Chilled Water Supply Temperature Research Articles

Predictive control optimization of chiller plants based on deep reinforcement learning

The energy consumption of HVAC systems is enormous, with chiller plants accounting for more than 50% of it. To improve energy efficiency, chiller systems are typically optimized at fixed intervals based on real-time building cooling loads, which usually assumes that the cooling load remains constant within the control interval. However, in many real applications, the current optimal control may be changed when the cooling load suddenly fluctuates. Although this issue can be addressed by shortening the control intervals, more frequent control can damage the equipment or increase energy consumption. To tackle this problem, this paper proposes a model-free predictive control method based on Reinforcement Learning (RL) control and Long Short-Term Memory (LSTM) prediction networks. The LSTM network aims to predict the future cooling load based on historical cooling load data, while the RL is used to make the best control for all chiller plants. By this way, the chilled water supply temperature setpoints can be optimized with the consideration of both instantaneous and predicted cooling loads to minimize energy consumption. In order to validate the effectiveness of the proposed method, an experimental simulation model was constructed based on actual chiller system parameters. Experimental results show that the energy-saving performance of the proposed method is superior to rule-based control (9% improvement), closely comparable to the model predictive control (0.73% difference), and further enhances energy-saving effects compared to non-predictive RL control without shortening the control interval. Additionally, the proposed method just learns by continuously interacting with the environment and does not require any accurate equipment models, making it a viable alternative for buildings when lacking extensive sensors and device information.

Mathematical modelling of passive displacement dual cooling coil (PDDCC) system in a tropical climate

Unlike the conventional displacement ventilation (DV) systems, the Passive Displacement Dual Cooling Coil (PDDCC) system generates airflow in indoor environments based on natural convection and eliminates the necessity of mechanical ventilation to meet the thermal comfort and indoor air quality requirements. Investigations on such PDDCC systems, though, have been largely limited and its performance capacity and fluid flow behaviour are not fully understood. A mathematical model was, hence, developed in the current study to explore the effects of various boundary conditions on the performance of the PDDCC system. Results obtained from the mathematical model were, first, compared and validated against the measurement readings gathered from an experimental investigation conducted on the PDDCC system in a controlled test environment. The validated mathematical model was then applied to a parametric evaluation where the effects of chilled water supply (CHWS) temperature, flow rate, and return air temperature on the PDDCC system performance were examined. The varying effect of the latent and sensible cooling load was observed at the different boundary conditions applied to the PDDCC model. Slightly more than 32% of rise in total cooling load was predicted by reducing the CHWS temperature which established a much more significant contribution in comparison to the CHWS flow rate which generated approximately 5% rise in the total cooling load. Findings were supported by examining the sensible heat ratio (SHR) and the ratio of Latent cooling load: Sensible cooling load which was then was attributed to the temperature difference between the supply and return air temperature. By ensuring the maximum temperature difference, the total cooling load performance of the PDDCC system can be maximised. The contribution of both the CHWS temperature and flowrate was essential to achieve the maximum cooling load, hence, improving the total cooling performance of the PDDCC as a result.

Energy Saving Optimization of Chilled Water System Based on Improved Fruit Fly Optimization Algorithm

Abstract As the main energy consumption part of the central air-conditioning system, the energy saving of the chilled water system is particularly important. In this paper, an improved fruit fly optimization algorithm (IFOA) is used to optimize the operating parameters of the chilled water system to reduce the energy consumption of the chilled water system. In IFOA, the 3-D position coordinate is introduced to expand the search space of the algorithm, the variable-step strategy balances the global search ability and local search ability of the algorithm and helps a single fruit fly jump out of the local optimization through chaos mapping. In order to verify the optimization effect of IFOA on the chilled water system, the energy consumption model of the chilled water system is established. With the lowest total energy consumption of the system as the goal, the operating parameters such as the chilled water supply temperature and the speed ratio of the chilled water pump are optimized. The simulation results show that the energy-saving optimization method of a central air-conditioning chilled water system based on IFOA can make the average energy-saving rate of the system reach 7.9%. Compared with other optimization algorithms, the method has a better energy-saving effect and is more stable.

Energy-saving potential analysis for a 24-hour operating chiller plant using the model-based global optimization method

Various studies have been conducted to develop energy efficient measures for chiller plants. In these studies, much attention has been paid to equipment retrofits and various optimal operation strategies, while some measures may interfere with building normal use, like conflicting with the factory's 24-hour production schedule. How much energy saving potential could be obtained from non-interfere measures remains unknown. Therefore, a full-performance simulation chiller plant model was first built based on the Modelica language and then hierarchically calibrated using measured data. Subsequently, the optimization variables were selected based on three principles proposed in this study. An optimization framework based on GenOpt-Dymola was established for the holistic optimization problem, and a chiller plant located in Shanghai, China was used to investigate the energy-saving potential. The results showed that the average global energy-saving rate of the chiller plant reached 8.5% just by optimizing the setpoints in the original control strategy. The largest contribution to energy savings came from the water pumps, which benefited from lowering the chilled water supply temperature, although this reduced the chiller efficiency. Overall, this work can help engineers select appropriate optimization variables that can be easily deployed and can achieve significant energy savings.

Energy-Efficient Approaches for a Machine Tool Building in a University through Field Measurement and Energy Modelling

The heating, ventilating, and air conditioning (HVAC) system consumes nearly 50% of the building’s energy, especially in Taiwan with a hot and humid climate. Due to the challenges in obtaining energy sources and the negative impacts of excessive energy use on the environment, it is essential to employ an energy-efficient HVAC system. This study conducted the machine tools building in a university. The field measurement was carried out, and the data were used to conduct energy modelling with EnergyPlus (EP) in order to discover some improvements in energy-efficient design. The validation between field measurement and energy modelling was performed, and the error rate was less than 10%. The following strategies were proposed in this study based on several energy-efficient approaches, including room temperature settings, chilled water supply temperature settings, chiller coefficient of performance (COP), shading, and building location. Energy-efficient approaches have been evaluated and could reduce energy consumption annually. The results reveal that the proposed energy-efficient approaches of room temperature settings (3.8%), chilled water supply temperature settings (2.1%), chiller COP (5.9%), using shading (9.1%), and building location (3.0%), respectively, could reduce energy consumption. The analysis discovered that using a well-performing HVAC system and building shading were effective in lowering the amount of energy used, and the energy modelling method could be an effective and satisfactory tool in determining potential energy savings.

On the performance study of a hybrid indirect evaporative cooling and latent-heat thermal energy storage system under commercial operating conditions

In this paper, a hybrid evaporatively-enhanced cooling system is proposed under commercial operating conditions. The indirect evaporative cooler (IEC) is first employed as a pre-cooling equipment to treat ambient air by recovering energy from the indoor exhaust air. The pre-cooled air is then treated to achieve the required thermal condition by employing the latent-heat thermal energy storage (LHTES) containing spherical balls encapsulated with a synthesized phase change material (PCM). Such a novel hybrid process takes full advantage of the IEC’s high cooling efficiency and LHTES’s power peak load shifting capability. Experimental system comprising the IEC integrated with the LHTES is designed and tested under various air conditions. Additionally, mathematical models are developed to simulate transport phenomena behind the system. Employing experimentally-validated models, thermodynamic analyses are conducted to study the system's performance. Results indicate that (1) the proposed hybrid system is able to pre-cool and dehumidify the ambient air in the tropics simultaneously; (2) the ambient air temperature and humidity ratio can be respectively reduced by 6–10 °C and 2–11 g/kg dry air under specific operating ranges; (3) the evaporative cooling technology facilitates the use of a higher chilled water supply temperature which results in an improvement on the chiller efficiency; and (4) a higher water flow rate leads to faster cold energy storage and release rates during the respective charging and discharging periods.

Optimized deep learning-based prediction model for chiller performance prediction

With the development of primary energy consumption and global climate change are encouraged air conditioning chillers to conventional air-conditioning methods. In the last previous couple of decades, several absorption technologies are introduced for absorption cooling systems and solar cooling to achieve higher chiller performance. In this paper, an efficient and optimal chiller performance prediction mechanism, named Spider monkey Bat Algorithm (SMBat)-based Generative adversarial network (SMBat-based GAN), is proposed. It is exploited to predict the chiller performance from time-series data. The proposed SMBat algorithm integrates the Spider Monkey Optimization (SMO) and the Bat algorithm. The GAN classifier is used to predict the chiller performance using the time series data based on the fitness function. In addition, extraction of the feature is achieved using the chilled water supply temperature, chilled water return temperature, condenser water return temperature, chiller water flow, cooling capacity, and power utilization. The features are selected by the wrapper method for selecting the appropriate features for better chiller performance prediction. Then these features are forwarded to the prediction module, where the prediction is done based on the proposed SMBat-based GAN. Finally, the experimental analysis exhibits that the proposed model offered better performance based on the metrics, such as Mean Square Error (MSE), R2-score, and Mean Absolute Error (MAE) when considering four datasets. The developed SMBat-based GAN attained better results for dataset-3 with a minimal MAE of 0.008, a minimal MSE of 0.0001, and a maximal R2-score of 0.999. The R2-score of the proposed method is 5.905%, 0.4%, 0.2%, and 1% higher when compared to the existing approaches, namely, regression model, ANN+ Bat, ANN, and Multi obj GA+NN.

A novel data-driven optimal chiller loading regulator based on backward modeling approach

This paper proposes a new method termed backward modeling approach (BMA) to achieve optimal chiller loading (OCL) for reducing energy consumption in industries running multiple-chillers with different efficiency. The developed OCL regulator (OCLR) based on novel BMA approach is composed of conditional generative network (cGAN) and deep neural network (DNN). Most works on the optimal chiller loading problem are to find out the setting of partial load rate (PLR) for each chiller. However, PLR for each chiller cannot be controlled directly and can only be achieved through setting chilled water supply temperature. A novel feedback control framework was developed to identify the relationship between chilled water supply temperature and the PLR. In light of this, the control instruction for chilled water supply temperature can be set to achieve the desired energy saving. The practical feasibility of loading optimization based on developed OCLR was evaluated by conducting field validation for 1 year in a reputed panel manufacturing factory based in Taiwan running multiple-chiller system. From the experimental results, it is evident that the developed data-driven OCLR based on BMA has very high performance and was able to conserve significant energy in the range of 81.9 MWh to 198 MWh per year.

Globally optimal control of hybrid chilled water plants integrated with small-scale thermal energy storage for energy-efficient operation

The integration of thermal energy storage in chilled water systems is an effective way to improve energy efficiency and is essential for achieving carbon emission reduction. However, the commonly used large-scale thermal energy storage needs significantly larger space, which hinders the wide application of thermal storage in large number of existing buildings. Unlike previous studies, this study integrated a small-scale stratified chilled water storage tank into chilled water plants and proposed a global optimal control strategy to enhance the overall system energy performance. The proposed strategy determines the optimal settings of stratified chilled water storage tank charging/discharging flow rate, chilled water supply temperature, and the number of chillers in order to minimize the daily energy consumption of the chilled water plants under varying load conditions. The stratified chilled water storage tank was modelled as a “virtual chiller” to quantify the energy consumption related to the charging/discharging. Multiple charging/discharging cycles were controlled for optimal chiller loading. The proposed control strategy was evaluated in a simulated complex central chilled water plant. The results show that the proposed optimal control strategy can save the daily energy consumption of the central chilled water plant by 4.35–7.67%, 2.10–3.90%, and 2.30–5.15% in three typical weather conditions.

Experimental study and performance analysis on a closed-cycle rotary dehumidification air conditioning system in deep underground spaces

High temperature and high humidity are challenges threatening human health in deep underground spaces. A rotary dehumidifying air conditioning system is an appealing way to solve this problem with free cooling and free heating from different depths of groundwater. A closed-cycle rotary dehumidification air conditioning system in deep underground spaces is proposed. The process and regeneration channels in this proposed system are merged into an optimal channel by recovering heat and harvesting water from the regenerated air. The effects of working conditions, such as the flow configuration of the chilled water, the chilled water supply temperature, the ambient air humidity and the regeneration temperature, on the energy and exergy performance are experimentally investigated. The recommended the flow configuration of the chilled water is 1:5:4. Under typical operating conditions (32 °C, 14 g kg −1 ), the moisture removal of the system driven by 80 °C hot air, reaches 4.9 kW, the thermodynamic coefficient is 1.1, and the exergy efficiency approaches 50.05%. This closed-cycle system with heat and water recovery improves the energy and exergy performance, which is an effective technology option for being applied in deep underground spaces.

Dynamic Calibration Method of Sensor Drift Fault in HVAC System Based on Bayesian Inference.

Sensor drift fault calibration is essential to maintain the operation of heating, ventilation and air conditioning systems (HVAC) in buildings. Bayesian inference (BI) is becoming more and more popular as a commonly used sensor fault calibration method. However, this method focused mainly on sensor bias fault, and it could be difficult to calibrate drift fault that changes with time. Therefore, a dynamic calibration method for sensor drift fault of HVAC systems based on BI is developed. Taking the drift fault calibration of the chilled water supply temperature sensor of the chiller as an example, the performance of the proposed dynamic calibration method is evaluated. Results show that the combination of the Exponentially Weighted Moving-Average (EWMA) method with high detection accuracy and the proposed BI dynamic calibration method can effectively improve the calibration accuracy of drift fault, and the Mean Absolute Percentage Error (MAPE) value between the calibrated and normal data is less than 5%.

Application of artificial intelligence (AI) control system on chiller plant at MTR station

Chillers account for up to 40% of total station energy consumption in the Hong Kong Mass Transit Railway (MTR) system. As part of green railway initiatives, a site trial was conducted to apply a fully automated AI system to control a chiller plant in order to optimise energy performance in real time while maintaining a level of passenger comfort that suits each station’s environment. Through the predictive power of the AI system, the plant power’s consumption and cooling demands can be forecasted based on actual chiller, station, and weather conditions, all of which vary over time. The optimal operational settings can then be determined using an optimisation model for real-time chiller plant control, including staging, sequencing, chilled water supply temperature set-point, etc. This paper presents the formulation of an AI system using data-driven machine learning models and numerical optimisation, and the comparison of the actual energy performance of the proposed system against rule-based control optimisation in a conventional building management system (BMS) through the site trial. The results revealed the proposed AI system achieves better energy efficiency with annual energy savings of approximately 8.7%.

Development and performance evaluation of a chiller plant predictive operational control strategy by artificial intelligence

Traditionally, chiller plants are controlled and monitored by a predetermined control strategy to ensure appropriate operation based on the designed system configuration. With the use of new technology of variable speed drive (VSD) for compressors, smart control strategies could be leveraged to enhance the system efficiency in lieu of traditional control strategies. For example, using orderly and straightforward switching procedures without considering various factors in switching the units, including the high-efficiency partial load range benefitted from the VSD, the actual performance of the units as a whole and the variable chilled water flow rate, result in the chiller plant not operating at maximum performance and efficiency. To address these issues, a hybrid predictive operational chiller plant control strategy is proposed to optimize the performance of the chiller plant. Artificial intelligence is employed as the data mining algorithm, with big data analysis based on the actual acquired voluminous operation data by fully considering the characteristics of chiller plants without additional installation of large-sized and high-priced equipment. Artificial neural network (ANN) was employed in the control strategy to predict the future outdoor temperature, building cooling load demand and the corresponding power consumption of the chiller plants. At the same time, particle swarm optimization (PSO) was applied to search for the optimized setpoints, e.g., chilled water supply temperature, operating sequence, chilled water flow rate, for the chiller plants. The developed control strategy has been launched in a chiller plant with a cooling capacity of 7,700 kW installed in a hospital in Hong Kong. The system coefficient of performance (COP) and overall energy consumption of the chiller plants were enhanced by about 8.6% and reduced by about 7.9%, respectively, compared with the traditional control strategy. This real-time, continuous, automatic optimization control strategy can determine the most efficient combination of operating parameters of a chiller plant with different control settings. This ensures that the chiller plant operates in its most efficient mode year-round under various operational conditions.

A Temperature & Humidity Setback Demand Response Strategy for HVAC Systems

• A T&H demand response strategy that adjusts the temperature & humidity is proposed • Indoor air, supply air and chilled water temperatures are adjusted in the strategy • The energy saving and the peak shedding of the strategy are up to 39.0% and 37.1% • The T&H strategy increases the energy saving by 22.9% compared with the T strategy • The strategy is more effective under lower sensible load or higher latent load Heating, ventilation and air conditioning (HVAC) system plays a significant role in demand response (DR) to reduce the peak load of the grid. Conventionally, only the sensible load is reduced by increasing the indoor temperature during DR periods and the latent load is rarely concerned. This study proposes a temperature & humidity (T&H) setback strategy for HVAC systems to reduce both the sensible and latent cooling loads by adjusting the supply air temperature and chilled water supply temperature simultaneously. The performance of the T&H setback strategy is analyzed from the perspectives of peak shedding, energy saving and indoor thermal comfort and compared with the common temperature setback strategy. The impact of weather conditions and occupant densities on the performance of T&H strategy is assessed. Results show that the energy saving of the HVAC system under the T&H setback strategy can be up to 39.0% and peak shedding can be increased to 37.1%, which are both higher compared with the temperature setback strategy. The peak shedding and energy saving of the T&H setback strategy is larger when the buildings have lower sensible or higher latent load. The T&H strategy can improve the flexibility of buildings during DR periods effectively.

Chilled water temperature resetting using model-free reinforcement learning: Engineering application

Optimal operation of chillers could be realized by controlling chiller on–off status and resetting chilled water supply/return temperature. The chilled water supply temperature could influence both chiller power and indoor comfort simultaneously. Hence, it is necessary to optimize its set point. By combining the reinforcement learning (RL) technique and expertise knowledge, a hybrid model-free chilled water temperature resetting method for chillers is proposed, to improve the robustness and learning speed of RL control. The proposed method and the comparative expert manual control are applied to control an actual HVAC system in Shanghai. Engineering practice results suggest: (1) the energy-saving performance of the hybrid model-free method is close to that of expert manual control, with limited comfort sacrifice, acceptable robustness, learning speed and control stability; (2) without model establishing and persistent manual overwatch work, the proposed control method could conserve manual labor before and after the intervention of optimized control. Based on the engineering practice, a supplementary simulation study is conducted, which indicates: (1) the performance of the proposed method only takes one cooling season to converge; (2) the proposed hybrid method outperforms both pure RL-based control and pure rule-based control in the first training cooling season, while inferior to model predictive control; (3) the proposed method is an acceptable alternative for systems without sufficient data or certain equipment condition, due to its online self-learning mechanism and low pre-condition requirement on historical data and models.

Control strategy for energy saving of refrigerating station systems in public buildings

In this work, a global energy saving optimal control strategy is proposed to address the high energy consumption of chilled water pumps caused by the small temperature difference between supply and return water, and the low efficiency of the chillers. The developed control strategy consists of two parts: the group control scheme and an optimal controller for variable water temperature/volume control. The group control scheme coordinates each chiller's operating time based on its running time and the chiller load rate. The optimal controller regulates the chilled water pump frequency according to the chilled water temperature difference and the valve opening at important terminal air conditioning units, and calculates the chilled water supply temperature under optimal comprehensive energy consumption of chillers and pumps. The proposed control strategy is easy to be embedded into the existing control systems of refrigerating stations. This scheme is then applied to the air conditioning system of a large public building in Beijing, China. Experimental results show that the proposed strategy can save up to 63% of energy consumption from the chilled water pumps and 6% from the chillers compared to the original control scheme.

A flexible and generic functional mock-up unit based threat injection framework for grid-interactive efficient buildings: A case study in Modelica

Grid-interactive efficient buildings (GEBs) have been considered as an important asset to support the power grid reliability by utilizing the demand flexibility offered by GEBs. GEBs are enabled by advances in sensors and controls, and the communication between building equipment, whole buildings, and the grid. The integration of different building technologies and network-based communication system makes GEBs vulnerable to passive threats such as equipment failure and active threats such as cyber-attacks. Modeling and simulation is an effective way to evaluate the impact of threats on the system performance. This paper proposes a generic and flexible threat injection framework for commonly-used building energy simulators such as EnergyPlus and Modelica to support threat modeling and evaluation. This framework leverages functional mock-up unit (FMU) to develop a general modeling interface for threat injection and simulation. A numerical case study using Modelica as a building energy simulator is conducted to demonstrate the capability of the framework for supporting single/multiple-order threat modeling and simulation of a GEB. Four threats and their combinations are injected on a Modelica-based threat-free building energy and control system, including operating supply fan at its full speed, remotely cycling the chiller on and off, blocking the chiller from receiving the chilled water supply temperature setpoints, and hijacking the global zone air temperature setpoint. Simulation results show that the cyber-attack that leads to short-term signal blocking has small effects on the system operation due to the “self-healing” feature of the heating, ventilation, and air-conditioning (HVAC) interactive control system. The threat that takes control of resetting the global zone air temperature setpoints has the most adverse impact on the system energy use, peak power demand, thermal comfort and the provision of demand flexibility. The combination of four threats have aggregative effects on the system but the effects are less than the additive effects of the individual threat.

Chiller system performance management with market basket analysis

PurposeThis study aims to apply association rule mining (ARM) to uncover specific associations between operating components of a chiller system and improve its coefficient of performance (COP), hence reducing the electricity use of buildings with central air conditioning.Design/methodology/approachFirst, 13 operating variables were identified, comprising measures of temperatures and flow rates of system components and their switching statuses. The variables were grouped into four bins before carrying out ARM. Strong rules were produced to associate the variables and switching statuses with different COP classes.FindingsThe strong rules explain existing constraints on practising chiller sequencing and prioritise variables for optimisation. Based on strong rules for the highest COP class, the optimal operating strategy involves rescheduling chillers and their associated components in pairs during a high load operation. Resetting the chilled water supply temperature is the next best strategy, followed by resetting the condenser water entering temperature, subject to operating constraints.Research limitations/implicationsThis study considers the even frequency method with four bins only. Replication work can be done with other discretisation methods and different numbers of classes to compare potential differences in the bin ranges of the optimised variables.Practical implicationsThe strong rules identified by ARM highlight associations between variables and high or low COPs. This supports the selection of critical variables and the operating status of system components to maximise the COP. Tailor-made optimisation strategies and the associated electricity savings can be further evaluated.Originality/valuePrevious studies applied ARM for chiller fault detection but without considering system performance under the interaction of different components. The novelty of this study is its demonstration of ARM’s intelligence at discovering associations in past operating data. This enables the identification of tailor-made energy management opportunities, which are essential for all engineering systems. ARM is free from the prediction errors of typical regression and black-box models.

Variable importance for chiller system optimization and sustainability

Very few studies have considered modelling the statuses of loading and component combinations (LC) in chiller systems. This study uses a decision tree (DT) and a neural network (NN) to model accurately six output levels of LC in terms of 11 predictor variables. The NN model gave a higher correct prediction of 89.52% than 84.44% in the DT model. A variable dropout analysis from the DT and NN models generalizes the top five significant variables: the statuses of cooling towers and primary chilled water pumps, and the supply temperature, return temperature and flow rate of chilled water. The system coefficient of performance (SCOP) modelled by NN has an improved R 2 of 84.18% versus 74.62% when the significant variables are included as inputs. The SCOP maximization brings three optimal strategies—implementing chiller sequencing, operating components in pairs and optimizing chilled water supply temperature—which reduce CO2 emissions by 99,815–346,987 kg.

Optimal setting parameters of cooling system under different climate zones for data center energy efficiency

The high rates of electricity consumption in data centers become a worldwide concern. Water-side free cooling is an effective solution to reduce the energy consumed by the data center's cooling system. However, the water-side free cooling system is not well planned, and the free cooling utilization rate is low in different climate zone. Moreover, the effects of local climatic conditions on the optimal setting parameters have not been researched. In this study, the optimal setting in chilled water supply temperature, temperature difference, and chilled water return temperature is analyzed to enhance the efficiency of the free cooling system. We observed that the optimal setting could reduce annual energy consumption by 23.5%–60.8% in various climate zones and has substantial energy-saving potential in the subtropical highland climate zone. The results show that the optimal setting parameters are highly affected by the annual accumulative wet-bulb temperature distribution. Furthermore, considering a monthly analysis of these enhanced parameters, we proposed a dual-working condition cooling system for the data center's future construction. This research work guides the design of such cooling systems in various climate zones.