Name（名称）：模型源的名称；
Active Parametric Model（活跃参数模型）：指示所显示的拟合是来自FIR结果拟合（F2P）的参数模型还是来自数据（PAR）的参数模型拟合；
Input x Output（输入x输出）：选中单元格对应的输入和输出；
Gain（增益）：选中单元格对应的模型增益；
Model（模型）：模型类型 — 稳定或斜坡；
Fit Stat（拟合度）：每个拟合对应输入输出关系的拟合度；
Output Fit Stat（输出拟合度）：对应整个输出的拟合度。
注意：比较表中的拟合度数据可能不会总是恰当的，因为拟合度数据是基于预滤波信号计算的，这些信号对每个拟合可能不同（取决于高级选项设置）。 请参阅第42页的“拟合统计”。
仿真
仿真选项卡由拟合组的输出趋势和拟合组中拟合（为分析选择）输出预测组成。（参见模型预测一节）。每个拟合的性能可以根据预测与预测输出的匹配程度来确定。选中用于分析的输入趋势也可在预测窗口中找到。
分析时使用：
•仿真—仅基于模型确定性部分的输出预测。它可以被认为是开环预测，没有对测量内容做出任何校正。请参阅第39页的“确定性预测（仿真）”。
•误差—误差是实际输出和预测输出（仅基于模型的确定性部分）之间的差值。如果误差是随机分布的，则认为模型是好的。请参阅第40页的“确定性仿真（误差）”。
•预测—基于输出的确定性部分以及基于反馈的校正项的预测输出。校正项等效于测量输出与上一时刻预测输出之间差的函数。计算的失配的频率是预测时域。该预测模仿了模型预测控制中的在预测时域上的输出。请参见第41页的“仿真更新”。预测也可用于更好地可视化斜坡响应;激活预测模式将去除斜坡输出变量的预测漂移。从这个角度来说，其行为类似于差分标签的行为。

原文：
**Name: **Name of the model source.
**Active Parametric Model: **Indicates that the fit displayed comes from the Parametric model fit to the FIR results (F2P) or the Parametric model fit from data (PAR).
**Input x Output: **Input and Output corresponding to the selected cell.
Gain: Model gain corresponding to the selected cell.
**Model: **Model type – stable or ramp.
**Fit Stat: **Fit Stat corresponding to the input output relationship for each fit.
**Output Fit Stat: **Fit Stat corresponding to the entire output.
NOTE: Comparing fit stats in this table may not always be appropriate, because fit statistics are computed based on prefiltered signals which may be different for each fit (depending on the advanced option settings). Refer to “Fit Statistics” on page 42.
Simulation
The Simulation tab consists of trends of the outputs of the Fit Group and the output predictions from the fits (selected for analysis) in the Fit Group. (See section on Model Predictions). The performance of each fit can be determined on how closely a prediction matches the output being predicted. The trends of inputs selected for the analysis are also available on the predictions window.
Perform analysis using:
• Simulations – A prediction of the output based only on the deterministic part of the model. It can be considered the open-loop prediction without any corrections made with measurements knowledge. Refer to “Deterministic Prediction (Simulation)” on page 39.
• Errors –Errors are the difference between the actual output and the predicted output (based only on the deterministic part of the model). A model may be considered good if the errors are randomly distributed. Refer to “Deterministic Simulation (Error)” on page 40.
• Predictions – Prediction of the output based on the deterministic part of the output along with a correction term based on feedback. The correction term is equivalent to a function of the difference between the measured output and the predicted output at a time instant in the past. The frequency at which the mismatch is computed is the prediction horizon. This prediction mimics the prediction of outputs over the prediction horizon in model-predictive control. Refer to “Simulation with Updating” on page 41. Predictions can also be useful for better visualization of ramp responses; activating the prediction mode will remove drifting of ramp output variable prediction. In this way, its behavior is similar to that of the difference flag.

2016.11.27