Main Content

MPC Prediction Models

Model predictive controllers use plant, disturbance, and noise models for prediction and state estimation. The different signal types are described inMPC Signal Types。MPC控制器中使用的模型结构出现在下面的图中。

植物模型

您可以以以下线性时间不变(LTI)格式之一指定植物模型:

  • 数字LTI模型 - 传输函数(TF),国家空间(SS),零极生(zpk

  • 确定的模型(需要系统标识工具箱™) -IDSS((system Identification Toolbox),,,,IDTF((system Identification Toolbox),,,,IDPROC((system Identification Toolbox),,,,andIDpoly((system Identification Toolbox)

MPC控制器使用具有无量纲输入和输出变量的离散时间,无延迟的状态空间系统执行所有估计和优化计算。因此,当您在MPC控制器中指定植物模型时,该软件会在需要时执行以下操作:

  1. Conversion to state space — TheSS命令将提供的模型转换为LTI状态空间模型。

  2. d一世scretization or resampling — If the model sample time differs from the MPC controller sample time (defined in theTS属性),以下一个发生:

    • 如果模型是连续的时间,C2D命令使用控制器示例时间将其转换为离散的LTI对象。

    • 如果模型是离散的时间,d2dcommand resamples it to generate a discrete-time LTI object using the controller sample time.

  3. 延迟删除 - 如果离散时间模型包括任何输入,输出或内部延迟,则absorbDelaycommand replaces them with the appropriate number of poles atz= 0,增加离散状态的总数。这inputdelay,,,,OutputDelay,,,,andInternalDelayproperties of the resulting state-space model are all zero.

  4. Conversion to dimensionless input and output variables — The MPC controller enables you to specify a scale factor for each plant input and output variable. If you do not specify scale factors, they default to1。该软件将工厂输入和输出变量转换为无量纲形式,如下所示:

    X p (( k + 1 = 一个 p X p (( k + b s 一世 p (( k y p (( k = s o - 1 C X p (( k + s o - 1 d s 一世 p (( k

    在哪里一个p,,,,b,,,,C,,,,andd是步骤3的恒定零延迟状态空间矩阵,也是:

    • s一世一世s a diagonal matrix of input scale factors in engineering units.

    • so是工程单元中输出量表因子的对角线矩阵。

    • Xp一世s the state vector from step 3 in engineering units (including any absorbed delay states). No scaling is performed on state variables.

    • p是无量纲植物输入变量的向量,包括操纵变量,测量的干扰和未得到的输入干扰。

    • yp是无量纲植物输出变量的向量。

    这resulting plant model has the following equivalent form:

    X p (( k + 1 = 一个 p X p (( k + b p (( k + b p v v (( k + b p d d (( k y p (( k = C p X p (( k + d p (( k + d p v v (( k + d p d d (( k

    这里, C p = s o - 1 C ,,,,bpu,,,,bPV,,,,andbpdare the corresponding columns ofBS一世。一个lso,dpu,,,,dPV,,,,anddpdare the corresponding columns of s o - 1 d s 一世 。最后,((k),v((k), 和d((k)are the dimensionless manipulated variables, measured disturbances, and unmeasured input disturbances, respectively.

    MPC控制器执行限制dpu= 0,,,,which means that the controller does not allow direct feedthrough from any manipulated variable to any plant output.

Input Disturbance Model

If your plant model includes unmeasured input disturbances,d((k),输入干扰模型指定了信号类型和特征d((k)。看Controller State Estimation有关模型的更多信息。

Getindistcommand provides access to the model in use.

输入干扰模型是影响以下控制器性能属性的关键因素:

  • dynamic response to apparent disturbances — The character of the controller response when the measured plant output deviates from its predicted trajectory, due to an unknown disturbance or modeling error.

  • 渐近拒绝持续干扰 - 如果干扰模型预测持续的干扰,控制器的调整将继续进行,直到植物产出恢复到所需的轨迹,并模仿经典的积分反馈控制器。

You can provide the input disturbance model as an LTI state-space (SS), 转换功能 (TF),or zero-pole-gain (zpk)object usingsetindist。MPC控制器将输入干扰模型转换为离散时间,无延迟的LTI状态空间系统,使用与转换的相同步骤plant model。结果是:

X 一世 d (( k + 1 = 一个 一世 d X 一世 d (( k + b 一世 d w 一世 d (( k d (( k = C 一世 d X 一世 d (( k + d 一世 d w 一世 d (( k

在哪里一个ID,,,,bID,,,,CID,,,,anddIDare constant state-space matrices, and:

  • XID((k)是nXID≥ 0输入干扰模型状态。

  • dk((k)是ndd一世mensionless unmeasured input disturbances.

  • wID((k)是nID≥ 1无量纲的白噪声输入,假定为零均值和单位方差。

If you do not provide an input disturbance model, then the controller uses a default model, which has integrators with dimensionless unity gain added to its outputs. An integrator is added for each unmeasured input disturbance, unless doing so would cause a violation of state observability. In this case, a static system with dimensionless unity gain is used instead.

输出干扰模型

这o你tput disturbance model is a special case of the more general input disturbance model. Its output,yod((k),一世s directly added to the plant output rather than affecting the plant states. The output disturbance model specifies the signal type and characteristics ofyod((k), 和一世t is often used in practice. SeeController State Estimationfor more details about the model.

getoutdist命令提供输出扰动model in use.

You can specify a custom output disturbance model as an LTI state-space (SS), 转换功能 (TF),or zero-pole-gain (zpk)object usingsetoutdist。使用与plant model,,,,the MPC controller converts the specified output disturbance model to a discrete-time, delay-free, LTI state-space system. The result is:

X o d (( k + 1 = 一个 o d X o d (( k + b o d w o d (( k y o d (( k = C o d X o d (( k + d o d w o d (( k

在哪里一个od,,,,bod,,,,Cod,,,,anddodare constant state-space matrices, and:

  • Xod((k)是nXOD≥ 1o你tput disturbance model states.

  • yod((k)是ny无量纲输出干扰将添加到无量纲的植物输出中。

  • wod((k)是nod无量纲的白噪声输入,假定为零均值和单位方差。

If you do not specify an output disturbance model, then the controller uses a default model, which has integrators with dimensionless unity gain added to some or all of its outputs. These integrators are added according to the following rules:

  • 对于未衡量的植物产量,没有估计估计不添加积分器的干扰。

  • 为每个测得的输出添加一个集成器,以减小输出权重的顺序。

    • For time-varying weights, the sum of the absolute values over time is considered for each output channel.

    • 对于相等的输出权重,遵循输出向量内的顺序。

  • For each measured output, an integrator is not added if doing so would cause a violation of state observability. Instead, a gain with a value of zero is used instead.

如果有输入干扰模型,则控制器在构造默认输出干扰模型之前将任何默认集成器添加到该模型中。

测量噪声模型

一个控制器设计目标是区分需要响应的干扰与测量噪声,这应该被忽略。测量噪声模型指定了预期的噪声类型和特征。看Controller State Estimationfor more details about the model.

使用与plant model,,,,the MPC controller converts the measurement noise model to a discrete-time, delay-free, LTI state-space system. The result is:

X n (( k + 1 = 一个 n X n (( k + b n w n (( k y n (( k = C n X n (( k + d n w n (( k

这里,一个n,,,,bn,,,,Cn,,,,anddn是恒定的状态空间矩阵,:

  • Xn((k)是nxn≥ 0no一世se model states.

  • yn((k)是nYMd一世mensionless noise signals to be added to the dimensionless measured plant outputs.

  • wn((k)是nn≥ 1无量纲的白噪声输入,假定为零均值和单位方差。

如果您不提供噪声模型,则默认值是统一静态增益:nxn= 0,,,,dn是一个nYM-by-nYM身份矩阵,一个n,,,,bn,,,,andCn是空的。

MPCcontroller object,mpcobj,属性mpcobj.model.noise提供对测量噪声模型的访问。

笔记

If the minimum eigenvalue of d n d n t 一世s less than 1x10–8,MPC控制器添加1x10–4to each diagonal element ofdn。this adjustment makes a successful default Kalman gain calculation more likely.

Related Topics