linear systems. Such local linearizations of a nonlinear system are e.g. 10 with multiple model approaches, e.g. local model networks (LMN). The use of a local model network (LMN) for this purpose is a known method in which interpolation between different local models that are valid in different operating ranges (input variables). For the various local models, then linear rules, such as e.g. PID controller, wherein the global controller output 15 is again determined by interpolation of the local controller outputs. An alternative approach is modeling with neural networks or fuzzy logic.
    A key criterion in any regulated system is the stability of the closed loop. A designed regulator must therefore have a certain stability criterion in the control loop, e.g. a known bounded-input-bounded-output (BlBO) or a 20 Lyapunov criterion.
    It is also desirable that a desired closed-loop behavior be achieved over the entire output range, e.g. In certain output ranges no undesirable high control deviations or overshoot occur. In the normal controller design, the stability and behavior of the closed loop is checked after the controller has been designed. If a quality value is not met, then the controller design has to be repeated, which of course is outlandish and ineffective in practice. Furthermore, therefore, no optimization of the two quality values stability and behavior of the closed loop is already possible in the controller design, or only 30 implement very expensive.
    It is therefore an object of the present invention to provide a method for designing a nonlinear controller for non-linear systems, which may take the form of a local model network.
    AV-3524 AT
    Plant, where certain quality values for the closed-loop control can be taken into account.
    This object is achieved according to the invention by determining a parameter set of the controller by means of an optimization with a multi-criteria evolutionary algorithm, in which a plurality of parameter sets, each representing a possible solution of the optimization, are determined in each evolution step and for each parameter set At least two quality values are determined and the quality values are optimized by the multi-criteria evolutionary algorithm. In this way, the desired quality values are already taken into account in the controller design and are not determined only after the controller design has been determined. However, the quality values are not only taken into account but also optimized in the course of the controller design, so that the designed controller even represents an optimium with regard to these quality values, ie a controller design which is the best possible in this regard.
    Advantageously, a quality value for the stability of the control loop and a quality value for the behavior of the control loop are determined and optimized as quality values, since these are the usual criteria for the evaluation of a control loop. For this purpose, the quality value for the stability of the control loop is preferably determined using a Lyapunov criterion with a decay rate, the quality value used being the decay rate. For the quality value of the behavior of the control loop, a setpoint signal and a permissible tolerance range of the output variable are defined around the setpoint signal, and a measure of compliance with this tolerance range is used as the quality value.
    In a particularly advantageous manner, a characteristic map can be created from the determined nonlinear controller for a determined controller parameter, which then can be used, for example, to calculate a control parameter. can be used in a control unit in a vehicle to control certain sizes. In this way, maps in vehicle control units can be directly determined and required and no longer need to be laboriously calibrated on the test bench, for which hitherto a large number of test runs with the test object (for example internal combustion engine, transmission, powertrain, vehicle) have been necessary. For this purpose, as a result of the optimization, a paroffoff with possible optimal controller parameters can be determined and from this a parameter set can be selected as a solution. This can then be e.g. be checked by a test run on the test bench for a test object.
    The subject invention will be explained in more detail below with reference to Figures 1 to 11, which show by way of example, schematically and not by way of limitation advantageous embodiments of the invention. This shows -2-
    AV-3524 AT
    1 shows an example of a local model network,
    FIG. 2 shows a local model network with a local governing network, FIG.
    3 shows a block diagram of a control loop,
    A block diagram of the control loop in state space representation, 5 10th
    5 shows a schematic representation of the sequence of the controller design,
    6 shows an example for the determination of the quality value for the behavior of the control loop,
    7 shows the solutions of the controller design in the form of a Pareto front,
    8 shows the result of the regulation of a non-linear process with the designed non-linear regulator and
    9 to 11 examples of characteristic diagrams of the controller parameters.
    The aim of the subject invention is to design a nonlinear controller for a non-linear controlled system or a non-linear process. For this purpose, it is first assumed that the nonlinear process is modeled in the form of a time-discrete, local model network 15 (LMN). LMN are well known and there is extensive literature on methods for the identification of a nonlinear process by means of an LMN, which is why in Fig.1 only briefly will be discussed. An LMN interpolates between local models LMi, each of which is valid in certain operating ranges (or ranges of the input variables uO), In which each i-th local model LMj of the LMN can consist of two parts, namely a validity function Φι and a model parameter vector 6j The model parameter vector Θ includes all the parameters for the ith model and the validity function Φϊ defines the scope of the ith local model within the partitioning space which is a subspace of the input space A local estimate of the output yt (k ) as the output of the i-th local model LMj at time k results from 25 y, (k) - xT {k) Oi, where x (k) denotes a regression vector, the current (k) and past (kn) and outputs Uj, yt (k). The global model output y (k) then results from a linear combination with a weighting of the M local model outputs by the validity function Φ, in the form y (k) = ^ Φ, (k) yt (k). For each local model LMj, a local, linear controller LCj is now created, as shown in FIG. 2. This results in a local area network (LCN). The global non-linear controller 1 then results again by linear combination of the local linear controller LCj as a function of their validity ranges, similar to the LMN.
    -3-
    AV-3524 AT
    The control loop is then known as shown in Figure 3. The output quantity>> (/ :) is fed back to the input and subtracted from a setpoint value w (k), which results in a control deviation e (k) which is fed to the controller 1 and uses this to produce a manipulated variable u (k). (= Model input) for the nonlinear process 2 or the process LM-modeling LM-5. In addition, disturbances z (k) can be taken into account. The well-known discrete-time linear law of control is known to be u (k) = u {k -1) + d0e (k) + d ^ eik) + d0e (k -1) + d2e (k-2) with the coefficients
    2 t, T, -1 and d2 = Kp ~. Here Kp, TN and Tv are the T Js
    Controller parameters and TN the Sampiingzeit. With e (k) = w '(i) - y (k), this rule of law can be reformulated io. For an i-th local PI D-controller, then the discrete-time linear law of control follows to m0) (A) = uU) {k -1) + / ^ (/:) ^ (^) - / ^ (^) ^ (^ ), where the matrix
    Kp] D (k) contains the controller parameters Kp, TN and Tv of the individual local PID controllers. The global matrix of the controller parameters KplD {k) is then again given on the basis of the LMN as a linear combination of KP1D (k) - £ Φ, Α ^, (A). This results in different controller parameters KPID (k) for different input variables ) of the non-linear regulator 1.
    As is known, the control law can be equivalently brought into the state space representation with the state x (not to be confused with the regression vector x from the LMN), in the form x (k + 1) = Λ (Φ) χ (£) + / ϊ (Φ ) Δη (£) + E (0) z (k), y (k) = Cx (k). With the system matrix Λ (Φ) = ^ Φ, 4 (k), the input matrix B (Φ) =] | ΓΦjBi (k), the perturbation matrix £ (Φ) = ^ Φ, ϋχ #) and the output matrix C Au (k) results
    Au (k) = K (£) w (£) - K (k) y (k). The state space representation is shown by way of example in FIG. In order to keep the number of state variables small, the time shifts of the disturbances are carried out externally {indicated by the block receiving the disturbances 25 z (k)). The disturbance matrix E contains the associated model parameters. Equally, there are equivalent control laws and state space representations for other linear controllers. This is standard control theory, which is why it will not be discussed further here.
    The controller parameters KPID (k) of the linear regulator 1 defined above are now determined by means of a multi-criteria evolutionary algorithm. Such algorithms are also well known, e.g. as a multicriteria genetic algorithm (multi-objective genetic algorithm or multiGA). In general, evolutionary algorithms work analogously to the natural
    AV-3524 AT allowed evolution where stronger individuals are more likely to prevail.
    The "strength" of an individual is measured with quality values fi (fitness function). In each step of the multi-criteria evolutionary algorithm, a large number of individuals are generated and the quality values defined for each individual are calculated. 5 In creating new individuals, two mechanisms come to mind, as in nature: inheritance and mutation. The individuals of one generation are being mixed and recombined to create the next generation of individuals. Due to the quality values f | better individuals have a higher chance to pass on their genes (controller parameters) to the next generation (the better one gets through). The han-10 mutation is a random change in genetic information (controller parameters). As a rule, optimization aims at minimizing the quality values f, whereby maximization can also be the optimization goal. Such algorithms are well known and there is also sufficient literature, e.g. K. Deb, "Multi-objective optimization using evolutionary algorithms", Cinchester John Wiley & Sons Ltd., 200Θ or 15 LD Li, X. Yu, X. Li, and W. Guo, "A Modified PSO Algorithm for Constrained Multi-objective Optimization," presented at the International Conference on Network and System Security, 2009, p , 462-467.
    In the concrete case of application, shown in FIG. 5, a parameter set KplD (k) of the controller 1 represents an individual and a stability criterion 20 and a criterion for the behavior of the control loop are used as quality values fs, fp. For initialization, any parameters KPID (k) can be specified. For initialization but you can also use existing methods, such. the well-known command pidtune from MATLAB®, which determines a good starting point of the controller parameter KPID (k) for the multicriteria evolutionary algorithm. In each step of the evolutionary algorithm, a pool of individuals U.. Represent population P of the evolutionary algorithm. For each individual of each pool of individuals h ... In then the quality values fs, fp are determined. On the basis of the quality values fs, fp the individuals of the next generation are then determined according to the rules of the multicriteria evolutionary algorithm. The termination of the optimization takes place either after a predetermined number of generations or when desired quality values fSl fp have been reached. For the two quality values fs, fp for the assessment of the stability and behavior of the closed-loop control system, e.g. the following approaches are made. -5-
    AV-3524 AT For stability, e.g. made a known Lyapunov approach on the basis of the above state space representation in which a Lyapunov function V (x) is sought that satisfies the following requirements: 0 V (x {k) = 0) = 0 ii) V (x (k)) > 0 for x (k) * 0 iii) V (x (k)) go / ο / · | χ (/ :) | - »< x > iv) AF (x (A)) = V (k + ) - a1 V (k) < 0 5 It is usual for LMN to limit the Lyapunov function V (x) to quadratic functions V (k) = xT (k) Px (k)> 0 with P> - 0. Where P is a positive definite matrix and α is a degree of decay. Alternatively, of course, other known Lyapunov functions come into question, such as e.g. Piece-wise square or fuzzy Lyapunov features.
    The criterion for stability then results for a quadratic Lyapunov function V (x) 10: P &gt; 0 inf {ο &lt; a &lt; 1: A ^ .PA ^ + Xü &lt; a 2p} inf {ο <a &lt; 1: A ^ PA ^ + Xy <a2p with 'Xu Xn ... xu' x = Xn * 22 ... x27 A Xu ··· Xn, V / e 3, Vr &lt; 7 &lt; 1 -r -Gv + Gß
    Gq = Aj -BtKpm jC Gn = A; - BjKp ^ C
    The LMN controlled by the controller 1 is exponentially stable when there are symmetric matrices P and Xj and a decay degree α satisfying the above conditions. The above system of equations can be solved with available equation solvers, where the quality factor fs for the stability here is the degree of decay α, ie fs = a (KTP1D1, / iz3). -6 ·
    For the behavior of the closed loop, at first, the output y (k) becomes 5 10
    Setpoint signal wp (k) with time length K generated, e.g. using a Design of Experiment method that preferably covers the entire output range of the LMN to capture the global closed loop lock. The closed-loop behavior quality value fp is then based on upper and lower limits yUb, y », the reference signal Wp (k), and conventional behavioral criteria, such as e.g. Overshoot Ayos, undershoot Ayui, rise time k ,, settling time kj and bandwidth bw- Thus, a permissible tolerance range is defined around the setpoint signal wp. The permitted range of the output variable is defined between upper and lower limit value yub, y <b, ie within the tolerance range, as exemplified in FIG. The quality value fp then results from fp = ΣΛ <*) with * = 1
    Aos + A sign {Awp (k ')) (y (k) - yut> (k)) &gt; 0 / "(*) = {0 and
    From + L * sign (& wp (k ')) (y (k) - ylb (k)) &gt; 0 f = <cMk) -y * (kt sign (& vp (k *)) = -1 c »IK *) -. Vä (*) | »5» (Awp (* ·)) =! 15 The coefficients Cos and describe the increase of the quality value fp when the system output y (k) leaves the permissible range between the limit values yub, y) b. The quality value fp for the behavior of the control loop is thus a measure of compliance with the tolerance range. For the determination of the quality value fp, the control loop (FIG. 3) is subjected to the desired value signal wp for each parameter set KPID (k). The limit values yub, ylb 20 are determined by means of usual parameters such as rise time kr, settling time ks, overshoot Ayos, undershoot Ay ^ and bandwidth bw, e.g. from a user.
    The multi-criteria evolutionary algorithm can optimize (minimize or maximize) multiple quality values and determine corresponding parameter sets Kpm (k), with several possible optimal solutions. The amount of optimal solutions can be represented as a pareto front. The Paretoffont is known to contain all solutions in which it is not possible to improve a quality value fj without
    -7-
    = 102013 ^ 0129
    AV-3524 AT at the same time to deteriorate another quality value f. For the case of the two quality values fs, fp, e.g. a pareto front P, as shown in Fig. 7. Behind everyone
    Point of the pareto front P is an associated parameter sets KMD (k) of the controller 1, which each represents a possible optimal solution of the controller design. Thus, it remains only 5 more to select a particular solution and thus a specific parameter set KPIb (k) from the possible optimal solutions ,
    In addition, the various quality criteria fj in the optimization could also be weighted to evaluate their significance.
    8 shows the result of the design of a PID controller according to the above method. In the diagram above, the desired value wp (k) and the output j> (£), together with the limit values y, y, represented for the quality value fp. The lower diagram shows the input quantities ui and u2. One recognizes the good approximation of the setpoint specification by the designed controller.
    The method is preferably performed on a computer programmed to implement the method 1s. Alternatively, a computer cluster can also be used to divide the required computing power among several computers.
    The determined non-linear regulator could then be e.g. be implemented in a control unit to determine the manipulated variables in real time, depending on the input variables, in order to regulate the nonlinear process. From the determined nonlinear controller 1 or the determined parameter set KPID (k), however, characteristic diagrams for the controller parameters Kp, TN and Tv (or do, di) can also be determined or stored as a function of the input variables Example in the Automotive Environment with Control Units 11 of Internal Combustion Engines 12 in Vehicles 10 Use (FIG. 12), where the structure of the 25 characteristic diagrams 13 implemented in the control unit 11 is predetermined and in the course of the calibration the characteristic diagrams 13 are to be filled with data Regulator structure of the control units 11 maps are used for the controller parameters, which are usually dependent on load (injection quantity of the fuel or torque T) and speed n.Also, other variables can be detected with sensors 14 and supplied to the control unit -30 dell-based fully automatic generation of these maps 13. With this map-based controller structure For example, the boost pressure is regulated by means of turbocharger and / or exhaust gas recirculation. For this purpose, the nonlinear controller for the non-linear path can be determined. With the aid of the controller, the wanted maps 13 of the controller para- * 8- IMUHMHl WKM 102013/50129
    AV-3524 AT meter Kp, TN and T. In Figs. FIGS. 9 to 11 show, by way of example, the characteristic maps 13 determined in this way for the controller parameters Kp, TN and Tv as a function of the input variables of torque T and rotational speed n. Alternatively, maps could be created for the parameters d0, di, d2. 5 * -9-
    权利要求:
    Claims (6)
    [1]
    A method for designing a nonlinear controller (1) for non-linear processes modeled in the form of a local model network (LNN), characterized in that the parameter set (KplD (k)) of the controller (1) be determined by means of an optimization with a multi-criteria evolutionary algorithm in which in each evolution step, a plurality of parameter sets {Km (k)), each representing a possible solution of the optimization, are determined and for each parameter set (KPJD (k)) at least two quality values (fj) are determined and the quality values <fs) are optimized by the multicriteria 10 evolutionary algorithm.
    [2]
    2. The method according to claim 1, characterized in that a quality value (fs) for the stability of the control loop and a quality value (fp) for the behavior of the control loop is determined and these Quaiitätswerte be optimized.
    [3]
    3. The method according to claim 1 or 2, characterized in that the quality value 15 (fs) for the stability of the control loop with a Lyapunovkriterium with a decay rate (a) is determined, being used as the quality value (fs), the decay rate (a).
    [4]
    Method according to one of Claims 1 to 3, characterized in that a setpoint signal (wp) and a permissible tolerance range of the output variable (j> (jfc>) are defined around the setpoint signal (wp) and as a quality value (fp) for the behavior a measure is used to maintain the tolerance range of the control loop.
    [5]
    5. The method according to any one of claims 1 to 4, characterized in that as a result of optimization a Pareto front (P) with possible optimal parameter sets (Kpm (k)) is determined and of which a parameter set {KPID (k)) is selected as a solution ,
    [6]
    6. Method according to one of claims 1 to 5, characterized in that, for a parameter {d0, di, d2, Kp, TN1, Tv) of the parameter set (KPID (k)), a characteristic field is created as a function of the input variables , -10-
    类似技术:
    公开号 | 公开日 | 专利标题
    AT512251B1|2014-08-15|Method of designing a nonlinear controller for non-linear processes
    DE602005002839T2|2008-05-08|PROCESSING OF INDUSTRIAL PRODUCTION PROCESSES
    DE10127788B4|2014-12-24|Integrated optimal model predictive control in a process control system
    EP2106576B1|2012-04-04|Method for the computer-assisted control and/or regulation of a technical system
    EP2108139B1|2010-06-30|Method for the computer-aided control and/or regulation of a technical system, particularly a gas turbine
    EP2185980B1|2013-07-31|Method for computer-aided control and/or regulation using neural networks
    EP3062176B1|2019-07-03|Method for adjusting a control loop, device for generating control parameters and control device
    DE102007001026B4|2008-09-04|Method for computer-aided control and / or regulation of a technical system
    EP2135140A2|2009-12-23|Method for computer-supported control and/or regulation of a technical system
    DE102008020379A1|2009-10-29|Method for computer-aided control and / or regulation of a technical system
    DE10341764A1|2004-07-15|Integrated model prediction control and optimization within a process control system
    EP2411735B1|2014-02-26|Method and device for regulating the temperature of steam for a steam power plant
    DE102011012710A1|2011-09-08|Fast identification and generation of process models
    DE102019126310A1|2020-06-04|System and method for optimizing the combustion of a boiler
    DE102016121338A1|2017-05-24|Time-discrete, rate-based, model-predictive control method for air path control of an internal combustion engine
    DE19547496C2|2003-04-17|Process for regulating internal combustion engines
    DE102005026441B4|2009-11-12|Method for adapting the pilot control map of a volume-flow-controlled diesel common rail pump
    DE102019126286A1|2020-06-04|System and method for optimizing the combustion of a boiler
    EP2780557B1|2017-06-21|Method and device for controlling a temperature of steam for a steam power plant
    DE4204047A1|1993-08-19|METHOD FOR POSITIONING AN ACTUATOR IN A MOTOR VEHICLE
    EP3376026B1|2021-12-15|Method for controlling the power output of a wind farm and corresponding wind farm
    EP0473914B1|1995-06-28|Control system for an actuator in an internal combustion engine
    DE102018213177A1|2020-02-13|Method for regulating the output of the internal combustion engine of a motor vehicle
    AT522290B1|2021-01-15|Method and control unit for controlling a non-linear technical process
    DE19510343C2|1997-05-28|Process for sequential feedforward control of a process
    同族专利:
    公开号 | 公开日
    JP2016510151A|2016-04-04|
    EP2962161A1|2016-01-06|
    AT512251A3|2014-03-15|
    US10466659B2|2019-11-05|
    EP2962161B1|2017-07-19|
    JP6285466B2|2018-02-28|
    AT512251B1|2014-08-15|
    US20160011571A1|2016-01-14|
    WO2014131661A1|2014-09-04|
    引用文献:
    公开号 | 申请日 | 公开日 | 申请人 | 专利标题
    DE102016205780A1|2016-04-07|2017-10-12|Volkswagen Aktiengesellschaft|Method and device for adjusting a controller of a vehicle and control system for a vehicle|
    DE102020114436A1|2020-05-29|2021-12-02|Dr. Ing. H.C. F. Porsche Aktiengesellschaft|System and method for the autonomous calibration of a control unit for the wheel control of a motor vehicle|US3618025A|1968-09-20|1971-11-02|Nippon Electric Co|Pulse phase control apparatus for pulse communications systems|
    DE3618025A1|1986-05-28|1987-12-03|Aachener Forsch Regelungstech|Device for determining the control precision and the stability of control loops from measurement values and parameters of the control loop|
    US6415272B1|1998-10-22|2002-07-02|Yamaha Hatsudoki Kabushiki Kaisha|System for intelligent control based on soft computing|
    DE10005424A1|2000-02-08|2001-08-09|Bosch Gmbh Robert|Control circuit for fitting to a controllable electro-magnetic valve for a motor vehicle braking installation feeds actual control range values like deceleration, wheel slippage or vehicle speed back into preset tolerance band|
    JP2002163005A|2000-11-29|2002-06-07|Nikon Corp|Method of designing control system, control system, method of regulating control system, and method for exposure|
    JP2005285090A|2003-12-24|2005-10-13|Yamaha Motor Co Ltd|Multiobjective optimization apparatus, method, and program|
    US7398257B2|2003-12-24|2008-07-08|Yamaha Hatsudoki Kabushiki Kaisha|Multiobjective optimization apparatus, multiobjective optimization method and multiobjective optimization program|
    AT7710U3|2004-06-30|2006-07-15|Avl List Gmbh|METHOD FOR CREATING A MODEL OF A CONTROL SIZE FOR NONLINEAR, STATIONARY REAL SYSTEMS, FOR EXAMPLE, COMBUSTION ENGINES OR SUBSYSTEMS THEREOF|
    DE202006002525U1|2006-02-17|2007-07-05|Brose Fahrzeugteile Gmbh & Co. Kommanditgesellschaft, Coburg|Anti-jamming protection of a motor-driven adjusting device for an adjusting device|
    JP5073268B2|2006-11-01|2012-11-14|富士重工業株式会社|Automatic adjustment device for control parameters|
    JP5098736B2|2008-03-25|2012-12-12|株式会社明電舎|Vehicle speed control device|
    CN101488010A|2009-03-06|2009-07-22|北京理工大学|Essentially nonlinear compensation controller of servo system|
    US8760461B2|2009-05-13|2014-06-24|Stmicroelectronics, Inc.|Device, system, and method for wide gamut color space support|
    TWI413006B|2010-06-01|2013-10-21|Nat Univ Tsing Hua|Method for buliding adaptive soft sensor|
    CN101900991A|2010-07-02|2010-12-01|长沙理工大学|Composite PID neural network control method based on nonlinear dynamic factor|
    AT511577B1|2011-05-31|2015-05-15|Avl List Gmbh|MACHINE IMPLEMENTED METHOD FOR OBTAINING DATA FROM A NON-LINEAR DYNAMIC ESTATE SYSTEM DURING A TEST RUN|CN104199437A|2014-08-15|2014-12-10|上海交通大学|Fractional order PI^lambdaD^mu controller parameter optimization method based on region pole index|
    CN107894714A|2017-11-10|2018-04-10|大连民族大学|The adaptive sliding-mode observer method of nonlinear system|
    JP6933585B2|2018-01-12|2021-09-08|株式会社豊田中央研究所|Information processing device, information processing method, computer program, control device|
    US10787984B2|2018-08-08|2020-09-29|Caterpillar Inc.|Power system optimization calibration|
    US10787981B2|2018-08-08|2020-09-29|Caterpillar Inc.|Power system optimization|
    CN110262248B|2019-07-08|2021-05-11|南京航空航天大学|Fault robust self-adaptive reconstruction method for micro gas turbine|
    DE102019216312A1|2019-10-23|2021-04-29|Robert Bosch Gmbh|Prediction and / or assessment of the emission quantity of substances in the exhaust gas of internal combustion engines|
    CN111077766A|2019-12-31|2020-04-28|福建（泉州）哈工大工程技术研究院|Control algorithm of random moving target tracking trolley based on nonlinear PID|
    法律状态:
    优先权:
    申请号 | 申请日 | 专利标题
    ATA50129/2013A|AT512251B1|2013-02-28|2013-02-28|Method of designing a nonlinear controller for non-linear processes|ATA50129/2013A| AT512251B1|2013-02-28|2013-02-28|Method of designing a nonlinear controller for non-linear processes|
    EP14707677.2A| EP2962161B1|2013-02-28|2014-02-19|Method for designing a nonlinear controller for nonlinear processes|
    US14/771,036| US10466659B2|2013-02-28|2014-02-19|Method for designing a non-linear controller for non-linear processes|
    PCT/EP2014/053180| WO2014131661A1|2013-02-28|2014-02-19|Method for designing a nonlinear controller for nonlinear processes|
    JP2015559467A| JP6285466B2|2013-02-28|2014-02-19|Design method of nonlinear controller for nonlinear process|
    [返回顶部]