Alert handling, basics

Alerts are sent whenever an original value or the valid online value of a data point element changes from one alert range to another (this does not apply to multi-instance alerts, see Alert handling for multi-instance alerts). An alert handling can be assigned to each data point element.

Alert handling must not be defined for data point elements with disabled last value storage, since this may result in inconsistent alarms.

Alerts tell the user of the control system when various, particularly critical, systems states occur. In many cases, the operator must send an acknowledgement to prove that he has seen the alert.

  • Alert ranges are intervals of the WinCC OA value range that together cover the entire range.
  • Each alert range has an assigned alert text (CAME text and/or WENT text) and alert class in addition to its limiting values.

Types of alert handling

There are three different types of alert handling, which are selectable depending on the data type of the data point element: Sum alert handling, Alert handling of continuous values, Alert handling of discrete values and Alert handling for multiinstance alerts.

An alert handling contains attributes for the configuration of alerts as well as attributes for an alert instance at runtime.

Sum alert handling

Sum alert handling combines several individual alerts of several alert handlings.

Sum alerts can be configured for non-numeric data point elements (string, struct, char, time, blob, DPID, langString, bit32).

Note that sum alerts are only archived if they contain an alert class. If an alert class is not assigned to a sum alert, the alert is not archived.

Alert handling of continuous values

An alert handling of continuous values is used when the whole value range shall be covered with specified intervals (limits).

An continuous alert handling can be configured for the following data types: uint, int, float and bool.

Additionally there can be defined a hysteresis for each alert range in order that a fluctuation of a value between two ranges does not trigger a host of alerts. See also Example of continuous alert handling.

Alert handling of discrete values

Alert handling of discrete values is used when discrete states should trigger an alert.

The alert handling of discrete values is used for data point elements of the data types uint, int, bool or bit32.

See also Example of discrete alert handling.

Alert handling of discrete values is not supported for data points of the data type float

Alert handling for multiinstance alerts

An alert handling for multiinstance alerts allows triggering multiple alerts from the periphery over a special WinCC OA driver or over a manager (see Multiinstance alerts). The alert properties (e.g. the alert class or the alert text) can be changed dynamically when the alert is triggered.

The alert handling for multiinstance can be used for analog and binary data types (uint, int, float, and bool).

The possible alert handling attributes can be found in the chapter _alert_hdl.

All chances of the configuration are usually saved in the configuration history.

An alert handling with more than twenty alert ranges cannot be configured with the current panels of the PARA module. But it is possible via the ASCII manager.

Additional values

Additional properties, so-called additional values, can be placed on an alert. These additional values are set by a special driver / manager or via Control / API functions.

It is recommended to write the alert value, text, priority and the quality code of the alert, which comes from the periphery, to the first four additional values ( _alert_hdl.._add_value_1 until _alert_hdl.._add_value_4) - what values are actually written to the values, is up to the driver developer. If this order is maintained, the first two additional values can be displayed in the corresponding columns ("Value" and "Alert text") of the alert screen (see Additional Values) when the config entry attributesFromAddValues is set.

The alert screen is not updated due to additional values. The additional values of an alert shown in the alert screen are not updated.

An alert can save up to 32 additional values that can be queried separately. These are addressable via config attributes - they can be written either all together to the _alert_hdl.._add_values attribute or individually to the _alert_hdl.._add_value_1 to _alert_hdl.._add_value_32 attributes. Although more than 32 additional values can be stored in the _alert_hdl.._add_values attribute, these can then be queried only together (by a query to _alert_hdl.._add_values; a direct access is only possible on the first 32 values using the attributes _alert_hdl.._add_value_1 to _alert_hdl.._add_value_32) and are available only in the historical database. Additional values are stored in the historical database (RAIMA or RDB). An additional value in the RDB may not exceed the size of 4 KBytes.

Additional values can be set for specific alerts by using alertSet/alertSetWait. If an alert was acknowledged, additional values can only be written for the alert by using the attribute "_add_values". In this case additional values cannot be set via _add_value_x (x = number).

'alert_hdl.._add_value_x' and 'alert_hdl.._add_values' may not be used in a same SELECT-Statement of a SQL query when RDB archiving is used.

General layout of the alert handling panel

The panels for the configuration of a continuous or discrete alert handling and the sum alert handling have similar functions. The differences are in the specific tabs of each panel. They are described in the following chapters separately.

Figure 1. Upper area of a panel

There is the DPE: field in the top right corner of each panel. It shows the name of the configured data point element (or description/alias).

In the right combo box in the bottom area of the alert handling panel the type of the alert handling has to be selected (this is not necessary in case of a sum alert, as for data point elements of a non-numeric data type this is automatically selected).

Figure 2. Combo box for alert handling type selection

Depending on the type of the alert handling to be configured on the data point element, the functions in the bottom area slightly differ.

Figure 3. Bottom area of an alert handling panel of continuous values
Figure 4. Bottom area of an alert handling panel of discrete values
Figure 5. Bottom area of an alert handling panel for multiinstance alerts
Figure 6. Bottom area of a sum alert handling panel

The fields/settings are described in the following:

active for

Must be ticked, when the alert handling shall be activated for the data point element. It can be chosen whether for the original value or for the online value shall be triggered an alert.

state

As soon as the alert is active, two state fields are visible. In the left one is shown the alert text in the corresponding alert color. In the right field is shown the currently alert state (no alert, CAME/acknowledged). The alert state is defined by the highest priority of the single alerts. The weight of the priorities is defined in the "Alert" tab.

Help

Opens the related help chapter from the WinCC OA online help.

OK

Saves the configuration and closes the alert handling panel (only possible if the alert handling is not active).

Acknowledge

By clicking on this button, a pending alert will be acknowledged, if one exists.

Apply

Saves the configuration in the alert handling panel (only possible if the alert handling is not active).

Close

Closes the alert handling panel without saving the configuration.

Settings/fields of continuous/discrete alert handling

Impulse

If an alert state (value range) is triggered several times by the same value, normally it results in only one alert (new state). If the "Impulse" checkbox is enabled, the same value always triggers a new alert with an updated alert time. For example, if 0 is set as good range and a value series of 1-1-1-... is received, a new alert is triggered every time. Depending on the selected alert handling (discrete or continuous), consider the following differences when using impulse alerts:

  • continuous: Since there are no WENT events, only the acknowledge types "acknowledge deletes" and "CAME is acknowledgeable" are allowed. In addition it is possible to handle a quick 0-1-0 pulse where the signal edge from good range (0) to alert range (1) triggers the alert. The WENT event is ignored. This means that the alert is displayed as "CAME/not acknowledged", though it is already gone. The alert remains till it is updated by the next impulse or deleted by acknowledge. Impuls alerts for alert handling of continuous values can only be used for data type bool.
  • discrete: In contrast to impulse alerts for continuous values, there are WENT events. Therefore the following acknowledge types are possible: "acknowledge deletes", "CAME is acknowledgeable", "CAME or WENT must be acknowledged" and "NOT acknowledgeable". Impulse alerts for alert handling of discrete values can be used for the following data types: bool, int, uint, float, char and bit32.

When using Impulse alerts the status bits can not be used.

Ranges

Number of alert ranges, which will be configured ("2 ranges",...,"20 ranges")

Settings/fields of sum alert handling

Filter

If single alerts shall be disabled/invisible, the check box "Filter" can be used. In the "Alert" tab an alert filter range value can be set for the sum alert handling. Thereby can be defined of how many single alerts a sum alert should be shown. If for example the range value 4 is set and 4 single alerts will be triggered, in the alert panel will be shown only one sum alert. By ticking the "Filter" check box an alert can be disabled (set to invisible), also then when the set range value has not been reached yet. The settings about filtering are required.

Note that you have to activate the automatic filtering in the panel "Properties of the alert table" -> on the Filter types tab of the alert screen settings panel in order to use the threshold parameter. Note also that if you use the alert filtering (the threshold parameter), you have to specify an alert class for the sum alert! Se chapter Alert filtering.

Changing the configuration

To change the configuration of an alert handling the active for check box must be disabled. This is only possible if there are currently no pending alerts for this DPE. Otherwise these needs to be acknowledged first by clicking the Acknowledge button.

Alerts and state displays

Alerts are displayed as:

  • Text on the alert screen, i.e. as alerts in the strict sense
  • State displays, for instance by flashing graphical elements.

The user acknowledges alerts on the alert panel, in which each line is one alert, by clicking on the appropriate "Acknowledge" cell. State displays are confirmed by acknowledging the appropriate flashing graphical object (in single or overall acknowledgement mode).

On the subject of acknowledgements, please read the section on Acknowledging state displays and Alert panel.

In WinCC OA the attributes of the config _alert_hdl are used to handle data point variable alerts. s.

  • 10:10:00 dp1 is set.

  • 11:00:00 alert handling will be deactivated (.._stime is still 10:10:00)

  • 12:00:00 dp1 is set with dpSetTimed("10:30:00", ..) -> the above error message will be shown (.._stime is now 10:30:00)

The error message:

WCCILevent (0), 2007.10.19 13:36:34.377, IMPL, INFO, 86, Invalid value, DP: <DPName>:_original.._value, AlertNonBinarySignal, handleNewValue, newValueTime (2007.10.19 13:36:34.205) set to: 2007.10.19 13:36:34.315

means that after the deactivation of the alert, the alert was activated with a time stamp before the deactivation of the alert:

EXAMPLE:

Also when you activate the alert handling with a dpSetTimed() before 11:00:00, this error message will be shown since the alert handling cannot be activated before it is deactivated.

General Notes

Alarm Instances

To get more information for the alarm instances in a system additional log messages and debug options are available for the Event Manager, see WCCILevent.

If the system detects more than 500 (1000, 1500, 2000, ...) alarm instances for a data point element an error is reported. If the number of alert instances is increasing or decreasing (by 500) additional log messages are written.

More alarm instances can occur if multi instance alarms or the option "acknowledge old alarms" for the alert class is used.

Example

Example for the log message when the number of alert instances is increasing, <DPE identifier> is substituted with the name of the data point element:

WCCILevent   (0), <TIMESTAMP>, PARAM,WARNING,    54,  Unexpected state, DP: <DPE identifier>, AlertAdministration, number
 of alarm instances greater than , 500

Example

Example for the log message when the number of alert instances is decreasing, <DPE identifier> is substituted with the name of the data point element:

WCCILevent   (0), <TIMESTAMP>, PARAM,WARNING,    54,  Unexpected state, DP: <DPE identifier>, AlertAdministration, number
 of alarm instances now less than , 500