{"id":58868,"date":"2023-02-17T14:58:10","date_gmt":"2023-02-17T06:58:10","guid":{"rendered":"https:\/\/oihelp.corporate.ifs.com\/help\/sentinel\/sentinel-processes\/drift-detection-process\/"},"modified":"2024-07-09T10:01:34","modified_gmt":"2024-07-09T02:01:34","slug":"drift-detection-process","status":"publish","type":"page","link":"https:\/\/oihelp.corporate.ifs.com\/help\/sentinel\/sentinel-processes\/drift-detection-process\/","title":{"rendered":"Drift Detection Process"},"content":{"rendered":"\n

The Drift Detection Process monitors the deviation between process variable data (inputs) and a reference input. The reference input can either be a fixed value, or it can be a dynamic value such as an entity.<\/p>\n

A primary deviation limit is applied to the reference input, as either a percentage value or an absolute value. This limit can be an upper limit, a lower limit, or both, forming a deviation band. In the same way, secondary and tertiary limits may be applied to the reference input, to raise secondary or tertiary states, respectively.<\/p>\n

Upper and lower primary deviation limits are created:<\/strong><\/p>\n\n\n\n\n
\"\"<\/td>\n\"\"<\/td>\n<\/tr>\n
The reference input data shown over several sample intervals <\/td>\nUsing an Absolute Value deviation mode, a primary deviation \u201cband\u201d is created. Upper and Lower deviations are used in this example.<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

Input Data is evaluated against upper and lower primary deviation limits:<\/strong><\/p>\n\n\n\n\n
<\/td>\n Input data
\n Primary deviation limits
\n Reference input data
\n Primary deviation \u201cband\u201d<\/td>\n<\/tr>\n
Input data is evaluated against the primary deviation limits at every sample interval.<\/td>\n A primary<\/strong> event is raised when data exceeds the limit.<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

As well as detecting data that breaches the primary, secondary and tertiary deviation limits (raising a primary, secondary or tertiary state, respectively), the drift detection process has a variety of option methods for raising a secondary or tertiary event.<\/p>\n

The Drift Detection process can measure duration above the primary limit, and also the number of data points above the primary limit. These measurements can be taken consecutively, or they can be summed in total over the preceding rolling sum period.<\/p>\n

The Drift Detection process is used for monitoring the deviation between a reference input and the test input. Unacceptable deviations may signify equipment failure or calibration errors.<\/p>\n

Drift Detection is a complex process capable of concurrently monitoring multiple conditions such as transgression of limits, state duration, and movement between states.<\/p>\n

\n

State Transition Rules<\/h2>\n

The Drift Detection process has clearly defined state transition logic paths. Transition from one state to another is equally dependant on the current evaluation of data, and on the current state.<\/p>\n

State transitions cause events to be raised, allowing for the escalation of actions via the Sentinel framework. Different actions can be assigned to different state outcomes.<\/p>\n

The following table outlines the allowable state transitions of the Drift Detection process.<\/p>\n\n\n\n\n\n\n\n\n\n\n
This state can transition<\/strong><\/td>\nTo this state<\/strong><\/td>\nUnder these conditions<\/strong><\/td>\n<\/tr>\n
Default<\/td>\nPrimary<\/td>\nData has transgressed the primary state upper or lower deviation limits.<\/td>\n<\/tr>\n
Primary<\/td>\nSecondary<\/td>\nAny one of the secondary state conditions have been met.<\/td>\n<\/tr>\n
Tertiary<\/td>\nOnly if the tertiary deviation limit is breached.<\/td>\n<\/tr>\n
Default<\/td>\nData is not erroneous, and is within the deviation limits of the curve.<\/td>\n<\/tr>\n
Secondary<\/td>\nTertiary<\/td>\nAny one of the tertiary state conditions has been met.<\/td>\n<\/tr>\n
Default<\/td>\nData is not erroneous, and is within the deviation limits of the curve.<\/td>\n<\/tr>\n
Tertiary<\/td>\nDefault<\/td>\nData is not erroneous, and is within the deviation limits of the curve.<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

The following diagram outlines the state transition logic of the Drift Detection process.<\/p>\n\n\n\n
<\/td>\nD<\/strong> Default
\nP<\/strong> Primary
\nS<\/strong> Secondary
\nT<\/strong> Tertiary<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n
\n

Test Outcomes<\/h2>\n

A number of outcomes are possible when the Drift Detection process is executed:<\/p>\n\n\n\n\n\n\n\n
Default State<\/strong><\/td>\nData is not in an erroneous state and is within the primary deviation limits.<\/td>\n<\/tr>\n
Primary State<\/strong><\/td>\nData is measured against the defined fixed or variable primary deviation limit (upper and\/or lower). If it breaches a primary deviation limit, a primary state is reached, and a primary event is raised.<\/td>\n<\/tr>\n
Secondary State<\/strong><\/td>\nA number of possible conditions can cause a secondary event, leading to a secondary state. These are explained in more detail in the Secondary State section.<\/td>\n<\/tr>\n
Tertiary State<\/strong><\/td>\nAs with the secondary state, a number of conditions can cause a tertiary event. These are explained in more detail in the Tertiary State section.<\/td>\n<\/tr>\n
Suppressed State<\/strong><\/td>\nThe monitor has been suppressed. For example, if the precondition has not been met.<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n
\n

Conditional Logic<\/h2>\n

The Drift Detection process provides the following conditional logic.<\/p>\n

Note: All of the example graphs in the following sections show tests that have used the Last Known Value sample method.<\/p>\n

Input Settings<\/h3>\n

The input and the reference input are defined in this section.<\/p>\n

    \n
  • Input<\/strong>: The input is the data that is monitored. Input data is defined as an attribute of the source entity, either as it is, or as part of a calculation.<\/li>\n
  • Reference Input<\/strong>: The fixed or variable reference input<\/em> is used to create deviation limits for the input data. The process will calculate limits around the reference input based on the deviation mode and the deviation limit settings, and then compare the input data stream to this model in order to determine drift.<\/li>\n<\/ul>\n

    Mode Settings<\/h3>\n

    Define whether the deviation mode is a percentage or an absolute value, to establish how primary, secondary and tertiary deviation limits are calculated. Also select whether to set upper or lower deviation limits, or both.<\/p>\n

    Deviation Limit Settings<\/h3>\n

    The primary state deviation limit must be set for the process to work. If the calculated primary state deviation limit is breached, a primary event occurs, and a primary state is reached. Optionally, a secondary state deviation limit and a tertiary state deviation limit may be set for the process.<\/p>\n

    Rolling Sum Period<\/h3>\n

    A secondary state rolling sum period can be defined for evaluating some of the secondary state conditions; likewise, a tertiary state rolling sum period can be defined for evaluating some of the tertiary state conditions.<\/p>\n

    The rolling sum period is a defined period (specified in days and hours). At every sample interval, the rolling sum period is that period preceding the sample interval. So, for example, if the secondary state rolling sum period is set at 1 day 2 hours then at 3pm on Tuesday 15 January 2013, the secondary state rolling sum period is from 1pm on Monday 14 January (covering the last 1 day and 2 hours). Any evaluations relating to that sample interval's secondary state rolling sum period conditions must fall within that time.<\/p>\n

    Primary State Deviation Limit<\/h3>\n

    The primary state deviation limit must be set for the process to work.<\/p>\n

    If the configured primary state deviation limit is breached, a primary event occurs, and a primary state is reached.<\/p>\n

    Example:<\/strong><\/p>\n

    In this example, a primary event is raised when the input data breaches the upper primary deviation limit. The sample interval is one hour.<\/p>\n\n\n\n
    <\/td>\n\u00a0Data item
    \n\u00a0Upper and lower primary deviation limits
    \n\u00a0Sample interval<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n\n\n\n
    \n

    \u00a0At\u00a05:00am<\/strong>, the input data exceeds the upper primary deviation limit. A primary state is reached, and a\u00a0primary<\/strong>\u00a0event is raised.<\/p>\n

    \u00a0At\u00a07:00am<\/strong>, the input data drops below the upper primary deviation limit. A default state is reached, and a\u00a0default<\/strong>\u00a0event is raised.<\/p>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

    Secondary State<\/h3>\n

    There are several possible ways to reach the secondary state outcome in the Drift Detection process:<\/p>\n

      \n
    • Secondary State Deviation<\/strong> Limit<\/li>\n
    • Secondary State Duration<\/strong> Limit<\/li>\n
    • Secondary State Sustained Value<\/strong> Limit<\/li>\n
    • Secondary State Rolling Sum Total Duration<\/strong> Limit<\/li>\n
    • Secondary State Rolling Sum Breach Occurrences<\/strong> Limit<\/li>\n<\/ul>\n

      You can choose to set one or more conditions to trigger the secondary state. For example, a test may have a secondary deviation limit defined, as well as configured variables for determining secondary state rolling sum breach occurrences.<\/p>\n

      Deviation Limit<\/h4>\n

      If the configured secondary deviation limit is breached, a secondary event occurs, and a secondary state is reached.<\/p>\n

      Example:<\/strong><\/p>\n

      In this example, a secondary event is raised when the input data breaches the upper secondary deviation limit. The sample interval is one hour.<\/p>\n\n\n\n
      <\/td>\n\u00a0Data item
      \n\u00a0Upper and lower primary deviation limits
      \n\u00a0Upper and lower secondary deviation limits
      \n\u00a0Sample interval<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n\n\n\n
      \n

      \u00a0At\u00a05:00am<\/strong>, the input data exceeds the upper primary deviation limit. A primary state is reached, and a\u00a0primary<\/strong>\u00a0event is raised.<\/p>\n

      \u00a0At\u00a07:00am<\/strong>, the input data exceeds the upper secondary deviation limit. A secondary state is reached, and a\u00a0secondary<\/strong>\u00a0event is raised.<\/p>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

      Duration Limit<\/h4>\n

      Monitor for data that continuously exceeds the primary deviation limit, for longer than the secondary state duration limit.<\/p>\n

      If data that continuously exceeds the primary deviation limit for longer than specified in the secondary state duration limit, a secondary event occurs, and a secondary state is reached.<\/p>\n

      Example:<\/strong><\/p>\n

      In this example, a secondary event is raised when the input data has exceeded the upper primary deviation limit for a period of 2 hours, exceeding the primary state duration limit of 1.5 hours. The sample interval is one hour.\u00a0<\/p>\n\n\n\n
      <\/td>\n\u00a0Data item
      \n\u00a0Upper and lower primary deviation limits
      \n\u00a0Sample interval<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n\n\n\n
      \n

      \u00a0At\u00a05:00am<\/strong>, the input data exceeds the upper primary deviation limit. A primary state is reached, and a\u00a0primary<\/strong>\u00a0event is raised.<\/p>\n

      \u00a0At\u00a07:00am<\/strong>, the input data is still above the upper deviation limit. The duration of data above the upper deviation limits is 2 hours, exceeding the 1.5 hour secondary duration limit. A secondary state is reached, and a\u00a0secondary<\/strong>\u00a0event is raised.<\/p>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

      Sustained Value Limit<\/h4>\n

      Monitor for when data exceeds the primary deviation limit for more than a specified number of times (secondary state, sustained value limit), consecutively. When this happens, a secondary event occurs, and a secondary state is reached.<\/p>\n

      Example:<\/strong><\/p>\n

      In this example, a secondary event is raised when the input data has exceeded the upper deviation limit, consecutively, for four sample intervals, exceeding the secondary state sustained value limit of 3. The sample interval is one hour.<\/p>\n\n\n\n
      \n

      <\/p>\n<\/td>\n

      		\n

      \u00a0Data item
      \n\u00a0Upper and lower primary deviation limits
      \n\u00a0Sample interval<\/p>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n\n\n\n
      \n

      \u00a0At\u00a05:00am<\/strong>, the input data exceeds the upper primary deviation limit. A primary state is reached, and a\u00a0primary<\/strong>\u00a0event is raised.<\/p>\n

      \u00a0At\u00a08:00am<\/strong>, the input data has exceeded the upper deviation limit for four sample intervals, consecutively, exceeding the secondary state sustained value limit of three. A secondary state is reached, and a\u00a0secondary<\/strong>\u00a0event is raised.<\/p>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

      Rolling Sum Total Duration Limit<\/h4>\n

      Monitor for a specified accumulated duration of all periods where data is in breach of the primary deviation limit, within the preceding specified secondary state rolling sum period.<\/p>\n

      If the total combined primary state duration is longer than the specified secondary state rolling sum duration limit, a secondary event occurs, and a secondary state is reached.<\/p>\n

      Example:<\/strong><\/p>\n

      In this example, a secondary event is raised when the input data is above the primary deviation limit for a total of four hours during the current secondary state rolling sum period, exceeding the secondary state rolling sum duration limit of 3.5 hours. The sample interval is one hour.\u00a0 Secondary state rolling sum period is 12 hours.<\/p>\n\n\n\n
      <\/td>\n\u00a0Data item
      \n\u00a0Upper and lower primary deviation limits
      \n\u00a0Sample interval
      \n\u00a0Current secondary state rolling sum period.<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n\n\n\n
      \n

      \u00a0At\u00a02:00am<\/strong>, the input data exceeds the upper primary deviation limit. A primary state is reached, and a\u00a0primary<\/strong>\u00a0event is raised.<\/p>\n

      \u00a0At\u00a04:00am<\/strong>, the input data drops below the upper primary deviation limit. A default state is reached, and a\u00a0default<\/strong>\u00a0event is raised.\u00a0The total duration of primary state for the current rolling sum period is 2 hours (two sample intervals). Timing stops while in the default state.<\/p>\n

      \u00a0At\u00a010:00am<\/strong>, the input data exceeds the upper primary deviation limit. A primary state is reached, and a\u00a0primary<\/strong>\u00a0event is raised. Timing resumes.<\/p>\n

      \u00a0At\u00a012:00pm<\/strong>, the combined total duration of primary state for the current rolling sum period is now 4 hours, exceeding the secondary state rolling sum duration limit. A secondary state is reached, and a\u00a0secondary<\/strong>\u00a0event is raised.<\/p>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

      Rolling Sum Breach Occurrences Limit<\/h4>\n

      Monitor for when a set number of values (secondary state breach occurrences limit) has breached the primary deviation limit, during the preceding specified secondary rolling sum period.<\/p>\n

      Example:<\/strong><\/p>\n

      In this example, a secondary event is raised when the input data has breached the primary deviation limit three times in the current secondary state rolling sum period, reaching the secondary state rolling sum breach occurrences limit of 3. The sample interval is one hour. Secondary state rolling sum period is 12 hours.<\/p>\n\n\n\n
      <\/td>\n\u00a0Data item
      \n\u00a0Upper and lower primary deviation limits
      \n\u00a0Sample interval
      \n\u00a0Current secondary state rolling sum period.
      \n\"\"\u00a0Occurrence Counter<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n\n\n\n
      \n

      \"\" At 2:00am<\/strong>, the input data exceeds the upper primary deviation limit. A primary state is reached, and a primary<\/strong> event is raised. The breach occurrences counter is one.\u00a0<\/p>\n

      \"\" At 4:00am<\/strong>, the input data drops below the upper primary deviation limit. A default state is reached, and a default<\/strong> event is raised. The breach occurrences counter remains at one.\u00a0<\/p>\n

      \"\" At 11:00am<\/strong>, the input data exceeds the upper primary deviation limit. A primary state is reached, and a primary<\/strong> event is raised. The breach occurrences counter is two.\u00a0<\/p>\n

      \"\" At 12:00pm<\/strong>, the primary deviation limit has been breached three times in the current rolling period, reaching the secondary state rolling sum breach occurrences limit. A secondary state is reached, and a secondary<\/strong> event is raised. The breach occurrences counter is three.\u00a0<\/p>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

      Tertiary State<\/h3>\n

      There are several ways to reach the tertiary state outcome in the Drift Detection process:<\/p>\n

        \n
      • Tertiary State Deviation Limit<\/li>\n
      • Tertiary State Duration Limit<\/li>\n
      • Tertiary State Sustained Value Limit<\/li>\n
      • Tertiary State Rolling Sum Total Duration Limit<\/li>\n
      • Tertiary State Rolling Sum Breach Occurrences Limit<\/li>\n<\/ul>\n

        You can choose to set one or more conditions to trigger the tertiary state. For example, a test may have a tertiary deviation limit defined, as well as configured variables for determining tertiary state rolling sum breach occurrences.<\/p>\n

        Deviation Limit<\/h4>\n

        If the configured tertiary deviation limit is breached, a tertiary event occurs, and a tertiary state is reached.<\/p>\n

        The graph demonstrates a breach of the tertiary deviation limit, causing a tertiary event to occur.<\/p>\n

        Example:<\/strong><\/p>\n

        In this example, a tertiary event is raised when the input data breaches the upper tertiary deviation limit. The sample interval is one hour.<\/p>\n\n\n\n
        <\/td>\n\u00a0Data item
        \n\u00a0Upper and lower primary deviation limits
        \n\u00a0Sample interval
        \n\u00a0Upper and lower secondary deviation limits
        \n\u00a0Upper and lower tertiary deviation limits<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n\n\n\n
        \n

        At 5:00am<\/strong>, the input data exceeds the upper primary deviation limit. A primary state is reached, and a primary<\/strong> event is raised.<\/p>\n

        At 6:00am<\/strong>, the input data exceeds the upper secondary deviation limit. A secondary state is reached, and a secondary<\/strong> event is raised.<\/p>\n

        At 8:00am<\/strong>, the input data exceeds the upper tertiary deviation limit. A tertiary state is reached, and a tertiary<\/strong> event is raised.<\/p>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

        Duration Limit<\/h4>\n

        Monitor for data that continuously exceeds the primary deviation limit, for longer than the tertiary state duration limit.<\/p>\n

        If data that continuously exceeds the primary deviation limit for longer than specified in the tertiary state duration limit, a tertiary event occurs, and a tertiary state is reached.<\/p>\n

        Example:<\/strong><\/p>\n

        In this example, a tertiary event is raised when the input data has exceeded the upper primary deviation limit consecutively for a period of 3 hours, exceeding the tertiary state duration limit of 2.5 hours. Secondary state duration limit is 1.5 hours. The sample interval is one hour.<\/p>\n\n\n\n
        <\/td>\n\u00a0Data item
        \n\u00a0Upper and lower primary deviation limits
        \n\u00a0Sample interval<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n\n\n\n
        \n

        At 5:00am<\/strong>, the input data exceeds the upper primary deviation limit. A primary state is reached, and a primary<\/strong> event is raised.<\/p>\n

        At 7:00am<\/strong>, the input data is still above the upper primary deviation limit. The duration of data above the primary limit is now 2 hours, exceeding the 1.5 hour secondary duration limit. A secondary state is reached, and a secondary<\/strong> event is raised.<\/p>\n

        At 8:00am<\/strong>, the input data is still above the upper primary deviation limit. The duration of data above the primary limit is now 3 hours exceeding the 2.5 hour tertiary duration limit. A tertiary state is reached, and a tertiary<\/strong> event is raised.<\/p>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

        Sustained Value Limit<\/h4>\n

        Monitor for when data exceeds the primary deviation limit for more than a specified number of times (tertiary state sustained value limit), consecutively. If this is the case, a tertiary event occurs, and a tertiary state is reached.<\/p>\n

        Example:<\/strong><\/p>\n

        In this example, a tertiary event is raised when the input data has exceeded the upper deviation limit for five sample intervals, consecutively, exceeding the tertiary state sustained value limit of 4. Secondary state sustained value limit is 2. The sample interval is one hour.<\/p>\n\n\n\n
        <\/td>\n\u00a0Data item
        \n\u00a0Upper and lower primary deviation limits
        \n\u00a0Sample interval<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n\n\n\n
        \n

        At 5:00am<\/strong>, the input data exceeds the upper primary deviation limit. A primary state is reached, and a primary<\/strong> event is raised.<\/p>\n

        At 7:00am<\/strong>, the input data has exceeded the upper primary deviation limit for three sample intervals, exceeding the secondary state sustained value limit of two. A secondary state is reached, and a secondary<\/strong> event is raised.<\/p>\n

        At 9:00am<\/strong>, the input data has exceeded the upper deviation limit for five sample intervals, exceeding the tertiary state sustained value limit of four. A tertiary state is reached, and a tertiary<\/strong> event is raised.<\/p>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

        Rolling Sum Total Duration Limit<\/h4>\n

        Monitor for a specified accumulated duration of all periods where data is in breach of the primary deviation limit, within the preceding specified tertiary state rolling sum period.<\/p>\n

        If the total combined primary state duration is longer than the specified tertiary state rolling sum duration limit, a tertiary event occurs, and a tertiary state is reached.<\/p>\n

        Example:<\/strong><\/p>\n

        In this example, the tertiary state rolling sum duration limit is exceeded, causing a tertiary event. Tertiary state rolling sum total duration is 3.5 hours. Tertiary rolling sum period is 12. The sample interval is one hour.<\/p>\n\n\n\n
        <\/td>\n\u00a0Data item
        \n\u00a0Upper and lower primary deviation limits
        \n\u00a0Sample interval
        \n\u00a0Current secondary state rolling sum period.
        \n\u00a0Previous sample interval\u2019s tertiary state rolling sum period<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n\n\n\n
        \n

        At 2:00am<\/strong>, the input data exceeds the upper primary deviation limit. A primary state is reached, and a primary<\/strong> event is raised.<\/p>\n

        At 4:00am<\/strong>, the input data drops below the upper primary deviation limit. A default state is reached, and a default<\/strong> event is raised.\u00a0The total duration of primary state for the current rolling sum period is 2 hours. Timing is paused.<\/p>\n

        At 9:00am<\/strong>, the input data exceeds the upper primary deviation limit. A primary state is reached, and a primary<\/strong> event is raised. Timing resumes.<\/p>\n

        At 11:00am<\/strong>, the combined total duration of data above the primary limit for the current rolling sum period is now 4 hours, exceeding the tertiary state rolling sum duration limit of 3.5 hours. A secondary state is reached, and a secondary <\/strong>event is raised. Note: The test outcome cannot go from primary state to tertiary state unless a tertiary deviation limit has been breached.<\/p>\n

        \"\" At 12:00pm<\/strong>, the combined total duration of data above the primary limit for the current rolling sum period is now 5 hours, exceeding the tertiary state rolling sum duration limit of 3.5 hours. A tertiary state is reached, and a tertiary <\/strong>event is raised.<\/p>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

        Rolling Sum Breach Occurrences Limit<\/h4>\n

        Monitor whether a set number of values (tertiary state breach occurrences limit) has breached the primary deviation limit, during the preceding specified tertiary rolling sum period.<\/p>\n

        Example:<\/strong><\/p>\n

        In the following example, the tertiary state rolling sum breach occurrences limit is exceeded, causing a tertiary event. Note that a secondary state must be reached first. Tertiary state breach occurrences limit is 3. Tertiary rolling sum period is 12 hours. The sample interval is 1 hour.<\/p>\n\n\n\n
        <\/td>\n\u00a0Data item
        \n\u00a0Upper and lower primary deviation limits
        \n\u00a0Sample interval
        \n\u00a0 Tertiary state rolling sum period (12 hours)
        \n\"\"\u00a0Occurrence Counter<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n\n\n\n
        \n

        \"\" At 2:00am<\/strong>, the input data exceeds the upper primary deviation limit. A primary state is reached, and a primary<\/strong> event is raised. The breach occurrences counter is one.<\/p>\n

        \"\" At 4:00am<\/strong>, the input data drops below the upper primary deviation limit. A default state is reached, and a default<\/strong> event is raised. The breach occurrences counter remains at one.<\/p>\n

        \"\" At 11:00am<\/strong>, the input data exceeds the upper primary deviation limit. A primary state is reached, and a primary<\/strong> event is raised. The breach occurrences counter is two.<\/p>\n

        \"\" At 12:00pm<\/strong>, the primary deviation limit has been breached three times in the current rolling sum period, reaching the tertiary state rolling sum breach occurrences limit. A secondary state is reached, and a secondary<\/strong> event is raised. Note: The test outcome cannot go from primary state to tertiary state unless a tertiary deviation limit has been breached. The breach occurrences counter is three.<\/p>\n

        \"\"\"\"\u00a0At 1:00pm, <\/strong>the primary deviation limit has been breached four times in the current rolling sum period, exceeding the tertiary state rolling sum breach occurrences limit. A tertiary state is reached, and a tertiary <\/strong>event is raised. The breach occurrences counter is four.<\/p>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

        \n

        Adding a Drift Detection Process<\/h2>\n

        The Drift Detection Process monitors the deviation between process variable data (inputs) and a reference input. If a limit or condition is breached when the process is executed, a new state is reached and an event is raised.<\/p>\n

        Setting Process Values and Limits<\/h3>\n

        Part of setting up the Drift Detection process involves selecting limits, such as the primary state deviation limit, secondary state deviation limit, tertiary state deviation limit, breach occurrences limits and so on. Limits can either be fixed values, or they can be variable data taken from entities.<\/p>\n

        The following limits are available. Select a limit type from the drop-down list, then type in or select a limit.<\/p>\n

          \n
        • Fixed Value<\/strong>: Type in a numerical value.<\/li>\n
        • Attribute<\/strong>: This option is only available if the test\u2019s Source Type<\/strong> is Entity <\/strong>or Hierarchy<\/strong>.
          \nClick the ellipsis button to select an attribute of the source entities.<\/li>\n
        • Source Tag<\/strong>: This option is only available if the Source Type<\/strong> is Tag<\/strong>.<\/li>\n
        • Calculation<\/strong>: Click the ellipsis button to open the Edit Calculation window.\u00a0\n
            \n
          • If the Source Type is Entity<\/strong> or Hierarchy<\/strong>: Type a calculation, prefixed by \u2018this\u2019 as the Source Entity<\/strong> token, for example: {this:THP} + 34<\/strong>.<\/li>\n
          • If the Source Type is Tag<\/strong>: Type a calculation, prefixed by \u2018this\u2019 as the Source<\/strong> Tag<\/strong> token, for example: {this} * 2<\/strong>.<\/li>\n<\/ul>\n<\/li>\n
          • Tag<\/strong>: Click the ellipsis button to select a tag.<\/li>\n
          • Entity Attribute<\/strong>: Click the ellipsis button to select an entity. From here, select an attribute, or attribute value, for the selected entity.<\/li>\n<\/ul>\n

            Adding the Process<\/h4>\n

            As with all Sentinel processes, the Drift Detection Process is defined within a Sentinel Test<\/strong> page.<\/p>\n

            In the Sentinel Test<\/strong> page:<\/p>\n

            1. Expand the Process<\/strong> panel.<\/p>\n

            2. Select Drift Detection<\/strong> from the drop-down list.<\/p>\n

            <\/p>\n

            Input Settings<\/h5>\n

            3. From the Input<\/strong> drop-down list, select an input from the following:<\/p>\n

              \n
            • Attribute<\/strong>: This option is only available if the Source Type<\/strong> is Entity <\/strong>or Hierarchy<\/strong>.
              \nClick the ellipsis button to select an attribute. You are limited to selecting an attribute of the test source monitor items. This attribute of each of the monitor items is a separate process input.<\/li>\n
            • Source Tag<\/strong>: This option is only available if the Source Type<\/strong> is Tag<\/strong>.<\/li>\n
            • Calculation<\/strong>: Click the ellipsis button to open the Edit Calculation window.\u00a0\n
                \n
              • If the Source Type is Entity<\/strong> or Hierarchy<\/strong>: Type a calculation, prefixed by \u2018this\u2019 as the Source Entity<\/strong> token, for example: {this:THP} + 34<\/strong>.<\/li>\n
              • If the Source Type is Tag<\/strong>: Type a calculation, prefixed by \u2018this\u2019 as the Source<\/strong> Tag<\/strong> token, for example: {this} * 2<\/strong>.<\/li>\n<\/ul>\n<\/li>\n
              • Entity Attribute<\/strong>: Click the ellipsis button to select an entity. From here, select an attribute, or attribute value, for the selected entity.<\/li>\n<\/ul>\n

                4. From the Reference Input<\/strong> drop-down list, select one of the following; then type in or select a corresponding value:<\/p>\n

                  \n
                • Fixed Value<\/li>\n
                • Attribute (only available of the Source<\/strong> type is Entity or Hierarchy)<\/li>\n
                • Source Tag (only available of the Source<\/strong> type is Tag)<\/li>\n
                • Calculation<\/li>\n
                • Tag<\/li>\n
                • Entity Attribute<\/li>\n<\/ul>\n
                  Mode Settings<\/h5>\n

                  5. Define whether the deviation mode is a percentage or an absolute value, to establish how primary, secondary and tertiary deviation limits are calculated. Also select whether to set upper or lower deviation limits, or both.<\/p>\n

                  Note: The mode settings apply to primary, secondary and tertiary deviation limits.<\/p>\n

                    \n
                  1. From the Deviation Mode<\/strong> drop-down list, select Percentage<\/em> or Absolute Value<\/em>. The deviation mode determines how the deviation limits will be calculated.<\/li>\n
                  2. Select the Upper Deviation<\/strong> check box to set an upper deviation limit.<\/li>\n
                  3. Select the Lower Deviation<\/strong> check box to set a lower deviation limit.<\/li>\n<\/ol>\n
                    Deviation Limit Settings<\/h5>\n

                    Note: The Primary Deviation Limit is mandatory for the Drift Detection Process.<\/p>\n

                    6. From the Primary Limit<\/strong> drop-down list, select one of the following and then type in or select a corresponding value:<\/p>\n

                      \n
                    • Fixed Value<\/li>\n
                    • Attribute (only available of the Source<\/strong> type is Entity or Hierarchy)<\/li>\n
                    • Source Tag (only available of the Source<\/strong> type is Tag)<\/li>\n
                    • Calculation<\/li>\n
                    • Tag<\/li>\n
                    • Entity Attribute<\/li>\n<\/ul>\n

                      7. Optionally define secondary limits in the same way (first select the Secondary Limit<\/strong> check box).<\/p>\n

                      8. Optionally define tertiary limits in the same way (first select the Tertiary Limit<\/strong> check box).<\/p>\n

                      Secondary State<\/h5>\n

                      Note: All Secondary State settings are calculated from the primary deviation limit.<\/p>\n

                      All secondary state settings are captured or selected in the Secondary State<\/strong> section of the process, with the exception of the Secondary Deviation Limit, which is set in the Deviation Limit Settings<\/strong> section of the process.<\/p>\n

                      <\/p>\n

                      Duration Limit<\/h6>\n

                      The Secondary State Duration Limit is used for monitoring where data is continuously beyond the primary deviation limit, for longer than the specified secondary duration limit.<\/p>\n

                      9. In the Secondary State<\/strong> section:<\/p>\n

                        \n
                      1. Select the Duration Limit<\/strong> check box.<\/li>\n
                      2. Type integer values in the days<\/strong>, hours<\/strong>, mins <\/strong>(minutes), and secs <\/strong>(seconds) Duration Limit <\/strong>boxes to define a duration period. The default value is zero.<\/li>\n<\/ol>\n

                        \"C:\\Users\\GABRIE~1.LAN\\AppData\\Local\\Temp\\SNAGHTML9bd3f4.PNG\"<\/p>\n

                        Sustained Value Limit<\/h6>\n

                        The Secondary State Sustained Value Limit is used for monitoring where data is beyond the primary deviation limit for more than a specified number of times (secondary state sustained value limit), consecutively.<\/p>\n

                        10. In the Secondary State<\/strong> section:<\/p>\n

                          \n
                        1. Select the Sustained Value<\/strong> Limit<\/strong> check box.<\/li>\n
                        2. From the Sustained Value<\/strong> Limit <\/strong>drop-down list, select one of the following and then type in or select a corresponding value:<\/li>\n<\/ol>\n