How can the quality of SCADA data be accounted for with SCADA signals in WaterGEMS?

Product(s): WaterGEMS
Version(s): CONNECT Edition, V8i
Area: Modeling

Problem

How can the quality of SCADA data be accounted for with SCADA signals in WaterGEMS?

For example, if data is questionable, how is that handled?

Background

The quality of data collected might sometimes be compromised due to incorrect reporting, faulty equipment or general errors in reading data. Often such data when identified, is classified to understand which of the values to consider for reference.

The data is captured against signals created for model elements.

The signals can be presented in two formats;

1. One value per row
2. Multiple values per row

In the “one value per row” format, each row should consist the signal value against a time stamp. Here the quality of the data can also be represented (E.g. Good, Bad or Questionable). A typical example of such a kind of data can be seen below;

Time

Pump

Flow (L/s)

Quality

12-09-16 1:39

PMP-1

43.2

Bad

12-09-16 2:00

PMP-1

45

Good

12-09-16 3:00

PMP-1

52.3

Questionable

For the “multiple values per row”, each row can have multiple signal values corresponding to single time stamp. A typical example of such a kind of data can be seen below; the order of the columns is not important.When defining the data,When defining the data, the order of the columns is not important. Also, if any unused columns are present it won’t affect importing the database.

Time

Tank Level (m)

Pump Flow (L/s)

Junction Pressure (psi)

12-09-16 1:39

121.2

43.2

19.3

12-09-16 2:00

124.7

45

21.2

12-09-16 3:00

135.5

52.3

35


The quality of the data corresponds to a single value being good, bad or questionable. Hence, for multiple values the quality field cannot be used.

Solution

An “Is Questionable?” field in your SCADA source can be accounted for in the preview and the questionable recordings will be skipped over in the graph, but currently a graph of the SCADA element will always show all values

For this article we will work on signal data captured which contains the flow values for a pump PMP-1 at different time steps. The data contains the columns of time, pump, flow and quality to be mapped to the SCADA signal.

First, set up the signals from SCADA Signals in the SCADA Connect Simulator by defining the Database Source. To know how to import data by specifying a database format refer the Help section in WaterGEMS (Help > Modeling Capabilities > SCADAConnect Overview > SCADA Signals Setup > SCADA Signals – Database). Although the data can be imported from several* different databases we have selected Excel as an example for the sake of simplicity.

* Refer Help section for a list of acceptable database sources

Once the database is imported specify the following parameters in the Database Source dialog box;

Table Name: Sheet1$. (Specify the “Sheet” in Excel to import the data)

Source Format: One value per row (Select this option to view questionable data)

Signal Name Field: Pump. (Name of the signal)

Value Field: Flow. (Value to be imported for selected signal field)

Time Stamp Field: Time (hours). (Define the time step to be followed)

Questionable Field: Quality. (This is the field where the data quality is expressed)

Under options, select “Historical” data.

Note: The Quality field should contain the word “Good” as a single word or part of a string “Good Data” so that the software can identify the data category. If the data in the database source is different from what can be read by the software, you may need to manipulate the field in your SCADA system first to make the data readable as “Good Data”.

Once the fields are selected appropriately, the dialog box would look like this;

After the data is successfully imported it would look like this;

From the above image you can see that the signal value is the flow from PMP-1. The right most column “Questionable?” maps the questionable data. The values which are ticked are the values which are questionable in nature. By default, only the good data is displayed in the inbuilt graph below and the bad/questionable values are skipped. When the “Include Questionable Values” option is selected, the questionable values will be displayed in the graph.

Note:

  • It is up to your SCADA system to make the determination of whether a reading is questionable or not. In other words, WaterGEMS doesn’t do this for you; the field needs to exist already in the SCADA source, or you need to set up your SCADA system to produce such a field.
  • The questionable values contain all values which are not “Good”. Since we have “Bad” and “Questionable” as other available categories, both these values will be considered as questionable values

Graphing the SCADA Element

Once the SCADA signals are configured you can now compare the historical data as imported through the SCADA database against the simulated model results. When a SCADA Element (linked to a model element) is graphed you can compare both;

The blue line represents the model data obtained after analysis and the red squares depict the historical data obtained from the excel database. In this data all the data values are represents irrespective of the category (good, bad and questionable). The result field for the questionable value are non-numerical and hence cannot be plotted on a graph at present. However, an enhancement to include this option in graphs in the future release of WaterGEMS has been filed (#127234).

Alternatively, you can color code the SCADA element or annotate to return the Signal Quality (Historical) values.

Note: It should be noted that the data quality should be collected and incorporated in the database by the user before importing the database in the Database Source. That way the “Quality” field will be available for mapping.

See Also

Help > Modeling Capabilities > SCADAConnect Overview > SCADA Signals Setup > SCADA Signals – Database

SCADAConnect Simulator for WaterGEMS (V8i SELECTseries 6 and CONNECT edition)

Recommended
Related