Running a Simulation ~Using the Scenario List-4

•关于外部输入位号的仿真,当前(PCTP 2009 Build 1)版本的一个重要限制是值中缺少原始的“存储器”。例如,如果已经开发了用于检查外部输入位号覆盖功能的计算,则在激活位号之后的仿真中,仿真器将永远不会恢复到原始值。

Running a Simulation ~Using the Scenario List-4_第1张图片

•由于实现方式不同,离线仿真(PCTP)行为与在线EXASMOC嵌入行为是不同的。当处理递归计算时,经常会遇到这个问题。例如,令[A] = [A] +1,并假设[A]的初始值= 10。则在每个内核执行时,离线仿真会增加[A]的值,即11,12,13 ...,然而,在线EXASMOC控制器将保持[A]的值恒定,即11,11,11,...
•在方案构建过程中必须指定许多重要的可在线修改的参数。否则控制器构建时将使用“默认”值。这些参数为:
 CV优先级
 CV设定点跟踪滤波器
 DV滤波器时间常数
 POV脉冲因子
•从PCTP 2010 Build 1版本开始,用户可以为白噪声指定随机数生成器种子。种子值可以是零或任何正数。默认种子值为0,其与先前版本中的行为一样:当每个仿真运行时种子将在内部改变,由一个仿真跳至另一个仿真时将生成不同的噪声序列。如果用户指定了一个非零值的种子,则除非用户更改种子,否则在运行时噪声序列是相同的(可重复的)。此功能对于比较不同整定参数时,控制器对相同噪声干扰响应性能非常有用。
•注意:即使噪声种子参数在通用选项卡中的每个干扰和子控制器选项卡中的MV出现,也不能为单个干扰或MV指定噪声种子。改变任何给定仿真方案的噪声种子参数将影响干扰和MVs的所有噪声干扰。如果用户要求从一个运行转到另一个运行时噪声序列自动改变,则用户必须确保所有干扰和MVs的噪声种子参数值被设置为零(0)。


原文:
• An important limitation of the current (PCTP 2009 Build 1) release with regards to the simulation of external input tags is the lack of original “memory” in the values. For example, if a calculation has been developed to check the override function of an external input tag, in simulation once the tag has been activated, the simulator will never revert back to the original value.
• Due to different implementations, the offline simulation (PCTP) behavior is NOT identical to the online EXASMOC embedding behavior. The problem is mostly observed when dealing with recursive calculations. For example, let [A]=[A]+1 and assume the initial value of [A] =10. The offline simulation increases the value of [A] at each kernel execution, i.e. 11, 12, 13, ..., however, the online EXASMOC controller keeps the value of [A] constant, i.e. 11, 11, 11, ...
•There are a number of important online modifiable parameters that MUST be specified during the scenario build process. Otherwise ‘default’ values are used in the controller build. The parameters are:
 CV priorities
 CV setpoint tracking filters
 DV filter time constants
 POV impulse factors
 •Starting with the PCTP 2010 Build 1 release, users can specify the seed of the random number generator for the white noise. The seed value can be zero or any positive number. The default seed value is 0, which has the same behavior as in previous releases: the seed is changed internally for each simulation run, generating different noise sequences from simulation run to simulation run. If users specify a non-zero seed, the noise sequence is the same (repeatable) from run to run until users change the seed. This functionality is very useful to compare controller performance with different tunings reacting to the same noise disturbance.
•NOTE: Even though the Noise Seed parameter appears with every disturbance in the general tab and MV in the sub-controller tabs one cannot specify noise seeds for individual disturbances or MVs. Changing any Noise Seed parameter for the given Simulation Scenario will affect all noise disturbances for the disturbances and MVs. If the user requires the noise sequence to automatically change from run to run then the user must ensure that the Noise Seed parameter values for ALL the disturbances and MVs are set to zero (0).


2010.10.16

你可能感兴趣的:(Running a Simulation ~Using the Scenario List-4)