Quantitative monitoring of temperature changes, ∆T, can be achieved by MR thermometry (MRT) based on proton resonance frequency (PRF) [1]. Additionally, MR velocimetry (MRV) is a well-established method to quantitatively measure velocities. Measurements of temperature and velocity fields are also of great interest to the engineering community, e.g. for the optimization of heat transfer performance. Conventional methods such as a combination of particle image velocimetry and particle image thermometry are not capable to provide 3D data of flow and temperature in macroscopic setups with one imaging modality and a single measurement setup within a similar amount of time as MRI does. The combination of MRT and MRV revealed promising results [2,3]; however, the capability of MRV and MRT has not yet been widely investigated in fluid flow. This work aimed to investigate 3D velocity and temperature fields in a countercurrent double pipe heat exchanger.The heat exchanger consists of two separate closed flow circuits, a copper pipe (outer diameter=15mm) inside a Plexiglass® pipe (inner diameter=50mm, length of both pipes=320mm) oriented parallel to the external magnetic field (Figure1). The inner hot flow circuit (flow2) transferred heat to the surrounding non-heated flow circuit (flow1=5l/min). The flow2 circuit was connected to a circulation heater/cooler (Julabo FC1200T and SE, class III, Seelbach, Germany) providing fluid temperatures of 21°C or 50°C with a constant flow rate. A detailed description of the flow circuits and the model can be found in [4]. Velocity and temperature fields were measured in flow1 when the temperature of flow2 was increased. 3D data was acquired at a 3T MR system (Prisma, Siemens, Erlangen, Germany) using a phase contrast GRE sequence for MRV and a velocity compensated GRE sequence for PRF shift thermometry (imaging parameters listed in Table 1). PRF data were acquired with 1.) both flow circuits at 21°C (heat off) and 2.) flow1=21°C and flow2=50°C (heat on). The phase images of each acquisition were subtracted and converted to $$\Delta T = \frac{\Delta \Phi}{2\pi \cdot f \cdot \alpha \cdot TE}$$ (f=123.2MHz, α_1%CuSO4=0.0097ppm/K). Reference phantoms filled with 5%Hydroxyethylcellulose+1%CuSO₄+dist. H₂O surrounded the flow model to allow correction for field drift. For MRV a flow off scan was acquired to correct for eddy currents.Figure 2 shows a superposition of temperature maps (TMs) and velocity vector fields at different cross-sectional views (distance 3cm) within the double pipe heat exchanger. Heated fluid raises to the top of the outer pipe depicted as a localized thin structure (plume). At the top, the heated fluid moves outwards forming side lobes which are warmer than the surroundings. The velocity field clearly represents the natural convection flow due to the temperature increase within the fluid. In Figure 3 two streamlines starting at locations X and O are superimposed on the color-encoded TMs. The distance between bullets represents the velocity magnitude. In areas with no temperature change only forced convection is present (represented by equally spaced bullets). In the heated area, a particle located closely at the plume follows a 3D trajectory into the lobe as it is traveling along Z. At the beginning, temperature and speed increase (wider spaced bullets), in the side lobe the speed is slower (bullets are packaged). Furthermore, it remains at a similar location when the forced convection exceeds the natural convection.A combination of forced convection (external pump providing laminar flow) and free convection (heating) as investigated in this work using MRI can contribute valuable new insights into heat transfer processes. The coupling between the successfully measured velocity and temperature fields can be clearly observed. The temperature increase causes a decrease of density of the fluid. Thus, the heated fluid rises (natural convection) in agreement with the measured velocity. The observation of the plume due to natural convection has been described in literature previously [5,6]. A detailed discussion on possible sources of errors in the TMs and a comparison to thermocouple measurements is given in [4]. The combination of MRV and MRT is a very promising technique to examine thermo-fluidic questions to gain a better understanding of devices such as heat exchangers. Currently, no other modality provides a similar performance and speed. Additionally, collaborations between fluid mechanical engineers and the MR community offer great potentials to investigate heat transport processes via blood flow for instance which occur during thermal treatment of tumors.