在這種情況下躏吊，我們模擬一個影響底部質(zhì)量的斜坡干擾。這一干擾會導(dǎo)致底部質(zhì)量瞬間超過上限誊锭。如果該計算打開，底部質(zhì)量CV違反上限后弥锄，我們預(yù)計有以下行動：
?計算3將頂部質(zhì)量CV的priority（優(yōu)先級）設(shè)定為10丧靡；
?計算5將reflux（回流）的max move size（最大動作幅度）設(shè)定為150；
?計算6將再沸器負(fù)荷的max move size（最大動作幅度）設(shè)定為1.5籽暇；
當(dāng)?shù)撞抠|(zhì)量CV值在范圍之內(nèi)時温治，我們預(yù)計：
?計算3將把頂部質(zhì)量CV的priority（優(yōu)先級）恢復(fù)到1；
?計算5將reflux（回流）的max move size（最大動作幅度）恢復(fù)到100戒悠；
?計算6將再沸器負(fù)荷的max move size（最大動作幅度）恢復(fù)到1.5熬荆。
若要查看計算結(jié)果，當(dāng)控制器拒絕了仿真底部質(zhì)量干擾時绸狐，我們將內(nèi)核調(diào)試文件打開并保持25步卤恳。
逐步運行仿真直到第15步。如下面屏幕截圖所示寒矿，在這一點突琳，我們可以通過單擊“bug”按鈕打開調(diào)試文件：
?我們注意到的第一點是，盡管底部質(zhì)量在范圍內(nèi)符相，EF tracking filter（EF跟蹤濾波）值依舊改為300秒拆融。注意在原始方案中我們將EF tracking filter設(shè)定為0.0。通過確認(rèn)EF tracking filter至少是5mins啊终，計算7確認(rèn)該過程不受到過于激烈的動作镜豹。
?對于內(nèi)核后計算，我們注意到底部質(zhì)量CV的穩(wěn)態(tài)約束行為表示了值101孕索。值101意味著由于經(jīng)濟(jì)函數(shù)原因（即經(jīng)濟(jì)優(yōu)化將一個變量推向約束）逛艰，在穩(wěn)態(tài)時約束是活動的。
當(dāng)?shù)撞抠|(zhì)量違反了上限搞旭，計算將分別把回流的最大動作和再沸器負(fù)荷設(shè)置為150和1.5散怖。
當(dāng)?shù)撞抠|(zhì)量違反了上限時，頂部質(zhì)量CV的優(yōu)先級將降為10肄渗。
一旦控制器拒絕了干擾镇眷，并將底部質(zhì)量帶回限制內(nèi)，計算將把變量恢復(fù)回標(biāo)稱值翎嫡。
仿真2
我們將觀察的下一個仿真稱為“Disconnected POVs”欠动。其仿真參數(shù)如下所示。
在第6步，我們預(yù)計計算1將移去底部質(zhì)量CV具伍。在第8步翅雏，我們預(yù)計由于計算2的原因，頂部質(zhì)量CV將被移去人芽。隨后由于計算4望几，子控制器將切換到非活動狀態(tài)。鑒于控制器只有一個子控制器萤厅，我們預(yù)計在第8步控制器將變成standby（待機(jī)）狀態(tài)橄抹。
原文：
In this case, we simulate a ramp disturbance that affects the bottom quality. This disturbance causes the bottom quality to exceed its upper limit in the transient. If the calculations are turned ON, we expect the following when the bottom quality CV violates its upper limit
? Calculation 3 sets the priority for the top quality CV to 10
? Calculation 5 sets the max move size for reflux to 150
? Calculation 6 sets the max move size for reboiler load to 1.5
When the bottom quality CV is within limits, we expect
? Calculation 3 reverts the priority for the top quality CV back to 1
? Calculation 5 reverts the max move size for reflux back to 100
? Calculation 6 reverts the max move size for reboiler load back to 1.5
To view the results for the calculations, we turn the kernel debug file ON for 25 steps while the controller rejects the simulated bottom quality disturbance.
Run the simulation stepwise until the 15th step. At this point, we may open the debug file by clicking on the “bug” button as shown in the screenshot below
?The first aspect we notice is that even though the bottom quality is within limits, the EF tracking filter value has changed to 300 seconds. Notice in the original scenario that we left the EF tracking filter at 0.0. By ensuring that the EF tracking filter is at least 5 minutes, calculation 7 ensures that the process is not subject to overly aggressive moves.
? For the post-kernel calculations, we notice that the steady-state constraint activity for the bottom quality CV indicates a value of 101. A value of 101 implies that the constraint is active at steady state due to economics (i.e. due to the economic optimization pushing the variable against a constraint).
When the bottom quality violates its upper limit, the calculations set the max moves for the reflux and reboiler load to 150 and 1.5 respectively.
The priority for the top quality CV is lowered to 10 when the bottom quality violates its upper limit.
Once the controller rejects the disturbance and brings the bottom quality within limits, the calculations revert the variables back to their nominal values.
**Simulation 2 **
The next simulation that we observe is called “Disconnected POVs.” The simulation parameters are shown below.
At step 6, we expect calculation 1 to remove the bottom quality CV. At step 8, we expect the top quality CV to be removed as a consequence of calculation 2. Subsequently, the sub-controller is turned inactive due to calculation 4. Since the controller has just the one sub-controller, we expect the controller to shed to standby at step 8.
2016.6.16
