Multi-chiller System Research Articles

A model predictive control for a multi-chiller system in data center considering whole system energy conservation

To meet the thermal environment requirements for the continuous and stable operation of information technology (IT) equipment in data rooms, the cooling system needs to operate all day, which has become the second largest energy-consuming system in data centers. The absence of an effective operational control strategy in the data center cooling process results in high energy consumption and poor power usage effectiveness (PUE) of the data center. This study proposes a model predictive control (MPC) strategy considering the constraint of whole system energy conservation for a multi-chiller system in a data center. This strategy aims to maintain the data center server room temperature stable in the IT equipment working environment, and at the same time, the energy consumption of the cooling system is significantly reduced, thus the PUE of the data center is lowered. The long and short-term memory (LSTM) neural network prediction models for data center cooling load and server room temperature were constructed using the mechanism-coupled data law method. The MPC algorithmic structure coupled with cooling load prediction model and server room temperature prediction model. It takes the balance between supply and demand of cooling capacity as a constraint. The cost function of multi-chiller system control considers the temperature-controlling error in the server room and the power consumption of the chilled-water system. A particle swarm optimization (PSO) algorithm is used to solve the optimal configuration strategy of the chilled water flow rate and the water supply temperature of each chiller, which realizes the dynamic control of the chilled-water system. A dynamic regulation model was developed for a data center in North China. TRNSYS was utilized for verification based on the actual measured data and relevant meteorological parameters of the case building. Compared with the operation results of the data center cooling system with proportional integral derivative (PID) control and fuzzy control. The results show that under MPC strategy, the stability of the server room temperature has improved by 27.77% compared to PID control and by 18.08% compared to fuzzy control, demonstrating the superiority of MPC in temperature stability control in the server room. In terms of chilled-water system energy consumption, compared to the PID control and fuzzy control strategies, the MPC strategy has achieved energy savings of 11.81% and 7.58%, respectively, showing significant energy-saving effects.

Particle Swarm Optimization for multi-chiller system: Capacity configuration and load distribution

This study applies Particle Swarm Optimization (PSO) to enhance the energy efficiency of a multi-chiller system in a large office building, with a focus on optimizing capacity configuration and load distribution. Given the rising demand for sustainable energy solutions in buildings, where HVAC systems, particularly chillers, account for a significant portion of energy consumption, this research aims to reduce energy use and improve system performance. EnergyPlus simulations, based on Baltimore weather data and a reference large office model, were conducted to build a baseline dataset. This dataset was then used in Python to calculate energy consumption and optimize load distribution and capacity configuration. PSO was applied in three stages: optimizing capacity configuration using five built-in EnergyPlus algorithms, optimizing load distribution, and simultaneously optimizing both. The results showed that optimizing capacity configuration alone improved performance, even using traditional load distribution methods. Load distribution optimization outperformed other algorithms and converged at a 0.7 Part Load Ratio (PLR) for staging. The integrated PSO application achieved an 11.3 % reduction in energy consumption, a 21.5 % improvement in Coefficient of Performance (COP), and a 12.8 % increase in the Seasonal Energy Efficiency Ratio (SEER) by precise operation of 15 different combinations throughout the entire load range. These results demonstrate the potential of PSO to significantly enhance the efficiency of multi-chiller systems, providing a novel approach to optimizing both capacity configuration and load distribution. This research contributes to a robust framework for improving energy performance in building systems and offers valuable insights for future sustainable energy solutions.

An optimal configuration strategy of multi-chiller system based on overlap analysis and improved GA

For design of the multi-chiller system, the configuration can be optimized that the system is high efficiency when it is operated under frequent load and operation condition in one year. Base on this idea, the paper presents a novel strategy of optimal configuration of multi-chiller system. The high hour-density range of annual load distribution and the high-efficiency region of combination of chillers are presented respectively, and their obtaining methods are developed. The overlap index is defined to evaluate the overlap between the high-efficiency region and high hour-density range, and it is maximized for the proposed optimal configuration strategy. GA is improved and employed for optimization. The proposed optimal configuration strategy is validated in a real multi-chiller system. The results show that the optimal configuration configured by the proposed strategy can improve annual Energy Efficiency Ratio (EER) about 5 % comparing with the original configuration of the real system. The configurations using equivalent strategy and typical non-equivalent strategy are compared, and the results show that both of hourly COP and annual EER of the configuration of the proposed strategy are better than them. 50 random configurations enumerated from chiller candidates are also compared. The results show that the overlap index is strongly correlated to annual EER, and it can be taken as optimization object for optimal configuration. Improved GA is compared with original GA, and the results show that of computation time cost can be reduced more than 60 % by use of Improved GA. The validation results show that the proposed strategy can obtain the optimal configuration of multi-chiller system directly from chiller candidates.

A Collaborative Neurodynamic Optimization Approach to Distributed Chiller Loading.

In this article, we present a collaborative neurodynamic optimization approach to distributed chiller loading in the presence of nonconvex power consumption functions and binary variables associated with cardinality constraints. We formulate a cardinality-constrained distributed optimization problem with nonconvex objective functions and discrete feasible regions, based on an augmented Lagrangian function. To overcome the difficulty caused by the nonconvexity in the formulated distributed optimization problem, we develop a collaborative neurodynamic optimization method based on multiple coupled recurrent neural networks reinitialized repeatedly using a meta-heuristic rule. We elaborate on experimental results based on two multi-chiller systems with the parameters from the chiller manufacturers to demonstrate the efficacy of the proposed approach in comparison to several baselines.

Optimal chiller loading considering the energy loss associated with the switching of chillers based on a novel hybrid fuzzy-metaheuristic method

In existing research, the energy loss associated with the switching of chillers has not been taken into account when addressing the traditional optimal chiller loading (tra-OCL) problem. Chillers incur a certain amount of energy loss during the switching process in practical operations. Hence, it is crucial to investigate the optimal chiller loading problem considering the energy loss associated with the switching of chillers (loss-OCL). In this study, we propose a novel parameter adaptation imperialist competitive algorithm based on parallel fuzzy inference systems (PA-ICA-PF) to solve the loss-OCL problem. Three operators—velocity adaptation, velocity divergence, and mutation-crossover—are introduced to enhance assimilation, revolution, and imperialist competition stages. Two fuzzy inference systems are utilized to concurrently guide searches in the solution space, ensuring a balance between diversification and intensification. The proposed method was experimentally verified in an actual multi-chiller system. Experimental results indicate that the total energy consumption of the system, when addressing the loss-OCL, is significantly lower than the actual value but slightly higher than the tra-OCL. This suggests that the proposed method is capable of obtaining the optimal partial load ratio of chillers while accounting for energy losses. To further validate the effectiveness of the proposed algorithm, we compared it with existing algorithms. Experimental results demonstrate that the proposed PA-ICA-PF method exhibits the best performance.

Optimizing energy efficiency in multi-chiller systems: A comprehensive Modelica-based approach

To achieve sustainable development of buildings, this paper investigates the energy-saving potential of multi-chiller systems by examining the interactions between sequencing optimization, load optimization, and parameter optimization in systems with varying capacities. A Modelica-based simulation platform, adaptable to various programming languages optimization engines, was developed for this purpose. The study introduces a sequencing optimization threshold function that considers both wet-bulb temperature and cooling load. Additionally, a two-stage optimization strategy that prioritizes parameter adjustments before sequencing is proposed. Case studies demonstrate that this threshold function and reversed optimization order can further enhance overall energy savings. Focusing on a system with three large chillers and one small chiller, the study reveals the energy benefits of activating the small chiller under specific conditions, improving overall system efficiency. The proposed optimization achieved a cumulative energy saving of 650,463 kWh over the summer, with sequencing contributing 5.64 %. Frequency conversion control of the water pump is identified as a significant energy-saving factor. This research provides a comprehensive approach to optimizing diverse chiller systems, highlighting substantial potential for energy conservation.

Energy characteristics of multi-chiller load distribution algorithms in a large office building

This study evaluates the energy efficiency of multi-chiller systems in large office buildings, focusing on their optimization across various climate zones as defined by ASHRAE. Using EnergyPlus for simulations, the research examines five different load distribution algorithms in multi-chiller systems that range from one to ten chillers, aiming to understand their effectiveness in 15 distinct climate zones. The primary objectives of the study include identifying the energy efficiency of multi-chiller systems in each climate zone, determining the appropriate number of chillers for each zone, and evaluating the performance of the load distribution algorithms. Based on the U.S. Department of Energy’s commercial building model, the results suggest that multi-chiller systems can significantly reduce cooling energy consumption in various climates. Among the algorithms evaluated, the Sequential Uniform Part Load Ratio (SUPLR) algorithm demonstrates notable efficiency, especially in the 4A climate zone (Baltimore), where it achieves substantial energy savings. Applying the SUPLR algorithm in a multi-chiller setup with four chillers in this zone leads to an estimated 24.5 % reduction in energy usage, equivalent to 183 MW annually. The research indicates that a range of 3 to 5 chillers is typically optimal for most climate zones. In-depth analysis in the 4A climate zone highlights the importance of minimizing operation hours at low Part Load Ratios (PLR) to ensure that chillers operate at a high Coefficient of Performance (COP). This strategy underscores the potential of well-designed multi-chiller systems to reduce cooling energy demand, particularly in climates with transitional seasons. This study provides an overview of the energy-saving potential of multi-chiller systems, applicable across a variety of climatic scenarios.

Attention mechanism-aided model ensemble method of chiller energy consumption prediction

The chiller energy consumption prediction plays an important role in reducing the building energy consumption. Continuously improving the performance of the prediction model is the key to ensuring the accuracy of the chiller energy consumption prediction. As a result, a new improved ensemble model of the chiller energy consumption prediction is proposed, based on the idea of algorithmic ensemble aided by an attention mechanism (AM). To guarantee that the proposed model can see information from many spatial and structural viewpoints, the improved ensemble model incorporates the benefits of several base prediction methods. The suggested approach chooses crucial features to simplify the model input by utilizing AM, thereby reducing the complexity of the samples. The improved ensemble model was experimentally verified in an actual multi-chiller system, and the results showed that the established model could obtain good prediction results. Finally, the proposed model was compared with existing, well-known prediction models. To evaluate the accuracy of these models, the mean absolute error (MAE), mean absolute percentage error (MAPE), mean square error (MSE), variance accounted for (VAF), and coefficient of determination (R2) were used as the evaluation indices. With an MAE of 2.858, an MAPE of 1.2755 %, an MSE of 20.1878, a VAF of 93.6058, and an R2 of 0.9923, the findings demonstrate that the improved ensemble model has a greater prediction accuracy than the other well-known prediction methods. The suggested model enhances the variety of empirical model algorithm libraries.

A bilevel fast-convergent optimizer via high-fidelity convex models: Application on optimal operation of all-parallel heterogeneous chiller-pump systems

To cope with buildings’ growing and diverse cooling demand, heterogeneous multi-chiller systems with non-dedicated pumps are commonly applied in large-scale centralized chiller plants. All-parallel heterogeneous systems offer high operation flexibility but pose challenges to optimal operation planning. For the deficiencies in convexity tractability, the classical neural networks (NNs) achieved great success in system identification yet were restricted in model-based optimization. In this study, an optimization-oriented convex modeling approach based on input convex neural networks (ICNN) to identify energy consumption models for energy units of chiller-pump systems was presented, which provides convex input–output mappings while leveraging the high-fidelity capability of NNs. Using the convex models, the energy minimization issue subjected to multiple constraints was formulated, the optimality of which was mathematically proven. A bilevel deterministic optimizer was developed to determine the global optimal. Comprehensive data experiments were conducted using practical and simulated operation data to evaluate the modeling and optimization performances of the proposed methods compared with conventional candidates. Numerical results suggest that the ICNN-based convex models exhibit superior modeling performances than physics-based models. A better overall energy saving ratio of 8.86 % and faster convergence time of 3.53 s for average achieved by the proposed bilevel optimizer compared with rule-based and meta-heuristic optimizers under identical external conditions.

A validated model for variable-speed multi-chillers to assess different operation modes

Environmental protection drives the replacement of environmentally harmful refrigerants with potentially flammable alternatives, necessitating efficient cooling strategies for varying loads. Legislation limits flammable refrigerant charge, leading to the adoption of multi-chiller systems with smaller charges and therefore also smaller loads per chiller. However, operating such systems requires complex strategies, addressable through fast and reliable models. This study presents a semi-empirical model predicting the behaviour of variable-speed multi-chiller systems. The model focuses on predicting cooling capacity and coefficient of performance (COP) for cooling, with input parameters of evaporation and condensation temperatures, and compressor speed. Experimental validation of the parameterized model shows COP predictions within 3 % for a two-module system using R-290 (propane) under different conditions and compressor speeds. The model assesses various operation strategies, including the conventional mode with fewer chiller modules running at high speed, compared to modes where all chillers run but at reduced speeds. Comparing COP values in these modes reveals energy-saving potentials of up to 14.57 % for constant condensation temperatures. Moreover, strategies involving condensation temperature reduction, or a combination with compressor speed reduction are analysed for energy-saving potentials. Reducing the condensation temperature by 3 K results in savings of up to 13.25 % and combined with the compressor speed reduction savings of up to 23.54 % are feasible. These findings emphasize the importance of considering compressor speed when operating multi-chiller systems to maximize energy savings. The presented model effectively evaluates various strategies and implements variable-load control in multi-chiller systems, serving as a valuable tool for optimising energy efficiency and promoting environmentally friendly cooling practices.

A novel energy consumption prediction method for chillers based on an improved support vector machine

The energy consumption prediction of the chiller is an important means to reduce the energy consumption of buildings. Therefore, a novel energy consumption prediction model for chillers based on an improved support vector machine (ICA-DE-SVM) is proposed. The imperialist competitive algorithm (ICA) is used to optimize the penalty coefficient and kernel function width of SVM, greatly improving the generalization ability and prediction accuracy of the SVM model. The assimilation process is very important in ICA. Colonies of empires move randomly toward imperialists during the assimilation process in ICA, which decreases population diversity and can lead to premature convergence. Therefore, to create more new locations for colonies and increase population diversity, the idea of differential mutation proposed by differential evolution (DE) was applied to ICA. The established model was experimentally verified in an actual multi-chiller system in a building, and the results showed that the ICA-DE-SVM model could obtain good prediction results. Finally, the proposed model was compared with SVM model, PSO-SVM model, GA-SVM model, WOA-SVM model, and ICA-SVM model. With an MAPE of 0.6%, an MSE of 2.3, and an R2 of 0.9998, the findings demonstrate that the ICA-DE-SVM model has a greater prediction accuracy than the other models.

Optimal chiller loading solution for energy conservation using Barnacles Mating Optimizer algorithm

This paper proposes an application of evolutionary optimization algorithm, Barnacles Mating Optimizer (BMO) to solve the optimal chiller loading (OCL) problem for minimization of the power consumption in the multi chiller system. BMO mimics the mating behaviour of barnacles in nature. To generate new off-springs, the concept of Hardy–Weinberg principle is adopted for exploitation process and the sperm-cast process is treated as exploration behaviour of BMO. OCL on the other hand is the problem of finding the combination of partial load ratio of each chiller to obtain the minimum energy consumption in air-conditioning system. To demonstrate the effectiveness of proposed BMO, it is tested on three different chiller systems. The obtained results from BMO are compared with other well-known optimization algorithms in the literature. The obtained comparison results indicate that proposed BMO is effective to reach minimum energy consumption for solving the OCL problem.

Discussion and Commentary on “Robust scheduling of multi-chiller system with chilled-water storage under hourly electricity pricing” in “Energy and Buildings” 218 (2020) 110058

A discussion of the “Robust scheduling of multi-chiller system with chilled-water storage under hourly electricity pricing” by Milad Sadat-Mohammadi et al. “Energy and Buildings” 218 (2020) 110058 is presented in this brief communication. A robust strategy for scheduling multi-chiller systems with chilled-water storage is provided in the mentioned work. The authors have discussed and presented the results for the optimal chiller loading considering the cooling load uncertainty in risk-neutral, risk-taker, and risk-averse strategies based on robust optimization method. RO takes a worst-case strategy to avoid ambiguity in probability distributions. Their proposed model (i.e., RO method), on the other hand, never can model risk-taking strategy in modeling the uncertainty of parameters that is cooling load in this study. Info-gap decision theory (IGDT) has a different formulation, which is an approach that deals with risk-neutral, risk-taker, and risk-averse strategies, which is formulated in this commentary. A clarification of the problem formulation and proper modeling of the problem are offered in this regard to attain the correct modeling of uncertain cooling load in energy systems. Also, based on the proposed formulation of the robust optimization approach, the assumption and simulation results achieved by their suggested model are analyzed and discussed.

Adaptive data-driven optimization of chiller loading with domain knowledge

In order to ensure energy saving and reliable operation for the actual multi-chiller systems, this paper proposes a novel chiller loading strategy that combines the operation data and domain knowledge. The operation data are used to establish the near-optimal performance map which can guide the optimization of chiller loading. The domain knowledge is used to formulate rules to keep the operation reliability of multi-chiller systems. Then a comprehensive method is developed to determine the near-optimal load allocation by considering the map, rules, and the current operation condition. Both the performance prediction study and on-site study are conducted to validate the proposed control strategy. In the performance prediction study, the strategy is compared with other two reference strategies during March to November. The results show that the proposed strategy has the highest energy efficiency among the three strategies. In field study, the proposed strategy is compared with the original strategy on the actual multi-chiller system. The on-site results show that the total energy saving of the proposed strategy is 5.3% compared with that of the original strategy. Then the proposed strategy can adapt to different operation conditions, ensuring energy savings while operating within the plant's constraints.

Robust scheduling of multi-chiller system with chilled-water storage under hourly electricity pricing

Multi-chiller systems with thermal storage units are widely used to supply the cooling load of residential and commercial buildings. Due to the significant energy consumption, economic dispatch has been a crucial issue in multi-chiller systems. This paper presents a robust optimization approach for the day-ahead scheduling of a multi-chiller system with a chilled-water storage unit. The objective function aims to minimize the overall electricity cost of the system under hourly electricity prices, considering the uncertainty associated with the cooling load. Two case studies with different cooling load profiles are considered in evaluating the effect of the storage unit in increasing the robustness of total cost against the changes in the uncertain parameter. The multi-chiller system with the storage unit is modeled as non-linear programming solved with the General Algebraic Modeling System (GAMS), and the optimal partial load ratio of chillers and charge/discharge status of the storage unit are determined. In both case studies, the performance of the system with/without the storage unit is analyzed. The results demonstrate that the integration of the chilled-water storage unit increases the robustness of total cost against the variation of cooling load, as well as reducing the overall electricity cost.

Optimal consumption modeling of multi–chiller system using a robust optimization algorithm with considering the measurement, control and threshold uncertainties

Recently, one of the best options for reducing energy consumption is optimizing energy management of multi-chiller systems in large scale and industrial plants. Different control methods have been implemented to achieve the best performance of the chiller plant but these methods not supply the cooling adequacy for indoor thermal comfort. In order to solve this problem, an optimal chiller loading is controlled in this paper, by applying the conventional and Lagrange methods. Since multi-chiller systems have some uncertainty elements, achieving a precise result is impossible. Three important parameters that cause uncertainty are measurement, threshold, and cooling demand. Using some various load characteristics, a novel hybrid sequential control method is proposed. With employ the historical data, in order to prove the superiority of the proposed hybrid control, its results have been compared with other control methods in the presence of different uncertainties. In the following, to achieve more accurate and optimal results, the controller parameters using a robust method have been improved. Since chiller plant optimization is a mixed-integer nonlinear problem. Then, a new algorithm, known as the camel traveling behavior algorithm, is used to obtain the best results. By analyzing the results, it is seen that the suggested method has the ability to compute in the inexact location.

Determinants of low energy performance in a multi-chiller system serving an educational premise

Abstract This study analyses an association between operating variables and low energy performance in terms of the coefficient of performance (COP) of a chiller system in daily operation. Year-round operating data were collected at 15-min intervals from a trend logging device for a system with five chillers of two different capacities. Two-step cluster analysis was applied to group operating data with high and low COP ratings. The odds ratio analysis ascertained that the key deficient control was operating extra chillers with one at particularly light loads. The optimal trade-off between the chiller power and condenser water pump power was not achieved due to the surplus operation of condenser water pumps. The low COP rating was also due to an imprecise control in the chilled water supply temperature and the condenser water entering temperature from their fixed set points. The significance of this study is to prioritize COP improvement opportunities by ranking the impact of deficient controls based on the actual configuration and operating practice of any given chiller system.

Energy efficient hourly scheduling of multi-chiller systems using imperialistic competitive algorithm

The hour-ahead optimization of the multi-chiller systems is necessary in order to minimize their power consumption and save energy. In this paper, an imperialistic competitive algorithm based optimal chiller loading problem is solved to find the partial load ratio, cooling flux, and the electricity requirement of the chillers and minimize the energy consumption of the whole system. The proposed approach simulates the socio-political competition using the initial population composed of the colonies and imperialists. The empires with their relative colonies and imperialists compete with each other to collapse the weakest empire and dominate the most powerful one achieving the minimum power consumption. The cooling capacity of the chillers and balance constraint are satisfied in hourly economic dispatch process. The optimization problem is tested on three standard systems. The imperialistic competitive algorithm based optimal chiller loading method finds better operating points for chillers than those obtained by recently published strategies.

Risk and Uncertainty Analysis of Cooling Demand in Multi-Chiller System Using Downside Risk Constraints Method

The optimal performance of a multi-chiller system (MCS) is the crucial factor in managing the expected power consumption (EPC). Also, the uncertainty of cooling demand in the industrial or residential sector plays a crucial role, which should be modeled and managed. So, the stochastic risk-constrained performance of the optimal chiller loading is studied in an uncertain environment in this paper. Scenario-based stochastic programming is applied to the provided case study to model the cooling demand uncertainty (CDU), and the downside risk constraints (DRC) are implemented to model the associated risks. The risk-averse performance of the MCS is compared with the risk-neutral one to show the positive effects of the DRC. The proposed model is implemented under the DICOPT solver in GAMS software. The comparison results show that the expected power consumption of MCS is increased slowly, while the expected risk-in-power consumption (ERIPC) is decreased promptly.

Optimal design of data center cooling systems concerning multi-chiller system configuration and component selection for energy-efficient operation and maximized free-cooling

The large data center electricity consumption is a growing global concern. To be environment-friendly and to enhance energy efficiency in operation, data center cooling systems adopt a variety of advanced cooling and renewable energy technologies such as free cooling. However, these free cooling systems are not optimally designed in field practices, and their energy efficiencies are much lower than that of the ideal case. In this study, optimal designs in water piping, pumps and equipment sequencing control are introduced to maximize the cooling efficiency of free cooling systems. It finds that the use of distribution headers around cooling towers and pumps, the maximization of the number of operating cooling towers, the minimization of the number of operating pumps and the mixed use of large and small single-speed pumps can reduce the system's power consumption by 60% under certain operating conditions. The results also show that the designs can reduce the annual energy consumption by 3–15% depending on the climate conditions.