建築空調負載預測模型開發與驗證

Abstract

本研究透過建立一套可即時連動中央氣象局開放資料平台之空調負載預測系統，以優化建築空調負載模擬之準確性與時效性。透過中央氣象局開放資料平台提供之API界面，定期擷取指定地區之氣象預報數據，涵蓋氣溫、露點溫度、相對溼度、風速、降雨量與日照量等關鍵氣象參數，並與氣象預報更新頻率同步自動執行。取得之即時預報與觀測資料將覆蓋於既有EPW（EnergyPlus Weather Data File）格式檔案中，更新其對應欄位以反映最新氣象情境，並將其作為空調負載模擬預測系統之輸入，進行建築空調負載之預測分析。本研究透過Python程式語言開發，確保氣象資料處理與EPW檔轉換之自動化與靈活性，結合以Matlab為基礎所開發之空調負載預測系統，達成動態預測空調負載之功能性。透過本系統之整合，將能有效反映氣象變化對建築能耗之即時影響，提供未來建築能源管理與智慧控制策略之決策依據。

This study establishes a real-time integrated air conditioning load prediction system utilizing the Open Data Platform of the Central Weather Administration (CWA), aiming to enhance the accuracy and timeliness of building HVAC simulations. Through the API services provided by the CWA, the system periodically retrieves weather forecast data for designated locations, including key meteorological parameters such as air temperature, dew point temperature, relative humidity, wind speed, precipitation, and solar radiation. The update frequency is synchronized with the release schedule of official weather forecasts.The retrieved real-time forecast and observational data are mapped onto existing EPW (EnergyPlus Weather Data File) templates, with relevant fields updated to reflect the latest weather scenarios. These updated EPW files are then used as inputs for the HVAC load simulation system to perform predictive analysis of building air conditioning loads.The system’s data processing and EPW file transformation modules are developed using Python, ensuring automation and flexibility. The HVAC load prediction module, implemented in Matlab, enables dynamic simulation based on real-time weather inputs. Through the integration of these components, the system effectively captures the immediate impacts of weather changes on building energy consumption, offering valuable insights for future energy management strategies and intelligent control systems in buildings.