I am to provide a graph that indicate deviation from a set point.
There are five significant points on the Y axis
* Upper scale limit.
* Upper alarm limit
* The desired point (set point).
* Lower alarm limit.
* Lower scale limit.
These can be, for instance
High Scale : 100
High Alarm: 80
Set Point : 50
Low Alarm: 20
Low Scale: 0.
They can also be,
High Scale : 100
High Alarm: 80
Set Point : 76
Low Alarm: 10
Low Scale: 0.
These two sets of axis settings should draw the exact same Y Axis. Even in the case where the distance from High Alarm to set point is just 4 units and distance from Set Point to Low Alarm is 56 units, they should take equal amount of space of the drawing area. A linear dataset would draw a gently sloping curve from Low Alarm to Set Point, but it would be very steep from Set Point to High Alarm as the whole distance would be covered in just 4 units.
How do I achieve this? I consider using four stacked curves that would each cover one area of the curve on the scale for its area so that it would only APPEAR to be a single curve, but would actually be four different curves. This, however functionally adequate, does not draw a continuous curve. Any better suggestions?
5 Answers, 1 is accepted
It is hard for us to provide you a proper help based on your description. Could you please send us a sample image showing what do you want to achieve?
Greetings,
Evgenia
the Telerik team
Rank 2
You can open a formal Support thread and send it or you can use a third-party website for files upload like 4shared.com for example.
All the best,
Evgenia
the Telerik team
Rank 2
Ok, here we go:
I am developing a solution for a large screen display on a processing plant.
The graphical layout for elements in this solution are specifically designed to draw attention to deviations from norm (ie, the set point).
Each value has an associated upper and lower alarm limit (High and Low), which will trigger alarms to which the operator should respond to avoid escalation and restore optimum production.
In addition, each value has another set of alarm limits (HighHigh and LowLow) that, when transcended, will cause operational consequence (ie, shutdown or revert to safe mode).
The end result looks something like this:
http://i.imgur.com/vcozn.png
However, since the various plots have different alarm limits that has no predefined relative size to the other limits, the end result quickly looks like this
http://i.imgur.com/7DDyr.png
It is not easy for the operator to spot deviations because nothing is aligned. Therefore, it was decided to force this graphical layout regardless of the actual areas.
http://i.imgur.com/8Y6Pi.png
Of course, this creates a sectioned Y Axis, so that when plotting a linearly increasing curve, which in a orthogonal plot area would create a straight line, it may end up looking like this.
http://i.imgur.com/qSGFE.png
So my question is how do I implement this display. My thought is to stack four distinct plots in a stackpanel, so that the one line is actually four different plots. Two things that sort of get in the way:
* Most of the drawing will happen in the boundary between two plots (around the set point). I worry that this will not look like a smooth line.
* If two adjacent data points are placed in two different plot areas, the line between them will be disconnected unless I create a "helping" point at the alarm limit.
That will not be a good solution because the alarm limits are dynamically changed from the process (they will be databound to values continuously being updated).
Currently RadChart doesn't support trend lines, financial indicators or alarm limits. However we have left our customers the opportunity to use it in various scenarios. You may find our Custom Sampling Function topic helpful. It demonstrates how to create a Standard Deviation function for the Range of numbers 0 - 1000 and creates ScatterSeriesDefinition with the result from this function.
Feel free to customize the source code from the topic so that it applies your scenario needs.
Greetings,
Evgenia
the Telerik team