The challenge of achieving net zero emissions by 2050 necessitates to drastically reduce the energy consumption of the buildings sector, which alone contributes for almost 40% of the anthropic CO2 emissions. In recent years, growing attention has been paid to the minimization of costs and emissions of large buildings by optimizing the design and operation of their energy systems. Most of the works available in the literature consider a limited set of technologies to meet only the electrical and heating demands. In this work an innovative model is developed for the optimization of the energy hub of a hospital, characterized by 5 distinct energy demands Electrical, Heating, Cooling, Steam and Sanitary Hot Water. To satisfy such demands, the proposed model considers a large set of technologies including cogeneration units, absorbers, reversible, and multipurpose heat pumps. The tool consists of two sequential Python codes for the size optimization and the operation optimization, respectively. The first code defines the size of each technology considering the energy needs of four representative days obtained using the k-medoid method, with two extreme days added to ensure load supply. A preliminary investigation demonstrated that this clustering technique leads to an overestimation in the objective function of less than 3%, compared to a 365-days simulation. Therefore, the computational time is drastically reduced with a small impact on the results accuracy. The optimal sizes obtained from the first part of the model are used as input for the second code, which defines the optimal operation of the system throughout the year. The optimization problems are solved using the Python extension of Gurobi, a software that allows for the assessment of MILQP and MILP models by exploiting the branch and bound method. Two different objective functions are considered: total annual costs and CO2 emissions. First code results show that, according to the considered objective function, a reduction of 14.4% in the annual costs or a reduction of 12.36% in the CO2 emissions can be reached respect a non-optimal reference configuration, with a difference between the two optimized solutions of 20.60% in the CO2 emissions and of 42.07% in the total costs. Between the two solution, the one provided by the economical optimization is chosen for the entire year simulation. Once selected the technologies to be used, the simulation of the model was performed resulting in a cost increment respect the first code of 4.44 %, and an increase of CO2 emissions of 4.5%.
La sfida di raggiungere zero emissioni nette entro il 2050 impone una drastica riduzione dei consumi del settore edilizio, che da solo contribuisce per quasi il 40% alle emissioni antropiche di CO2. Negli ultimi anni, è cresciuta l’attenzione verso la minimizzazione dei costi e delle emissioni degli edifici di grandi dimensioni attraverso l’ottimizzazione della progettazione e del funzionamento dei loro sistemi energetici. Gran parte degli studi presenti in letteratura considera un numero limitato di tecnologie per soddisfare esclusivamente la domanda di energia elettrica e di riscaldamento. In questo lavoro è stato sviluppato un modello innovativo per l’ottimizzazione del polo energetico di un ospedale, caratterizzato da cinque diverse tipologie di domanda energetica: elettricità, riscaldamento, raffrescamento, vapore e acqua calda sanitaria. Per soddisfare queste domande, il modello proposto considera un ampio set di tecnologie, tra cui unità di cogenerazione, assorbitori, pompe di calore reversibili e multiuso. Lo strumento sviluppato consiste in due codici Python sequenziali per l’ottimizzazione delle dimensioni e del funzionamento del sistema. Il primo codice determina la dimensione di ciascuna tecnologia in base ai fabbisogni energetici di quattro giorni rappresentativi, ottenuti mediante il metodo k-medoid, con l’aggiunta di due giorni estremi per garantire la copertura del carico. Un’analisi preliminare ha dimostrato che questa tecnica di clustering porta a una sovrastima della funzione obiettivo inferiore al 3% rispetto a una simulazione su 365 giorni, riducendo drasticamente il tempo computazionale con un impatto minimo sull’accuratezza dei risultati. Le dimensioni ottimali ottenute nella prima fase vengono poi utilizzate come input per il secondo codice, che definisce il funzionamento ottimale del sistema durante tutto l’anno. I problemi di ottimizzazione vengono risolti utilizzando l’estensione Python di Gurobi, un software che consente la valutazione di modelli MILQP e MILP sfruttando il metodo branch and bound. Sono considerate due diverse funzioni obiettivo: il costo totale annuo e le emissioni di CO2. I risultati del primo codice mostrano che, a seconda della funzione obiettivo adottata, si può ottenere una riduzione del 14,4% dei costi annuali o una riduzione del 12.36% delle emissioni di CO2 rispetto ad una configurazione di riferimento non ottimizzata, con una differenza tra le due soluzioni ottimizzate del 20,60% nelle emissioni di CO2 e del 42,07% nei costi totali. Tra le due soluzioni, quella ottenuta dall’ottimizzazione economica è scelta per la simulazione annuale. Una volta selezionate le tecnologie da utilizzare, i risultati del secondo codice mostrano un incremento dei costi rispetto al primo codice del 4,44%, e un aumento di emissioni del 4.5%.
Development of a tool for the sizing and operation optimization of the energy hub of a new hospital in Legnago
PURI, TIBERIO
2024/2025
Abstract
The challenge of achieving net zero emissions by 2050 necessitates to drastically reduce the energy consumption of the buildings sector, which alone contributes for almost 40% of the anthropic CO2 emissions. In recent years, growing attention has been paid to the minimization of costs and emissions of large buildings by optimizing the design and operation of their energy systems. Most of the works available in the literature consider a limited set of technologies to meet only the electrical and heating demands. In this work an innovative model is developed for the optimization of the energy hub of a hospital, characterized by 5 distinct energy demands Electrical, Heating, Cooling, Steam and Sanitary Hot Water. To satisfy such demands, the proposed model considers a large set of technologies including cogeneration units, absorbers, reversible, and multipurpose heat pumps. The tool consists of two sequential Python codes for the size optimization and the operation optimization, respectively. The first code defines the size of each technology considering the energy needs of four representative days obtained using the k-medoid method, with two extreme days added to ensure load supply. A preliminary investigation demonstrated that this clustering technique leads to an overestimation in the objective function of less than 3%, compared to a 365-days simulation. Therefore, the computational time is drastically reduced with a small impact on the results accuracy. The optimal sizes obtained from the first part of the model are used as input for the second code, which defines the optimal operation of the system throughout the year. The optimization problems are solved using the Python extension of Gurobi, a software that allows for the assessment of MILQP and MILP models by exploiting the branch and bound method. Two different objective functions are considered: total annual costs and CO2 emissions. First code results show that, according to the considered objective function, a reduction of 14.4% in the annual costs or a reduction of 12.36% in the CO2 emissions can be reached respect a non-optimal reference configuration, with a difference between the two optimized solutions of 20.60% in the CO2 emissions and of 42.07% in the total costs. Between the two solution, the one provided by the economical optimization is chosen for the entire year simulation. Once selected the technologies to be used, the simulation of the model was performed resulting in a cost increment respect the first code of 4.44 %, and an increase of CO2 emissions of 4.5%.File | Dimensione | Formato | |
---|---|---|---|
Puri_Tiberio.pdf
accesso riservato
Dimensione
6.57 MB
Formato
Adobe PDF
|
6.57 MB | Adobe PDF |
The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License
https://hdl.handle.net/20.500.12608/82352