You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: docs/CROWDIN/zh_CN/AdvancedOptions/FullClosedLoop.md
+38-38Lines changed: 38 additions & 38 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -120,82 +120,82 @@ IEEE Control Systems Magazine, ResearchGate [The Artificial Pancreas and Meal Co
120
120
121
121
在**HCL**中实施了关于闭环自动推注量的安全限制。
122
122
123
-
**FCL** loopers no longer need to give a sizable bolus around meal start. The impact of this means that restrictions in size limits for **SMBs** must be widened to make the loop capable of delivering large enough **SMBs**.
If you are operating with **AAPS** in the Master release, it is suggested **AAPS**' Preferences are set up with the maximum allowed **SMB** size so that **FCL** can give (maxUAMSMBBasalMinutes=120, i.e. 2 hours worth of basal at that daytime).
If your basal rate is very low, the resulting **SMB** limits might be too low to allow sufficient control to tackle postprandial **BG** rises. One possible solution is to avoid diets that cause strong **BG** spikes and later switches to a **AAPS** dev variant that offers a new parameter in **SMB** delivery settings: smb_max_range_extension. This will expand the standard maximum of 2 hours worth of basal by a factor of >1. (Additionally, the default 50% **SMB**delivery ratio might be elevated in dev. variants).
**Follow the instructions to enable AAPS to mimic your bolussing via a couple of SMBs**.
129
+
**按照说明使AAPS通过若干SMB模拟您的推注**。
130
130
131
-
Check the **SMB**tab periodicallu to see whether your **SMBs**are allowed to be sufficient enough to deliver the required insulin needed for the loop around meal starts.
If not, your tuning efforts will sometimes come to nothing!
133
+
否则,您的调校努力有时会白费!
134
134
135
135
136
136
```{admonition} Boosting **ISF** can become dangerous
137
137
:class: danger
138
138
139
-
Carefully observe/analyse the **SMB** sizes shortly after your meal commences. Tune in steps, and do not vary more than 1 or 2 parameters at a time.
139
+
密切观察/分析餐后**SMB**量。 逐步调校,每次最多调整1-2个参数。
140
140
141
-
Your **AAPS'** setting must be sufficiently set up to cope with your (!) variety of meals.
141
+
您的**AAPS**设置必须充分适应您的各种餐食。
142
142
```
143
143
144
-
## Meal detection/your Automations for boosting
144
+
## 进餐检测/您的增强自动化规则
145
145
146
-
For successful **FCL**, **ISF** is the key tuning parameter. When utilising **AAPS** Master + **Automations**, a **> 100% profile change must automatically be triggered upon meal recognition** (via glucose deltas), and provide the sharpened **ISF**.
-Step 1 - review the **ISF** applicable for this meal time hour within the **Profile**, and see whether e.g. Autosens suggest a modification that takes care of the current (last few hours’) insulin sensitivity status of the body..
151
-
-Step 2 - apply a factor (1/Profile%, as set in **Automation**) to boost **ISF**.
152
-
-Step 3 - check that the suggested **ISF**falls within set safety limits.
-In your **Automation** list, you can tick the check-mark (to the left of each field) OFF => This de-activates that **Automation**. For instance you can do this for all breakfast related **FCL****Automations** to go to **HCL** for breakfast(s).
-For each **Automation** rule you can tick the box for User action => then the defined Actions will not automatically be executed when Conditions apply. Rather, the **AAPS**main screen will alert you whenever your **FCL**would automatically give a **SMB**. You have the opportunity then to say ‘yes’ or ‘no’. This is extremely useful in your tuning phase.
This feature can be useful for certain situations like “foot to floor” syndrome whher there is a sudden rise in **BG** when getting up in the morning, but the user wants to prevent a fully automatic “breakfast started” response.
The section below provides guidance how to bundle **Automation’s** Conditions and how to tackle situations in which the **AAPS** should increase (or decrease) insulin delivery. As **ISF**cannot directly be tuned, raising **Profile Percentage** over 100% will do the same for our purposes.
The key to successful **FCL****at the beginning of BG increases from meals, very large automatic SMBS must be given by the loop as quickly as possible** “to catch up” with the required **IOB** needed (compare with your typical administered bolus for similar meal in h**HCL**!)
To achieve this, data from your **HCL** should be analysed to determine which **deltas** might be not meal –related and those delta which might be.
170
+
为此,需分析**HCL**数据以确定哪些变化率(delta)可能与进餐无关,哪些相关。
171
171
172
-
-As you can define the **Automation** within a pre-defined time-window, you need only to analyse there.
173
-
-If you do very different kinds of meals (e.g. a rather high carb breakfast, but low carb lunch) you can choose to do two different (sets of) **Automations** for each of the time slots.
174
-
-Exclude the nights if you see occasional jumps from a compression lows
175
-
-Usually, just using the delta of the past 5 minutes suffices.
176
-
-But you can also make use one of the average deltas. By comparing the deltas in the conditions of your **Automations** you could even define actions of different aggressiveness depending on whether the **BG** rises in an accelerated way or not.
A**CGM**with patchy data puts the user in a bad spot because, to be on the safe side, you need to „sandbag“ your definition which delta is surely a sign of a started meal. This means:
-**FCL**loses additional time, resulting in higher **BG**peaks and lower %**TIR**;
184
-
-you cannot use an earlier or smaller delta which could trigger, also without a meal, the **SMBs** that are supposed to make up for a user bolus in **FCL**.
Furthermore, first rises after a meal are characterized by **low iob** present. With that in mind, an Automation(#1) for a dinner might look like this:
If Conditions apply, **AAPS** would give 1 or 2 **SMBs** in the next 12 minutes, using a boosted **ISF** according to the set elevated **Profile Percentage** (in the example, a 30% boost of insulinReq). As long as these Conditions apply, the **Automation** rule extends by another 12 minutes. A low carb meal might have slower **BG** rise characteristics. It would benefit from another Automation (#2) that kicks in at lower delta, and gives a weaker insulin boost.
The same **Automation** probably will kick in also in higher carb meals, once the steep rise as defined in Automation#1 is over.
196
+
一旦自动化#1中定义的急剧上升结束,同样的**自动化**可能也会在碳水化合物含量较高的膳食中启动。
197
197
198
-
You need to “stage” these two (+ maybe a third) **Automations** to fit with what you see in your meal (variety) => Setting appropriate jump sizes, **iob**criteria, and amplifications will be an iterative tuning process. Also, if you include appropriate time slots in the Conditions, you can easy do different Automations for your different daily meals times (breakfast, lunch, dinner).
Note that, still in the rise phase (!), the "overflow" of **iob** must be blocked so that the late effects of the **insulin** (the "**tail**" after 3-5 hours) will not exceed the braking capacity of the loop through zero-temping (“taking away” basal, to reduce hypo risk).
0 commit comments