From the morphological to functional layout: comprehensive analysis and CNN-based prediction frames for carbon emissions in the cold region of University Campus

At the moment, more than half of the world's population lives in the metropolese regions, a number that grows to two thirds until the middle of the century [1]. Urban centers are associated with 80 % of the global energy requirements and 75 % of the emissions associated with climate change. When more and more people hike into cities [2]. Buildings as important urban components contribute significantly to energy consumption and emissions, which corresponds to 40 % of the total energy requirement and 30 % of the greenhouse gas product [3]. The combating of the energy requirements of buildings requires the optimization of urban morphology at regional level, since the structures are fundamentally connected to their surroundings. Urban energy consumption is shaped by factors such as shading effects and microclimatic conditions [4,5]While the efficiency of photovoltaic systems (PV) on the building surfaces depends on their spatial use in various geographical and climatic environments [6]. Therefore, it has become increasingly a focus of research [7]. With their high population density and carbon intensity, university grounds offer unique challenges to examine how morphological features are related to their complex layouts and diverse functions on emissions [8].

Many scientists have examined [9,10]Whereby some concentrate on how to build layout influences the energy intensity [11]. Vartholomaios et al. Simulations carried out to examine the effects of point, plate and closed layout patterns for the use of heating and cooling in residential areas of Thessaloniki, Greece [12]. Deng et al. Compared to the difference in the heating energy consumption between the courtyard and row layouts of residential complexes in Jinan, China by simulation [13]. Song et al. Analyzed the effects of five block layout parameters on the heating of the energy consumption of commercial and residential areas in Harbin, China, and found that the average height (AH) and the form coefficient (SC) had the most important effects [14]. Shareef et al. Examined the morphological elements that influence the cooling energy consumption in urban districts in the United Arab Emirates and show that a certain degree of heterogeneity reduces the solar radiation and lowers the air temperatures outdoors [15]. Manganese et al. Examined the morphological layout parameters that influence the energy consumption in residential areas in Istanbul, Turkey, and identify the average height as a key factor [16]. Leng et al. Discovered that factors such as the structure of footprint, the property ratio (PR), AH and the total area of the walls influence the energy that is required for warming in office buildings in Harbin, China [17]. Xu et al. emphasized the remarkable effects of factors such as SC, PR, AH and average depth on the energy consumption of office structures in Wuhan, China [18]. Xie et al. found that the building length, the SC and the building density (BD) are critical layout factors that influence the energy requirements in the dormitory in Wuhan [19].

In addition, research has examined the role of architectural layout in improving the capacity of solar energy [20]. Zhang et al. examined the PV potential of typical urban quarters in Singapore and found that morphological differences could increase PV generation by up to 200 % [6]. Zhao et al. Evaluates the potential of solar energy production for the construction of surfaces in residential areas of Adelaide, Australia, and discovered that roofs offer the highest potential of energy generation, while facades in the north take second place [21]. Xu et al. Analyzed the PV potential in residential areas of Wuhan and found that AH and Sky View Factor (SVF) significantly influenced electricity generation [18]. Tian et al. Examined comparison layout parameters that influence the solar power potential of residential areas in Wuhan, which shows that PR, BD, AH and building track had significant effects [22]. Xu et al. Researched further [23]. Studies by Xie et al. About a dormitory in Wuhan showed that AH, SC and SVF were the most influential parameters for the production of solar energy [19].

Two primary methods are currently being used to predict building services on a regional scale in urban areas: traditional methods and the AI methods for artificial intelligence (AI) [24]. The physical modeling is one of the traditional approaches [6]statistical methods [17]and regression methods [25]which are applicable to predict linear relationships. However, these methods can be less effective for complex, non -linear problems [26]. University site with their diverse layouts and complex functions [27]Current challenges that traditional methods may have difficulties to tackle because the physical modeling consumes a lot of time and computing power [28,29]and predictions based on linear relationships [30]. As a result, AI methods offer a valuable alternative. Deep Learning uses multi -layer neural architectures as a branch of AI to extract hierarchical characteristics from data records [31,32]. This approach, in particular by folding networks (CNNS), Excels in the feature extraction and non -linear [33]. In connection with the establishment of morphology in the urban regional scale, CNNS can code the spatial layouts of building clusters and geometries, which effectively extracts and quantified characteristics [34,35]. Research has successfully integrated CNNS with urban morphology data records to develop predictive models for temperature, heat comfort and air pollution [[36]Present [37]Present [38]]. Despite these progress, CNNs were less often applied to regional carbon emission models based on urban layout patterns [39].