How You Manage Maintenance Forecasts

A forecast collects work requirements from all the active work requirements and suggests due dates based on the asset operating organization, maintenance organization, forecast method, concurrent work options, and last maintenance work order.

This includes assets that operate in the contextual program’s maintenance organization as well as other organizations when enabled.

Note: If an asset has the indicators, Allow maintenance programs or Allow work orders set to No, then it won’t be considered by the forecast scheduled process.

When generating the maintenance forecast, the application considers the value set for the Forecast Horizon in Days. This value is defined in the maintenance organization plant parameters, program header and work requirement, controlling how far into the future the forecast will generate due dates. The number of days value is used based on the lowest level of definition in a program. After a forecast is run, if some due dates aren't displayed in the calendar or Gantt chart, you can review the setup and adjust the horizon window values. For additional details, refer to the Guidelines to Set Up Maintenance Plant Parameters topic in the Implementing Manufacturing and Supply Chain Materials Management guide.

If an asset is operating in a maintenance organization, the due dates will be forecasted in the same organization. If an asset is operating in a non-mx organization, the program won't forecast in that organization as work orders can’t be created. Instead, the due dates forecast in either the program contextual maintenance organization or a primary maintenance organization.

Generate Maintenance Forecasts

Create and regenerate forecasts regularly to provide latest data to your planners. A forecast is used to verify if the modeling methods are correctly translated into a cadence of due dates into the future. It considers all active programs, work requirements and assets, and then generates the future preventative maintenance due dates. It is recommended as part of implementation as well as through on-going support that you fully understand all the modeling capabilities so that you know how to define, create, and update an asset’s preventative maintenance forecast.

You can create and periodically update the forecast by one of two methods:

• Within a single program, you can verify the forecast methods for the work requirements by selecting from the Actions drop-down the Generate Forecast button. This will generate a scheduled process to create and update the forecast.
• Once a program is validated, you should set up the Forecast to periodically run using a scheduled process. The job can be accessed from the Tasks pane on the Maintenance Management landing page by clicking the Generate Maintenance Forecast link. For additional details about this scheduled process, refer to the Scheduled Processes for SCM guide.

When the Generate Maintenance Forecast scheduled process is submitted, it locks all the unique Maintenance Programs covered by the program parameters. When it completes, it unlocks the program so that its forecast can be generated again by resubmitting the scheduled process. In the rare case where a scheduled process ends in error or when a process is manually stopped by user intervention, there's a chance the program is still locked. This prevents further updates of its forecast using the scheduled process. If this occurs, you can unlock the program by this method.

• Create a small program with 1 work requirement in the same organization.
• Ensure the scheduled process is not running for the organization or program.
• For the small program, run the Generate Forecast in-session, this should unlock all the programs in the organization.
• Then try running the Scheduled Process again, perhaps for one program at the time until the issue is resolved and confirmed. You should also coordinate efforts with Oracle Support if you have any questions or concerns.

Note: It is recommended to have a coordinated effort for creating and updating the forecast of individual programs as well as programs across each organization. If individual updates are performed that are not coordinated across users on a regular basis, it can be difficult to maintain consistency in updating the forecasts. Scheduled processes should be used to consistently update the forecast on a routine interval that is known by your planners.

Validate a Maintenance Forecast

After a forecast is generated, you can view and confirm a forecast within the program using the Gantt or Calendar tabs, by searching for an asset and/or the work requirement on the Maintenance Forecasts page or reviewing an OBTI analysis or report. Depending on the size and complexity of a forecast, you will pick the appropriate method.

Within the program, you can confirm that a forecast has been successfully created or updated. There are 2 key indicators located on the Overview tab of each program and tell you when it was last updated and forecasted.

You can then validate the forecast modeling outcome of each asset using the Gannt Chart or Calendar tabs within the program, or by searching the Maintenance Forecasts page. For most modeling solutions, you can easily verify the cadence and expected due dates match your forecast modeling. This includes viewing additional details about a due date, such as the source of the forecast, interval, and work definitions that are included.

During the creation of new work requirement for an asset, you may need an iterative approach where you generate and review the forecast, make small modifications, then regenerate the forecast to verify the outcome. However, once a forecast has been established and work orders have been created, care should be taken when modifying the forecast methods of the source work requirement. Additional details and consideration around updating a work requirement is covered in the next section.

It is also recommended to have an OTBI analysis or report available to verify the latest updates to a program using the Maintenance Program and Work Requirement subject areas. These updates can then be validated against the corresponding OTBI report for the Forecast Line subject area, allowing you to confirm if the latest programs updates are reflected in the forecast or not. This comparison provides you the ability to understand the resulting cadence and due dates in the forecast and help guide any future updates to the modeling methods. Additional details on using OTBI are covered in this user guide in the section entitled Reports and Analytics.

Here are the key fields in an OTBI report for the Maintenance Program and Work Requirement definition subject areas that can be used to validate against the Forecast Lines subject area:

• Maintenance Program - Header
  • Creation Date – useful for verifying when the program was created. This is the baseline date for the creation and editing of work requirement details.
  • Last Updated Date – useful to know when the last time header attributes were updated. This date does not reflect work requirement updates.
  • Last Forecasted On Date – as mentioned, this is key to knowing when the last forecast successfully completed either in-session or using the scheduled process for this Program. This same field is also tracked at the Work Requirement and Forecast line level.
• Work Requirement - Header
  • Creation Date – useful for verifying when the requirement was created. Comparing this to the Last Updated Date will reveal if updates have been made.
  • Last Updated Date – useful to know when the last time header attributes were updated. There is no auditing history in the page, so it can be difficult to know what changes were made by a user.
    • Comparing the historical Forecast cadence against the latest modeling method is only way to understand if changes were made over time.
  • Last Forecast Date – the individual requirements are also updated with the last time the forecast was generated. This is useful to know if the requirement was in an active status when the last time a program was updated or not by comparing to the Program Last Forecast Date.
    • Comparing the Creation and Last Updated dates on the requirement to the Last Forecast Date will confirm if the latest version of the requirement was considered in the forecast
    • Additionally, the requirement’s Last Forecasted Date can be compared to the Last Forecasted Date in each resulting Forecast Line. This will provide some indication if the latest or an earlier version was used to create the line and its due date.
  • Start Date – the start date is important to determine the first due date. However, once the first work order is created, the start date shouldn’t be updated, as it will not re-sequence the due dates.
    • If a user has updated it to a different value, then you can usually verify it against the historical forecast by on Last Forecast Dates.
• Work Requirement - Work Definitions
  • Creation Date – useful for verifying when the work definition was added to the requirement.
  • Last Updated Date – useful to know when the last time the definition was updated.
    • You can compare a Work Definitions Last Updated Date to the requirement’s Last Updated Date and Forecast Last Date to verify if the current Work Definition details were considered in the forecast.
    • It is possible to delete a work definition, add a work definition or edit is interval value, along with merge/suppress option between or across different forecast updates. This could cause forecast confusion and issues in understanding the intended modeling and cadence
    • Additionally, if the forecast has not been refreshed since the last work definition or requirement update, then the future forecast will not be in sync per the modeling
  • Repeats At Interval – this value can be updated over time, so it is useful to compare to past forecast history to understand any changes
• Work Requirement - Affected Assets
  • Creation Date – useful for verifying when the asset explicitly defined for the requirement.
  • Last Updated Date – useful to know when the last time the asset was updated. You can compare it to the requirement’s Last Updated Date and Forecast Last Date.
  • Historical Details – these can only be updated prior to the first work order, so they should be easy to compare to the forecast to verify any changes in the modeling and the affected asset details

Use an OTBI report for the Forecast Lines subject area to compare against the latest updates to the Maintenance Program and Work Requirement definition subject areas. As covered in the Reports and Analytics section of this guide, the topic OTBI for Oracle Maintenance provides guidance on all the subject areas, including an example of an OTBI analysis for the Forecast Lines. Here are some key fields from the analysis example to review:

• Forecast Lines for a program should be sorted by Work Requirement, then Asset, Forecast Sequence, Due at Cycle Interval and Forecast (Due) Date. This will provide a logical order of forecast history to compare against the modeling methods.
• Forecast Line ID - a Forecast Line will be deleted and recreated by each forecast update until a work order is created, resulting in a new unique Line ID. Therefore, you can compare these values over time to confirm from a previous forecast to a later forecast if the lines were recreated or not.
• Last Forecast Date – you will see the last date in which a forecast was updated for a line with a work order. Once a work order is created, the forecast line is frozen. This date can be compared against the corresponding dates in the work requirement to verify if the latest updates are reflected in the forecast lines or not.
• Forecast Sequence and Due Date – The Forecast Sequence will begin as of the Start Date at a value of 1 and will calculate forward for the forecast horizon number of days. If the Start date is in the past, then the sequence will calculate from that date, but it will only produce due dates from the application date forward. Therefore, it may be common to see a forecast sequence for a cycle of intervals starting in the forecast at a value other than 1.
  • The cadence of a requirement for an Asset over time should be analyzed by comparing the Start Date, Cycle, and Intervals (if used), Work Definition Due Intervals (if cycle-based) against the Forecast Lines. This will verify if the current modeling is correctly applied for all future due dates without a work order.
  • Additionally, if the forecast method is utilizing a last completion option, then forecasted next due will not be properly recalculated until the next work order is completed.
• Forecast Meter Reading – if utilization meters for an Asset are used to calculate the due dates, then this value will reflect where the forecast anticipates the reading of the asset each future due date. This value is calculated using the base interval value and meter utilization rates. The due dates are then updated based on the actual meter reading history when a forecast is updated.

Update a Work Requirement

As you are validating a forecast, you may identify due dates are not falling on the correct sequence for one or more assets. For a newly created work requirement, you may need an iterative approach where you generate and review the forecast, make small modifications, then regenerate the forecast to verify the outcome. This is an important step in validating the forecast modeling for an asset’s preventative maintenance during its initial definition.

However, for an existing forecast, you may wish to manage small changes in the Maintenance Forecasts page before considering any changes to its source work requirement. Care should always be taken when modifying the forecast methods of the source work requirement once work orders are created. Once a cadence is established using the work requirement start date, it may be difficult to adjust it depending on the modeling methods.

Note: It is also essential that you verify all your forecast modeling methods prior to creating the first work order for an asset in a work requirement. Else, you will be limited to the adjustments you can make to future due dates.

When editing a work requirement, it's important to understand the impact to the existing forecast, future forecast, and in general its history.

• Forecasts are always created and recreated from the date of the scheduled process into the future, per the forecast horizon in days.
• Existing forecast due dates, without work orders, are deleted and recreated. Past dates will be preserved and not updated.
• If a forecast due date has been manually set to be skipped, or a requested due date or organization is defined, then the forecast preserves the due date and it won't be deleted and recreated. Additionally, all previous due dates before this date are also preserved, which means the forecast won't delete or recreate them. This is necessary to maintain history of the manual edits and allow the forecasted due dates to be later edited as required.
• If a forecast due date is manually reset to not be skipped, have a requested due date or organization defined, then the forecast due dates will be reconsidered for deletion and recreation. This includes the individual due dates and all due dates before it in history that are not preserved due to manual edits.
• Work Orders are created only for due dates and times in the future. Therefore, it's possible that due dates in the past are skipped and not considered for work order creation if that scheduled process isn't scheduled to run on a routine basis. You can use the Manage Forecasts page or REST API to manually create these work orders, as required.
• Changes to the work requirement forecast method or parameters may fundamentally change the future forecast cadence. It becomes difficult to verify the original or last modeling in the past. Therefore, it may be advisable to create new work requirements when a significant change in cadence is required.
• If you use a Day or Meter interval, you must choose a method to calculate the next due dates as either using the base interval or the last completion. Once work orders are created for a forecast, changing this method could result in dramatic changes to a future forecast. Therefore, care should be taken when updating the method.

Manage the Forecast Using the Maintenance Forecast Page

You can also view and update the forecast using the Maintenance Forecasts page or REST API. For an asset’s due date, you can view key details about the asset and forecast method. Additionally, you can:

• Define and update a requested work order start date if a work order has not been created yet. This is helpful to manually define when a work order will be created, instead of its due date, by the work order creation scheduled process. The requested start date can only be set to a value between the last and next due date in the forecast.
• Define and update a requested maintenance organization where the work order will be created. If the due date is forecasted by a program that is enabled for cross-organizations, then you can request the work order to be created in another maintenance-enabled organization, instead of its forecasted organization. This requested organization must have the same work definitions in order to be defined.
• Skip a due date for an asset. In case you don't want to create a work order for a due date, set the skipped indicator to Yes. This won't consider the due date for work order creation using the work order creation scheduled process. You can also un-skip a due date, which allows the due date to be reconsidered by the next run of the schedule process if the due date is in the present or future. Else, you can manually create a work order.
• Manually create a work order for a due date. For a due date without an active work order, you can create a work order based off the forecasted due date and location, or by using the requested start date and location, if defined. This action will launch a scheduled process request to create the work order for the due date. This action is useful for due dates with a status of Planned, Unplanned, Skipped or Cancelled.
• Manually cancel a work order for a due date. You can cancel the work order if no transactions have been created in execution. Once cancelled, you can still view the work order reference, but it will be set to a status of Cancelled and will not be available for execution. The work order and due date will be considered by the forecast scheduled process to generate future due dates. Additionally, for a Cancelled work order on a due date, you can manually create a new work order in the same or requested location.

Note: If a work requirement is defined to create only the next work order, you can’t manually create any future work orders unless the last work order in history is in a status of completed or canceled.

Note: Care should be taken when you manually update a forecasted due date in the Manage Forecast UI or using the REST API. These updates fundamentally change the behavior of the forecast scheduled process. If you update a forecast due date with a requested due date or organization, or skip the due date, it won't be deleted and recreated by the forecast scheduled process. Additionally, all previous due dates before this date will also be preserved, which means the forecast won't delete or recreate them. Therefore, if you have a forecast method that relies on the last completion of a work order to recalculate the drive the next due date, you won't see an updated forecast due date if the date is preserved due to manual editing..