> For the complete documentation index, see [llms.txt](https://docs.myrspoven.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.myrspoven.com/ai-features/ai-penalty.md). # AI penalty myCoreAI balances several competing goals on every optimization cycle — occupant comfort, equipment efficiency, and operating cost. Each undesirable outcome carries a penalty, and myCoreAI selects the candidate plan with the lowest total penalty over the optimization horizon. This page describes how those priorities are ordered. ## How the penalty system works On every cycle, myCoreAI evaluates many candidate setpoint plans across a 72-hour horizon, predicts the outcome of each against the building model, sums the penalties, and selects the lowest-cost plan. Only the first hour is dispatched to the Building Management System (BMS); the remaining horizon ensures the current decision leads to a coherent trajectory. Penalties are not equal in weight, and several are conditional on season, time of day, or occupancy. ## Priority 1 — Comfort Comfort penalties dominate. myCoreAI weights them far above any energy or cost term, so it accepts higher energy use rather than allow comfort to drift outside the comfort bounds. Comfort penalties cover: * **Indoor temperature** outside the comfort bounds * **CO₂ levels** above the configured limit * **Humidity** above the configured limit, where measured For buildings with sensor groupings configured, myCoreAI evaluates comfort per group (typically one group per heating or cooling circuit) rather than per individual sensor. ## Priority 2 — Air handling unit behavior For each air handling unit (AHU), myCoreAI chooses between heating, cooling, or ventilation-only based on outdoor temperature and season. | Season | AHU role | Penalty | | ---------- | ------------------------ | ------------------------------------------------------------------------------------- | | Winter | Heating via AHU | High — discouraged unless comfort requires it or **Heat with ventilation** is enabled | | Summer | Free cooling via AHU | Low — encouraged before mechanical cooling | | Year-round | Fan / pressure operation | Small constant — discourages unnecessary ventilation | ## Priority 3 — Heating source myCoreAI assigns a low penalty to the primary heating circuits (typically district heating via radiators) and a high penalty to AHU-based heating. This ranking produces the winter logic above. A smoothness term discourages rapid hour-to-hour changes in heating setpoints. The **Max change heating circuit** setting in AI settings caps how fast setpoints can move. ## Priority 4 — Cooling source Cooling sources carry light penalties, ordered by cost and environmental impact: free cooling via ventilation, then district cooling, then mechanical cooling. ## Energy price modulation When [myLoadshift](/ai-features/loadshift.md) is active, myCoreAI multiplies energy-related penalties by the time-varying spot price and shifts consumption toward cheaper hours. Comfort penalties are not price-modulated. ## Energy cap When energy cap is configured, an additional penalty activates as the predicted daily average approaches the cap. myCoreAI weights it heavily — the cap behaves as a near-hard constraint. ## Putting it together myCoreAI scores each candidate plan on: 1. Comfort penalties — high, usually zero when comfort is maintained 2. Equipment-use penalties — moderate, season- and temperature-dependent 3. Smoothness penalties — small, keep operation stable 4. Cost-related penalties — small, favor cheaper energy when other constraints are satisfied 5. Cap penalties — active only near the energy cap The lowest-total-cost plan is dispatched. ## Worked example A district-heated office on a mild February afternoon: outdoor temperature is 6 °C, indoor sensors sit mid-band, and the spot price drops sharply between 13:00 and 15:00. With myLoadshift active, myCoreAI lifts the radiator supply slightly during the cheap window to pre-charge the building mass, then lets it coast through the early evening price peak. Comfort penalties stay at zero throughout — the move is driven entirely by the cost term, bounded by the smoothness cap on the heating circuit. ## Interpreting unexpected behavior When setpoints look counterintuitive — radiator supply unchanged on a mild day, AHU running warmer than expected — a higher-priority penalty is typically active: a comfort sensor near its limit, the smoothness term preventing a fast drop, or the building approaching its energy cap. To bias myCoreAI toward more aggressive energy savings, adjust AI settings: widen the adaptive min/max, enable AI Freedom, or increase off-hours offsets. --- # 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 dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter: ``` GET https://docs.myrspoven.com/ai-features/ai-penalty.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal. 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.