> For the complete documentation index, see [llms.txt](https://docs.buildings.ability.abb/collection/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.buildings.ability.abb/collection/platform/platform-core/baseline-and-factors/portfolio-benchmark-and-targets/create-a-baseline.md).

# Create a Baseline

Create a baseline to define the historical operation used to calculate target values. Choose the reference according to the question you want to answer: performance against a known past period, departures from recent operation, or when historical operation changed.

You need a [-> saved B\&T config](https://docs.buildings.ability.abb/collection/platform/platform-core/baseline-and-factors/portfolio-benchmark-and-targets/create-a-target.md) and historical data for the source and any selected feature attributes.

<figure><img src="https://2013058394-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FOqGnkz52o2zVUsEGEZGk%2Fuploads%2Fgit-blob-4b65357c3b2bc10de1d37d7835955cb82e09b0bd%2Fbenchmark-targets-configure.png?alt=media" alt="Target Line Settings with saved periods, manual date fields, Save target, and Add another date range"><figcaption><p>Fig. 3: Baseline period and general settings</p></figcaption></figure>

## Open the baseline form

1. [-> Open Benchmark and Targets](https://docs.buildings.ability.abb/collection/platform/platform-core/baseline-and-factors/portfolio-benchmark-and-targets/open-benchmark-and-targets.md).
2. Click **Configure target** for the required B\&T config.
3. Click **Add another date range** **\[A]** in Fig. 3 if the B\&T config already has saved baselines.

<figure><img src="https://2013058394-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FOqGnkz52o2zVUsEGEZGk%2Fuploads%2Fgit-blob-b2c69a964e872a884242f58ea35c9cddd2134a67%2Fbenchmark-targets-baseline-choice.png?alt=media" alt="New baseline form with Auto marked A, Manually marked B, Set Date marked C, and Moving baseline marked D"><figcaption><p>Fig. 13: Choose automatic detection, fixed dates, or a moving baseline</p></figcaption></figure>

Adding a range opens **Reference period of best correlation**. In Fig. 13, select **Auto** **\[A]** to find historical periods automatically, or **Manually** **\[B]** to choose **Set Date** **\[C]** or **Moving baseline** **\[D]**. Selecting an option does not save a new baseline.

## Choose a baseline approach

| Approach                               | Use it when                                                                            | Example and tradeoff                                                                                                                                                                                            |
| -------------------------------------- | -------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Manually**, then **Set Date**        | You know which historical period represents the operation you want to compare against. | Keep a stable period before an energy-efficiency project as the reference for measuring its effect. The reference does not advance with time.                                                                   |
| **Manually**, then **Moving baseline** | You want the comparison to adapt to a rolling period of recent operation.              | Monitor unexpected energy use relative to recent building behaviour. A lasting deterioration can eventually enter the reference period, so keep a fixed comparison as well when long-term drift matters.        |
| **Auto**                               | You are unsure where meaningful changes occurred in the historical data.               | Find candidate periods before and after a change in plant, occupancy, or operating schedules, then review which period is appropriate. Detection is retrospective and does not establish the cause of a change. |

## Use fixed dates

Choose a period with representative operation and sufficient source and feature data. For an energy-reduction project, use a period before the intervention rather than one that already contains the improvement. Where seasonal behaviour matters, include the relevant operating seasons rather than selecting only an unusually quiet week.

1. Select **Manually**.
2. Select **Set Date**.
3. Enter the start date in **From**.
4. Enter the end date in **To**.\
   ◈ The end date must be later than the start date.
5. Click **Save target** **\[B]** in Fig. 3.\
   ▷ The saved period appears in the baseline list.

{% hint style="info" %}
**Apply** also saves the date range, but leaves range editing open. It is not a preview-only action. **Save target** saves the range and completes range editing.
{% endhint %}

## Use a moving baseline

Use this approach for a comparison that advances with time, such as operational monitoring with alarms on generated deviations. It is not an unchanged reference for proving long-term savings.

1. Select **Manually**.
2. Select **Moving baseline**.
3. Select **Offset days**.
4. Select **Baseline length**.
5. Click **Save target** **\[B]** in Fig. 3.\
   ▷ The baseline list shows the moving period's length and offset.

The reference period ends the selected number of offset days before today's start. Its start is the selected baseline length before that end. For example, a length of 365 days and an offset of 30 days uses a 365-day reference period ending 30 days before today.

### Offset days

**Offset days** leaves a gap between today and the end of the reference period. Use a gap to keep the most recent operating change out of the comparison temporarily, or to avoid a recent period whose readings are not yet complete. A larger gap preserves older conditions for longer but may make the reference less representative of current operation.

This is a time offset, not the percentage reduction configured under **General**. It does not create a 10% reduction goal; use the [-> baseline's Offset setting](https://docs.buildings.ability.abb/collection/platform/platform-core/baseline-and-factors/portfolio-benchmark-and-targets/configure-a-baseline.md#offset) for that.

### Baseline length

**Baseline length** controls how much history contributes to the moving reference. A longer period can include more seasonal and operating variation and changes more slowly as new days enter it. A shorter period follows recent operating conditions more closely but is more influenced by temporary events.

For example, a year-long reference can be useful for a seasonal building, while a shorter reference can help investigate recent operating changes. Choose it according to the available data and the behaviour you need to compare, then inspect the preview rather than assuming one duration suits every building.

## Detect baselines automatically

Use automatic detection when you do not know where a building's historical energy behaviour changed. Instead of making you choose every date boundary, the analysis looks for periods with a reasonably consistent relationship between measured operation and the modelled target.

The analysis compares measured values with an expected operating pattern, smooths the differences over the chosen window, and looks for departures large enough to indicate another operating period. It proposes periods that you can inspect and use as baselines. Smoothing helps distinguish sustained changes from isolated unusual days; the results still need your review alongside known changes in the building.

This analyses the past. It is not a continuously running alarm, does not determine the cause of a change, and does not prove that every detected period is suitable as your reference. For ongoing monitoring, activate an appropriate baseline and use its generated deviation attributes in an alarm rule.

{% hint style="warning" %}
Running **Auto detect** replaces this B\&T config's previous automatically detected baselines, including any that were active. Their generated outputs are subject to background cleanup. Manually created baselines are not replaced. Review the new baselines and activate the ones you want to use.
{% endhint %}

<figure><img src="https://2013058394-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FOqGnkz52o2zVUsEGEZGk%2Fuploads%2Fgit-blob-ff42e02ad9a82dfb16b68bb3d2e09762a094e7d0%2Fbenchmark-targets-auto-detect.png?alt=media" alt="Automatic detection form showing drift window, minimum baseline duration, drift threshold, and Auto detect"><figcaption><p>Fig. 4: Automatic baseline detection</p></figcaption></figure>

1. Select **Auto** in the reference-period selector.
2. Enter **Drift Detection Window (Days)** **\[A]** in Fig. 4.
3. Enter **Minimum Baseline Duration (Days)** **\[B]**.
4. Enter **Baseline Drift Threshold (%)** **\[C]**.
5. Click **Auto detect** **\[D]**.\
   ▷ The detected periods appear as saved baselines. New automatically detected baselines are paused.
6. Review each period in [-> Correlation period](https://docs.buildings.ability.abb/collection/platform/platform-core/baseline-and-factors/portfolio-benchmark-and-targets/review-a-target.md#review-the-correlation-period) and compare it with known changes in operation.

### Drift Detection Window (Days)

This is the number of days used to smooth the comparison. Enter a positive whole number.

Use a longer window when ordinary day-to-day fluctuations are being mistaken for operating changes. It gives more weight to sustained behaviour, but can make a shorter change harder to distinguish or its timing less precise. Use a shorter window when investigating shorter-lived changes, accepting that more normal variation can influence the result.

### Minimum Baseline Duration (Days)

This setting controls the duration used to establish a candidate operating period. Enter a positive whole number. Use a longer duration when you want long-lived operating regimes rather than short interruptions; consider a shorter duration when a relevant occupancy pattern or operating phase lasted only a shorter period.

Check the dates of the returned periods against the source history. Do not assume the label means that every returned range is an exact requested duration or that a shorter period is automatically a fault.

### Baseline Drift Threshold (%)

This sets how large a departure must be before it is treated as a change rather than ordinary variation. Enter a value from 0 to 100; decimals are allowed.

Lower the threshold when meaningful small changes are being missed. Raise it when normal variation produces too many candidate changes. For example, a small drift threshold may help investigate a subtle change in steady demand, while a larger threshold focuses attention on more pronounced changes. It is a detection sensitivity setting, not the percentage-reduction target configured under **General**.

### Review and tune the result

Start with the available defaults and inspect the returned periods. Change one setting at a time so you can see its effect. If a brief exceptional event splits otherwise similar operation, consider more smoothing or a higher threshold. If an important change is absent, consider less smoothing or a lower threshold, provided the historical data supports the comparison.

Before rerunning detection, remember that it replaces the B\&T config's previous automatically detected baselines. If detection is unavailable, confirm that the B\&T config is saved and its historical data has loaded, and review any displayed data-loading or validation error.

Next, [-> configure the baseline](https://docs.buildings.ability.abb/collection/platform/platform-core/baseline-and-factors/portfolio-benchmark-and-targets/configure-a-baseline.md). Saving the baseline does not activate it.


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.buildings.ability.abb/collection/platform/platform-core/baseline-and-factors/portfolio-benchmark-and-targets/create-a-baseline.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
