• The main objective of this project is to analyze the performance of a heat pipe by Deep Convolution Neural Network (DCNN)

• This method reduces the time and cost of analyzing any thermal device by experimental analysis.

A heat pipe is an ingeniously designed device engineered to efficiently transfer heat from a heat source to a heat sink. This transfer is achieved through the evaporation and condensation of a working fluid contained within a sealed system.

The structure of a heat pipe is divided into three distinct sections:

• Evaporator Section: This is the heating zone of the pipe, where the working fluid absorbs heat from the external source, causing it to evaporate.

• Adiabatic Section: Also known as the transport section, this part of the pipe is where the vaporized fluid, now carrying the absorbed heat, travels without significant loss or gain of heat.

• Condenser Section: In this cooling zone, the vapor releases its heat to the external sink and condenses back into a liquid form.

At its core, a heat pipe is essentially a sealed container, often tube-shaped, lined with a wicking material along its inner walls. This wicking material plays a crucial role in the effective functioning of the heat pipe by facilitating the return of the condensed fluid back to the evaporator section, thereby maintaining a continuous cycle of heat transfer

A heat pipe stands out as a remarkably efficient yet simple mechanism capable of transferring substantial amounts of heat across considerable distances. What distinguishes this device is its ability to maintain a nearly constant temperature throughout the process, and it accomplishes this feat without the need for any external power input.

At the heart of a heat pipe's operation are two critical factors: the choice of fluid and the internal operating pressure. These elements are intricately linked to the heat pipe's operating temperature.

• Selection of Fluid: The working fluid inside a heat pipe is chosen based on its ability to undergo phase changes (evaporation and condensation) efficiently at the desired operating temperatures. The fluid must have suitable thermodynamic properties, ensuring it can evaporate when absorbing heat at the source and condense when releasing heat at the sink.

• Operating Pressure: The pressure within the heat pipe is precisely controlled and tailored to complement the characteristics of the chosen fluid. This pressure regulation is crucial as it directly influences the boiling and condensation points of the fluid, thereby dictating the efficiency and effectiveness of the heat transfer process.

By balancing these two aspects – the fluid type and the operating pressure – a heat pipe can function optimally within its intended temperature range, making it an invaluable tool in a myriad of thermal management applications.

The first consideration in the identification of the working fluid is the operating vapor temperature range.

The prime requirements are:

1. Compatibility with wick and wall materials

2. Good thermal stability

3. High latent heat (2.24 x 106 )

4. High thermal conductivity (0.68 W/mK )

5. Low liquid viscosities (2.28 x 10-4 )

6. Low vapour viscosities (1.28 Kg/m3 )

7. High surface tension (5.84 x 10-2 N/m)

Working Fluid used in this study are & hybrid Nano fluid.

• Ag - Silver
• Al2O3 - Aluminium Oxide

Silver Nanoparticles (Ag) and its Preperation

Silver nano-particles are nano particles of silver of between 1 nm and 100 nm in size.

Silver nanoparticles are prepared by the following methods

• Synthesis Methods

  • Wet Chemistry Method

  • Monosaccharide Reduction

  • Sodium Citrate Reduction

  • Reduction via Sodium Borohydride

  • Light – Mediated Growth

  • Silver Mirror Reaction

• Ion Implantation

• Biological Synthesis

SODIUM CITRATE REDUCTION METHOD HAS BEEN ADOPTED FOR THE PREPARATION OF SILVER NANOPARTICLE.

Silver nanofluid solutions were prepared through the reduction of silver nitrate in an aqueous solution using citrate. 10 mg of AgNO₃ was dissolved in 50 ml of double-distilled water, and this solution was brought to a boil. Subsequently, a 1% solution of sodium citrate was added dropwise over a period of 10 minutes. The solution was maintained at a boil for approximately 1 hour. The final product exhibited a greenish-yellow color and had an absorption maximum at 444 nm.

Aluminium Oxide (Alumina) Nanoparticles (Al₂O₃) and Its Preperation

Aqueous solutions of aluminum nitrate and urea were prepared at the desired concentrations. These solutions were then mixed together and heated until the temperature reached 100°C. The reaction produced aluminum hydroxide, a gelatinous precipitate, which was subsequently filtered and heated to temperatures above 250°C.

The final product obtained was alumina nanoparticles in powder form.

The Alumina nanoparticles were then mixed with distilled water and placed in an ultrasonic vibrator to form an alumina nanofluids.

A heat pipe charged with hybrid nanofluid was investigated for its performance by varying parameters such as

• Heat input,

• Fluid Ratio,

• Inclination, and

• Flow rate.

The heat inputs are varied from 40 to 100 W, two fluid ratios such as 80:20, and 60:40, the inclination angle of are 0°, 30°, and 45°, and the flow rates of 200 and 300 ml/min. The two nanofluids used in this investigation are silver nanofluid and Al₂O₃. They were mixed in the ratio of 60:40 and 80:20 and prepared as hybrid nanofluid and used in this investigation.

Deep Convolutional Neural Networks (DCNN) is the most admired deep learning architecture. Researchers are using deep learning to solve many problems since they are computationally efficient. It utilizes effective convolution operation followed by pooling operation and performs sharing of parameters.

WHAT WE DO :

We will be generating a DCNN MODEL using python language which will be predicting the output data.

For this , the computer first require data which is already available to learn how the process is being done and what is actually undergone to get such output .

The code contains two major parts

1. Training

2. Testing

TRAINING:

Training is basically the process, where we provide the available data to the computer through which it learns the process.

TESTING:

Testing is the process where we just give the input values and the computer predicts the output with the help of training it has undergone.

After importing the training and test data, we decided to use the Keras model and set the epochs to 300, as it demonstrated a lower percentage of error compared to others.

In this project work, an analysis of a heat pipe using a hybrid nanofluid was conducted using neural network techniques.

A Deep Convolutional Neural Network (DCNN) was adopted for the analysis. The experimental analysis clearly revealed that while the surface temperature of the heat pipe increases with an increase in input heat, it gradually decreases when moving away from the evaporator section.

Using the experimental data, a DCNN model was developed for the heat pipe utilizing the hybrid nanofluid. This model has been validated against a set of experimental data.

The DCNN model, with a precision of 0.991, accurately predicts the outlet temperature of the heat pipe using a hybrid nanofluid composed of Ag and Al₂O₃. The model takes into account different inclinations of the heat pipe, flow rates, fluid ratios, and heat inputs.

The temperatures predicted by the model have been used to estimate the efficiency of the heat pipe. A coefficient of determination of 0.991 obtained in the DCNN model strongly indicates that the DCNN approach can be effectively extended to predict the performance of heat pipes.

The analysis shows that the efficiency of the heat pipe increases with a decrease in the water inlet temperature, while it decreases with an increase in heat input. Furthermore, it is evident from the analysis that the water outlet temperature increases with an increase in fluid ratio and also rises with an increase in the water inlet temperature.

• Link to the code: https://github.com/sivkhiran/Neural-Network-Analysis-of-Heat-pipe-using-Hybrid-Nano-Fluids/blob/main/7_final_final_mechcode.ipynb