Jonas Lundsted Poulsen, Senior Specialist at the Danish Technological Institute (DTI), speaking at 14th IEA Heat Pump Conference, Chicago, Illinois
Jonas Lundsted Poulsen, Senior Specialist at the Danish Technological Institute (DTI), speaking at 14th IEA Heat Pump Conference, Chicago, Illinois

Danish Researchers Develop Model for Optimizing Large Ammonia Heat Pump

The model can improve the resilience and performance of large-scale heat pumps under variable operating conditions.

Danish researchers have developed a dynamic model for a 1MW (284TR)-scale single-stage ammonia heat pump system to optimize its design for fast-changing external conditions.

Jonas Lundsted Poulsen, Senior Specialist at the Danish Technological Institute (DTI), presented results of the heat pump study – “Analysis of Large-scale Ammonia Heat Pumps in Transient Operating Conditions” – at the 14th IEA Heat Pump Conference, held May 15–18 in Chicago, Illinois.

In addition to DTI, researchers from the Technical University of Denmark and Johnson Controls Denmark worked on the study.

Large-scale heat pumps are being identified as key components in the crucial move towards the decarbonization of district heating and the industrial heat supply. However, when these heat pumps are integrated with ambient heat sources or industrial processes, they may face challenges operating under variable boundary conditions such as rapid changes in the sink or source temperature or flow rate, including variations in production processes and sudden changes of weather conditions.

The dynamic operating conditions may result in unwanted dynamic effects, risking suboptimal performance and component damage, the study noted.

Addressing this issue, the researchers created a dynamic model to develop and validate a 1MW-scale single-stage Johnson Controls ammonia heat pump system, equipped with a screw compressor and a flooded evaporator. The goal of the model was to optimize both the design and control of similar large-scale systems, thereby ensuring optimal performance under rapidly changing conditions.

The researchers employed the Dymola (Dynamic Modeling Laboratory) tool.

“All the results from the simulation indicate that the dynamic model, systematically developed by modeling each individual part of the heat pump system against its design data, is a robust tool,”  said Poulsen.

“Furthermore, the finetuning of the controllers to match the dynamic response of the measured data underscores the model’s potential in developing holistic control strategies and design guidelines,” he added. “It signifies a substantial leap in our ability to improve the resilience and performance of large-scale heat pumps under variable operating conditions.”

“This dynamic model stands as a testament to the potential of integrating technology and complex scientific modeling,” Poulsen continued. “By optimizing both the design and control of large-scale heat pump systems, this solution made possible a robust blueprint for future improvements in the field of heat supply and distribution systems.”

However, model validation with measured or empirical data is critical for accuracy checks, overfitting prevention, model improvement, confidence in predictions, avoiding bias and understanding uncertainty, he noted.

Model description

The heat pump, illustrated below, comprised a component screw compressor (C1), oil cooler (HX6), condenser (HX3), liquid separators (R1 and R2), sub-cooler (HX4), economizer (HX5), evaporator (HX1), superheater (HX2) and expansion valve (V4).

The experimental setup included multiple pressure, temperature, flow-rate and level sensors. Validation was carried out across three different cases: case 0: stable operating conditions (steady state scenario); case 1: receiver level changes (a single set-point change); and case 2: ramping up of the compressor (multiple simultaneous changes).

A key aspect of the model is the detailed representation of the heat exchanger. In this model, the plate heat exchanger was divided into 10 cells. The waterside was fitted with a constant heat transfer coefficient, while the ammonia side used a heat transfer model with a fitted correction factor. Comparisons were made at various design conditions, including heat transfer rate, water outlet temperature, and pressure drop for ammonia.

The screw compressors aspect of the model, which includes the economizer part and oil cooling, is based on a finite volume method and involves the conservation of mass and energy. The mass flow was defined based on the pressure drop, following the Saint Venant and Wantzel approach.

The design data included a combination of evaporating temperatures (20 to 45°C/45 to 113°F) and condensing temperatures (65 to 95°C /149 to 203°F) with a 5°C (41°F) resolution at 3,000 rpm.

Layout of heat pump system, including input parameters(written in green)
Layout of heat pump system, including input parameters(written in green)

Simulation outcome

The dynamic response for case 0, featuring some of the system’s key parameters, revealed no significant differences between the model and the measured data. The parameters such as discharge pressures, evaporation temperature and the oil inlet temperature, when viewed as a function of time, all displayed close alignment.

In case 1, despite the set point change in the receiver, other input parameters such as compressor speed and secondary flows on the sink and source side remained relatively stable, demonstrating the system’s stability under the modified conditions.

Likewise, case 2 showed that the model and measured data correspond for changes in temperature on the sink and source side as well as changes on the compressor side.

Related Products

Recent News

Related Partner

Related Products

Newsletter

Go to top