This study employs the Non-dominated Sorting Genetic Algorithm II (NSGA II) to optimise the operational strategies of the Hoa Binh reservoir, situated on the Da River in Vietnam. Commencing operations with its first turbine activated in December 1989 and achieving full functionality in 1994, the reservoir serves as a crucial project for energy generation, flood regulation, and downstream water provisioning to the Red River Delta (RRD). Because of its pivotal role, there have been several scholarly investigations into the reservoir's operation, but their examinations were mostly implemented using data before 2010. Given the changes in the upper reservoirs system, shifts in water usage demands, and changes in the morphology of the Red River downstream, updating research on the operation of the Hoa Binh reservoir is necessary. To better understand the problem, the study first identified the significant changes in the river system and how these changes have affected Hoa Binh reservoir operations. Notably, the construction of several major upstream reservoirs, including Lai Chau, Son La, Huoi Quang, and Ban Chat since 2011, has influenced on the flow regime and sediment transport dynamics of the Da River. Among these, we verified how the Son La reservoir has affected the Hoa Binh reservoir inflow regime and operational strategies. Downstream in the RRD, pronounced alterations in river morphology have been observed. The sedimentation reduction due to the deposition in the upstream reservoir, has resulted in riverbed scouring, culminating in a substantial decrease in water levels, for a given discharge, at the key monitoring station Ha Noi. Consequently, to reach a certain water level in Ha Noi, more water release from reservoirs is required. After identifying these main issues, the NSGA II model was applied to optimise the operational strategies of the Hoa Binh reservoir. This optimization framework encompasses multiple objectives: focusing on the dry season the minimization of water release deviation from the demand and the maximization of hydropower generation . Concurrently, the optimisation framework must satisfy operational constraints, including storage continuity, minimum reservoir elevation, and the daily average minimum release specified by governmental regulations. Key input parameters for the optimisation model include daily reservoir inflow, water demand within the Red River Delta, and the reservoir's storage-water level equation. The decision variables are the daily outflow from Hoa Binh. The optimisation model was run for a 22-year period, segmented into two sub-periods (2001-2011 and 2012–2022), mirroring the commencement of Son La reservoir operations. Results show notable increments in energy production and reservoir water levels, while simultaneously satisfying downstream demands in the RRD. Flood control objectives were not optimized within this study; however, safety measures during flood seasons were upheld, ensuring the normal water levels of the reservoir and the third alarm water level threshold downstream in Ha Noi. On average the optimised regulation ensured a 10% increase of energy generation over the 2001-2011 period and a 2% increase in the 2012-2022, ensuring always the consumptive water demand and the dam safety. Finally, the study presents a comparative graph depicting the median trajectory of optimal and historical reservoir water levels across the aforementioned periods. In the present contexts, operators of the Hoa Binh reservoir face end-of-dry-season water shortages, sometimes causing the reservoir to operate at its dead level, risking electrical network stability and causing load shedding in the northern region. Derived from optimal water levels, the maximum reservoir water level at the end of the dry season can be increased. Moreover, operators may utilize these graphical representations as support while making decisions regarding reservoir operation.
Questo studio utilizza l'Algoritmo Genetico di Ordinamento Non Dominato II (NSGA II) per migliorare le strategie operative del bacino di Hoa Binh, situato sul fiume Da in Vietnam. Il bacino rappresenta un progetto cruciale per la generazione di energia, la regolazione delle inondazioni e la fornitura di acqua a valle del Delta del Fiume Rosso (RRD). Sono stati condotti diversi studi; tuttavia, la maggior parte di essi si è basata su dati precedenti al 2010, prima dell’entrata in funzione di serbatoi a monte. Considerando le modifiche nel sistema di bacino superiore, i cambiamenti nella domanda di uso dell'acqua e le alterazioni nella morfologia del Fiume Rosso a valle, diventa imperativo aggiornare la ricerca sull'operatività del bacino di Hoa Binh. Questo studio mira innanzitutto ad identificare i cambiamenti significativi verificatesi nel sistema fluviale e come questi cambiamenti hanno influenzato le operazioni del bacino. In particolare, la costruzione di diversi bacini a monte, ha influenzato il regime di flusso e la dinamica del trasporto di sedimenti del Fiume Da. Tra questi, il bacino di Son La ha gli effetti più diretti sulle strategie operative del bacino di Hoa Binh. A valle nel RRD, sono state osservate significative alterazioni nella morfologia del fiume. La riduzione della sedimentazione ha causato, infatti, un approfondimento del letto del fiume, con una sostanziale diminuzione dei livelli idrici presso la stazione di monitoraggio chiave Ha Noi, a parità di portata. Di conseguenza, per raggiungere un certo livello dell'acqua a Ha Noi, è necessario rilasciare più acqua dai bacini a monte. Dopo aver identificato il problema principale, il modello NSGA II è stato applicato per ottimizzare le strategie operative del bacino. Questo framework di ottimizzazione comprende molteplici obiettivi, tra cui la minimizzazione della deviazione del rilascio dell'acqua dalla domanda e la massimizzazione della generazione di energia idroelettrica durante la stagione secca. Contestualmente, il framework di ottimizzazione deve soddisfare vincoli operativi, tra cui la continuità dello stoccaggio, il livello minimo di invaso, il rilascio minimo medio giornaliero e il non superamento del livello di invaso massimo specificato dalle normative governative. I parametri di input chiave per il modello di ottimizzazione includono il flusso giornaliero nel bacino, la domanda di acqua all'interno del RRD e l'equazione di conservazione dei volumi idrici per il bacino. Le variabili decisionali sono il deflusso giornaliero da Hoa Binh. Il modello di ottimizzazione è stato eseguito per un periodo di 22 anni, suddiviso in due sottoperiodi (2001-2011 e 2012-2022) che riflettono l'inizio delle operazioni del bacino di Son La. I risultati mostrano notevoli incrementi nella produzione di energia e nei livelli dell'acqua del bacino, soddisfacendo contemporaneamente le domande a valle nel RRD e i vincoli operativo. Gli obiettivi di controllo delle inondazioni non sono stati ottimizzati all'interno di questo studio; tuttavia, le misure di sicurezza durante le stagioni delle piene sono state rispettate, garantendo i normali livelli dell'acqua del bacino e la soglia del terzo allarme per i livelli idrometrici a valle, ad Ha Noi. In media la regolazione ottimizzata ha garantito un aumento del 10% della produzione di energia nel periodo 2001-2011 e un aumento del 2% nel periodo 2012-2022, garantendo sempre la domanda irrigua, potabile e industriuale e la sicurezza della diga. Infine, lo studio presenta un grafico comparativo con la traiettoria mediana dei livelli ottimali e storici del bacino corrispondenti ai periodi sopra menzionati. Mantenendo livelli di invaso ottimali, l’invaso alla fine della stagione secca può essere significativamente aumentato. Gli operatori possono utilizzare le rappresentazioni grafiche ottenute come supporto per prendere decisioni riguardanti la regolazione del bacino.
OPTIMISATION OF THE MULTI-PURPOSE HOA BINH RESERVOIR DURING THE DRY SEASON USING GENETIC ALGORITHMS
NGUYEN, HAI YEN
2025
Abstract
This study employs the Non-dominated Sorting Genetic Algorithm II (NSGA II) to optimise the operational strategies of the Hoa Binh reservoir, situated on the Da River in Vietnam. Commencing operations with its first turbine activated in December 1989 and achieving full functionality in 1994, the reservoir serves as a crucial project for energy generation, flood regulation, and downstream water provisioning to the Red River Delta (RRD). Because of its pivotal role, there have been several scholarly investigations into the reservoir's operation, but their examinations were mostly implemented using data before 2010. Given the changes in the upper reservoirs system, shifts in water usage demands, and changes in the morphology of the Red River downstream, updating research on the operation of the Hoa Binh reservoir is necessary. To better understand the problem, the study first identified the significant changes in the river system and how these changes have affected Hoa Binh reservoir operations. Notably, the construction of several major upstream reservoirs, including Lai Chau, Son La, Huoi Quang, and Ban Chat since 2011, has influenced on the flow regime and sediment transport dynamics of the Da River. Among these, we verified how the Son La reservoir has affected the Hoa Binh reservoir inflow regime and operational strategies. Downstream in the RRD, pronounced alterations in river morphology have been observed. The sedimentation reduction due to the deposition in the upstream reservoir, has resulted in riverbed scouring, culminating in a substantial decrease in water levels, for a given discharge, at the key monitoring station Ha Noi. Consequently, to reach a certain water level in Ha Noi, more water release from reservoirs is required. After identifying these main issues, the NSGA II model was applied to optimise the operational strategies of the Hoa Binh reservoir. This optimization framework encompasses multiple objectives: focusing on the dry season the minimization of water release deviation from the demand and the maximization of hydropower generation . Concurrently, the optimisation framework must satisfy operational constraints, including storage continuity, minimum reservoir elevation, and the daily average minimum release specified by governmental regulations. Key input parameters for the optimisation model include daily reservoir inflow, water demand within the Red River Delta, and the reservoir's storage-water level equation. The decision variables are the daily outflow from Hoa Binh. The optimisation model was run for a 22-year period, segmented into two sub-periods (2001-2011 and 2012–2022), mirroring the commencement of Son La reservoir operations. Results show notable increments in energy production and reservoir water levels, while simultaneously satisfying downstream demands in the RRD. Flood control objectives were not optimized within this study; however, safety measures during flood seasons were upheld, ensuring the normal water levels of the reservoir and the third alarm water level threshold downstream in Ha Noi. On average the optimised regulation ensured a 10% increase of energy generation over the 2001-2011 period and a 2% increase in the 2012-2022, ensuring always the consumptive water demand and the dam safety. Finally, the study presents a comparative graph depicting the median trajectory of optimal and historical reservoir water levels across the aforementioned periods. In the present contexts, operators of the Hoa Binh reservoir face end-of-dry-season water shortages, sometimes causing the reservoir to operate at its dead level, risking electrical network stability and causing load shedding in the northern region. Derived from optimal water levels, the maximum reservoir water level at the end of the dry season can be increased. Moreover, operators may utilize these graphical representations as support while making decisions regarding reservoir operation.File | Dimensione | Formato | |
---|---|---|---|
Nguyen tesi corretto.pdf
embargo fino al 24/01/2026
Dimensione
7.13 MB
Formato
Adobe PDF
|
7.13 MB | Adobe PDF |
I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/20.500.14242/190472
URN:NBN:IT:UNIBS-190472