Development of an Intelligent Control System for Building with Self-respiration Capability

[[abstract]]為提供人體所需，建築內部需能適時進行空氣調節。典型的空調系統雖能因應溫濕度的變異，以調節室內空氣；然其易損耗相當的成本與能源，更何況此種封閉式架構，實難以適時的排除危害氣體。相對的，開放式通風架構雖消耗較少能源，可提供相當的室內空氣條件，然其所導致的噪音與粉塵，亦將降低整體室內品質。因此，如何同時考量空氣安全、舒適條件與操作成本等多方面需求，實為目前建築控制系統的重要議題。 本論文提出一智慧型控制系統，以提供建築具有調節呼吸的能力。在架構規劃上，本系統可根據需求，彈性設定不同感測器於不同位置，包含保護人體安全的一氧化碳和烷系感測器；保持環境舒適度的溫濕度、二氧化碳與空氣品質感測器；與監視能源消耗的電流計等。經設定後，系統將可對應其指標如安全、舒適度與成本等進行監控。在空氣調節分析上，可根據自適應性網路模糊推論系統(Adaptive Network-based Fuzzy Inference System ANFIS)，以自動建立建築內氣體動力模型。最後在空氣調節控制上，則可透過灰預測方法，基於建築氣體模型的推估，以事先或動態調節設備，以達到舒適度與成本的最舒適控制。 在實驗上，透過兩種不同建築結構，分別為小木屋與獨棟木屋，前者具自動窗與多重不同風扇，後者則配以動態通風井與多重抽風扇，兩者均透過所提系統與感測器，以整合其自然與強制對流機制。在低碳屋中利用煙囪效應使室內降溫的實驗下，可以得知未使用智慧型控制方法時室內溫度會低於室外溫度一度。若是再藉由抽風扇加強對流的情況下，室內可較室外溫度低約三度，因此利用建築設計架構與自然對流結合機械通風可以增加房屋室內的對流性，以達到室內降溫與空氣循環的效果。透過多階段累積危害氣體濃度的狀態機，本系統除可提供特定氣體自動警示與排除，具安全保護能力外；系統更結合ANFIS與灰預測控制法，在保持室內外溫差達三度以上，提供具維持舒適度的能力，並進一步較條件控制法減少16%的能源消耗。[[abstract]]A building should timely regular air conditions for human body needs Although an air conditioner is popularly used to regular variations of temperature and humidity it costs high investment and energy Moreover this kind of close structure fails to timely extract dangerous gas to outside An open ventilation structure consumes less power while providing well air conditions; however it reduces indoor quality by causing high noise and more particles Therefore an important issue of development of a building control system is how to reach requirements i e safety comfort and cost simultaneously This work proposes an intelligent control system enabling a building with self-respiration capability In structure configuration this system allows users to flexibly configure various locations and sensors such as safety protection e g CO and alkane sensors; comfort keeping e g temperature humidity CO2 and air-quality; energy monitoring by current meter After configuration of sensing sources this system then can monitor corresponding indices including safety comfort and cost In gas analysis an adaptive network-based fuzzy inference system (ANFIS) is adopted to model gas dynamics of a building Finally by using a grey prediction with the gas dynamics estimation the system can regulate devices in advance for reaching both objectives of comfort and cost In experiments two different buildings without any air conditioner were performed in which a small cabin is with auto windows and various fans; a two-floor townhouse is with a ventilate well and four fans Both buildings are integrated by the proposed control system to provide ventilation for self-respiration Experimental results show that the proposed system can detect and extract dangerous gas for protecting safety based on a multi-level a state machine which transits operating states by accumulating densities of dangerous gas Furthermore while comparing to a condition-based method this system keeps lower temperature than outdoors 3℃ and decreases lower energy consumption 16 % by integrating ANFIS estimation and grey predictio