Configuration

Rules are YAML files. Each rule has name, type, flag, inputs, and optionally params and expression.

Rule types

Type	Description
bounds	Value outside [low, high]; supports units: metric
flatline	Sensor stuck (rolling spread < tolerance)
expression	Pandas/numpy expression
hunting	Excessive AHU state changes (PID hunting)
oa_fraction	OA fraction / design airflow error
erv_efficiency	ERV effectiveness out of range

Expression rule

name: my_rule
type: expression
flag: my_flag

inputs:
  col_a:
    column: actual_df_column_name
  col_b:
    column: other_column

params:
  thres: 0.1

expression: |
  (col_a < col_b - thres) & (col_a > 0)

Bounds rule (bad data)

Inputs use BRICK class names; column_map keys match:

name: bad_sensor_check
type: bounds
flag: bad_sensor_flag

params:
  units: imperial

inputs:
  Supply_Air_Temperature_Sensor:
    brick: Supply_Air_Temperature_Sensor
    column: Supply_Air_Temperature_Sensor
    bounds:
      imperial: [40, 150]
      metric: [4, 66]
  Return_Air_Temperature_Sensor:
    brick: Return_Air_Temperature_Sensor
    column: Return_Air_Temperature_Sensor
    bounds:
      imperial: [40, 100]
      metric: [4, 38]

Flatline rule

name: sensor_flatline
type: flatline
flag: flatline_flag

inputs:
  Supply_Air_Temperature_Sensor:
    brick: Supply_Air_Temperature_Sensor
    column: Supply_Air_Temperature_Sensor
  Outside_Air_Temperature_Sensor:
    brick: Outside_Air_Temperature_Sensor
    column: Outside_Air_Temperature_Sensor

params:
  tolerance: 0.000001
  window: 12

BRICK metadata in YAML

Rules can include brick and equipment_type for documentation and future filtering:

name: oat_too_high_free_cooling
type: expression
flag: fc9_flag
equipment_type: [AHU, VAV_AHU]

inputs:
  Outside_Air_Temperature_Sensor:
    brick: Outside_Air_Temperature_Sensor
    column: oat
  Supply_Air_Temperature_Setpoint:
    brick: Supply_Air_Temperature_Setpoint
    column: sat_setpoint
  Damper_Position_Command:
    brick: Damper_Position_Command
    column: economizer_sig

• brick — Brick class name. Used for column resolution: column_map can be keyed by Brick class (e.g. Supply_Air_Temperature_Sensor), and the runner resolves columns via BRICK first. When the same Brick class appears multiple times (e.g. two Valve_Command), use BrickClass|rule_input in the column map.
• equipment_type — Equipment types this rule applies to. When using run_all_rules_brick.py with a Brick TTL, only rules whose equipment_type matches the model’s ofdd:equipmentType are run. See Data Model & Brick.

---

Next: API Reference — RuleRunner, brick_resolver, example scripts