Mechanical engineering Thesis/Dissertation revision

DUAL AXIs SOLAR TRACKING SYSTEM

BY

Date:

ABSTRACT

Photovoltaic projects are preferably carried out in large scale with thousands of the huge panels laid out on flat surfaces facing the sun. However, the sun is constantly in motion across the sky hence there is the need for a sun tracking system included in the photovoltaic panels. These systems are subjected to some loads including the wind which contributes to both the aerodynamic cyclic stresses and mechanical stresses.

In this paper, we will thus look into the designing process of a sun tracking system considering the structural stresses and fatigue stresses which ultimately lead to shortening the systems lifetime. These static pressures are obtained by determined using Finite Element Analysis (FEA) which is carried out using static approach while fatigue is then performed using the stress life-method L. From simulations, it is deduced that stress resistance of the most fragile material should be checked with a safety factor higher than two.

CHAPTER ONE

INTRODUCTION

Whereas renewable energy has been preferred for the sake of environmental conservation, solar power has been the fastest growing means of this energy globally. According to the National Centre for Policy Analysis, grid-connected solar capacity has increased by 60% annually from 2004 to 2009. However, solar system does not contribute much to the total energy produced for example in the US where it contributed to 0.18% of the total energy produced in 2011.

The photovoltaic (PV) solar cell energy has become more competitive with other types of solar energy thereby contributing to 0.10% of the total US energy contribution from 2010 to 2011. Its continued use and growth have been promoted by continued federal tax subsidies one as well as the renewable standard protocol which has resulted in the decrease in its cost. The PV solar power is considered very active and is expected to reach the national grid very fast.

Tracking the sun is the major way of increasing the energy collected from the PV panels. It is preferred to keep the solar panel perpendicular to the sunlight by precisely controlled polar azimuth angle of the panel. The power output of four different trackers was compared to the power output of a fixed panel and the result shown in figure 1. The trackers included a dual axis, North-South, vertical and East-West. The dual axis trackers were found to produce maximum output compared to the other trackers.

Figure 1. A graph of maximum output power for the various PV trackers used.

An STS 444 Dual-Axis solar system is currently being developed by French Development Enterprises(FDE). The solar system apart from tracking the sun in two axes can also be transported to other places.

Carrying out a simulated analysis on the full-scale STS 444 tracker and design a small-scale functional model of the same for demonstration purposes was the project's primary objective. The first part of the report focuses on static and dynamic analysis while the second part concentrates on the design and fabrication of the small-scale functional model.

Existing CPV Tracker: An Overview

The market is flooded with photovoltaic trackers that have different prices depending on the robust nature of each, and also, the material of the frame and type of solar panel will cost a fluctuation of the system. Below are some of the characteristics of the designs used in making these systems.

1. C140 Gearless CPV tracker from Mechatron

The C140 falls under the beetle product line in the Mechatron's product listing categories. It name beetle is based on biomimicry of the actual exoskeleton of a beetle. The exoskeleton is boasted of having a superior strength to weight ratio hence reliable for the case of making a system which will be exposed to strong winds and snow loading during the winter. The biomimicry model of the C140 solar tracker is of an elliptical tube design, and this structure is similar to the wing of a turbine with a simple lattice structure. The lattice structure can take preferred since it easily transfers the structural loading in strong winds of up to 17.9 ms1 velocities.

It comprises of a hydraulic motion mechanism with zero-backlash and also a continuous slip capability that keeps the drive mechanism sufficiently stable even under hurricane conditions. Additionally, this tracking device has got a telemetry system that gives a real-time information output about the machinery operation and as if not enough there is a smart control software that tracks the products output and consequently optimizes the products path to maximize the yield.

2. C7 Tracker by Sunpower

This tracker concentrates solar power up to seven times the insolation hence making it a lowest leveled cost of electricity utility available in the market. It has a single axis tracker positioned horizontally with rows of parabolic mirrors that are reflecting light onto solar cells. It has high efficient solar cells made from silicon powered by the Maxeon cell technology. It can thus reach a tracking of one hundred and fifty degrees, seventy-five degrees on both sides of the axis and maximum power of 12.4kWp and finally being able to withstand winds with a velocity of 44ms1. The system gets to the consumer while preassembled hence making installing first and easy.

3. CPV system from Suncore

The system has got triple-junction solar cells with a twenty-eight percent efficiency. Its module rests on a low profile tilt and dual roll axis tracking system where the motor and modules can easily be accessed from the ground for cleaning and maintenance. Needless to say the CPV has a much higher power output compared to the C& from Sunpower by over one kilowatt of power per system. As for being autonomous it had a provision to be remotely controlled and operated for the need of better output.

4. BSQ-D280/53

The system has a high degree pointing accuracy coupled with a high quality of stiffness and also higher tracking range. It has a large tracking range of fifty-two square meters that consists of an open loop tracking control which achieves a tracking accuracy of 0.1 degrees on average.

Tracker Design optimization

There is need to reflect on the efficiency and workability of the single axis solar tracker and consequently compare it to the dual axis to justify the reason why the latter is preferred. The single axis can withstand a wind loading reaching nine hundred and thirty-four and snow loads of

BACKGROUND

The efficiency of the solar cell and the intensity of light falling on the cell determines the amount of the power output of a solar system. The price of the solar panel is determined by the cell technology as far as fabrication is concerned. The amount of solar power output also depends on it.

Different types of cells available for use today includes crystalline silicon cell which is the most common and widely used. It is 13%- 20% efficient. This means it can produce about 200W power for use per square meter. One way to increase the power output of a solar system is by increasing the amount of light falling on the panel. A PV panel, for example, is most useful when it receives light from a source at the perfectly perpendicular angle. The interdependence of the power output of a solar panel and the angle of incidence of the light is demonstrated in figure 2. The PV must move with the sun to keep the angle of incidence at zero hence the importance of solar tracking.

Figure 2: A graph of measured output power against angle of incidence.

Although solar tracking system can provide more power output as compared to stationary PV panel, it is however costly hence may not always be the best option for a particular application. Using fixed PV panel means placing it facing the sun. For efficiency, the angle is put in a location where it will receive the sunlight at most of the time and also at the optimal angle about the equator. Stationary solar panel orientation is shown in figure 3 below.

Figure 3: Orientation of stationary PV panel.

Due to the tilted axis and elliptical rotation of the earth, a stationary PV panel's output will vary throughout the day thereby resulting in less power output as compared to the tracking system.

A solar tracking PV panel is more efficient with the potential of doubling the stationary panel under ideal conditions regarding power output.

TYPES OF SOLAR TRACKER SYSTEM

Solar tracker systems can be classified into two main groups including single-axis trackers and dual-axis trackers.

Single-axis trackers

They can have either a horizontal or a polar axis and follow the sun's East-West or even north-south movement. The flat types are majorly used in areas near the equator where the sun is scorching at midday hence does not require much adjustment to the vertical axis. The horizontal type is shown in figure 4 below.

Figure 4: A horizontal tracker type.

In very high altitudes where the sun does not get high very, a polar type single axis trackers are used where vertical compensation is not as much as horizontal position.

Dual-axis tracker

It follows the sun's movement irrespective of the axis of rotation. They have both vertical and horizontal axes hence can track the sun's apparent motion in the sky. Dual-axis motion has the advantage of maximizing the total power output by keeping the panel in the direction of the sunlight longer. A standard dual-axis tracker is shown in figure 6 below.

Figure 5: Standard dual-axis tracker.

Tracking system

The main types of solar tracking systems classifications are passive tracking and activity monitoring.

Passive tracking

Passive trackers in most cases use compressed gases to move the tracker. The movement is aided by the creation of gas pressure depending on the angle of sunlight and the gas containers. The tracker is transferred to an equilibrium position as a result of the pressure. The major disadvantage of passive trackers is that they do not use a controller and are slow and vulnerable to external effects such as the wind.

Active tracking

Uses an electromechanical system to position the solar tracker to keep the panel in the perpendicular angle to the sun. It also has a controller which drives the motors and actuators to position the tractor. The trackers may use a solar map, sensors or both. For example, during sunny, whether the sensors would be used to track the sun while during cloudy conditions the information from the solar map would be utilized.

GOAL STATEMENT

The two primary targets and objectives of this project were;

To conduct wind and snow load analyses on the STS 444 to determine the maximum force and stress acting on the system at different panel polar angles.

To design and build a small-scale functional prototype to demonstrate tracking abilities of the STS 444. The sponsor would be using this model to showcase to its future customers.

CHAPTER TWO

LITERATURE REVIEW

Introduction

Solar energy was developed to compensate for the insufficient energy from other sources as well as improve the green technology. Solar tracking is a technology aimed at improving power output for photovoltaic solar panels. A dual-axis solar tracking system 444(STS 444) has been developed and patented by the French Development Enterprises (FDE). The tracker is capable of withstanding wind and snow loads of up to 89.4m/s and 200kg/m2.

About the future expansion of solar energy, various governments and institutions have spearheaded research in the area as the alternative to the natural gas which has been considered as unclean.

Efficiency

Under ideal conditions, the effectiveness of a dual-axis PV solar system is 20-35% higher than fixed PV solar systems which are said to double with tracking solar system. The average electrical power obtained from solar systems depends on the day whether such as clouds and sunlight variations.

Solar cells have a lifespan of approximately 20-25 years. Output power, however, is reported to decline by 0.5% annually. This means that the solar panels are expected to produce about 80% of their rated capacity by 20 years of function.

Reliability

Parallel connection of solar systems is the most efficient connection so that the defective panel may not affect the functionality of the other solar cells and also replacing it is made comfortable without interfering with the whole system. The significant disadvantage associated with solar energy is the low power generation and energy storage ability.

Feasibility

A major concern in solar electricity generation is the time required for the subsidies to remain in place for these systems to be autonomously profitable. The amount of power produced depends on the amount of sunlight received in a particular area, the orientation of the solar array, whether the solar arrays are fixed or track the sun, construction costs and financing options.

Despite this economic factors as well as practical concerns, solar power could become the primary source of energy if the technology catches up to increase the efficiency of the photovoltaic collection and the means by which the energy is harvested.

Chapter III

3.0 SOLAR TRACKER SYSTEM

Area of study

This project is to be conducted in two stages

· Laboratory (simulation and analysis)

· Field (prototyping)

Experimental Model

The solar tracker system 444 falls under second generation dual-axis solar tracker with regards to the development by the FDE. This report includes analyses which were accomplished using preliminary blueprints of the STS 444 which again is provided by the FDE with approximate dimensional estimations. The following sections will thus contain a detailed discussion of the snow load and wind analyses using the STS 444 model. The Data obtained were then compared concerning non-dimensional values to make it easy application in similar systems.

Wind load analyses

Wind load is the most important load on the solar tracker system compared to all the other loads, and it is responsible for very strong forces which again vary in direction and can also cause mechanical damage in the process due to the resonance. Thus this report will analyze only the most important configurations of the solar tracking system while comparing it to a wind velocity of 200mph under a specific direction.

1. The STS experiences forces which were calculated using three approaches namely;

2. Analytical

3. Simulation

4. Experimental method

An estimate of the forces was then obtained using the full-scale tracker system then the obtained forces were then used to obtain the scaling factor for the third approach which is the experimental analysis approach. The boundary conditions set for the experimental approach while analyzing the wind tunnel were then used to conduct a simulation, and the values obtained were compared to those obtained in the previous. Upon data validation, necessary corrections were made on the simulation model to achieve desired results. The data obtained from these three approaches is from the force of wind acting on the system at different angles of tilt on the system.

1. The parameters used for this analysis include

2. the force of wind,

3. area and

4. Force coefficient.

Area

To determine the force coefficient the reference area is first obtained since the wind is often in direct contact with the system, in this analysis the references area is the surfaces of the Photovoltaic panels installed on the system. However, there are certain times that blockage occurred due to a change in the projection area, this error was later accounted for using a correction on wind velocity blockage. Having the fixed reference area will eventually aid in the comparison between drag and lift forces caused by the wind.

Forces

The vertical and horizontal forces acting on the STS 444 accounted for the lift and drag forces due to the horizontal wind blowing on the system. Comparison between the three approaches of analyses invoked the use of the lift and drag forces and upon verification an absolute value is obtained the value is used for the calculation of the von Mises stresses that are acting on the structure based on a few assumptions made such as the exclusion of small components and setting of some boundaries.

Force coefficient

The drag and lift forces are once again used to calculate the lift and drag coefficient. Force coefficient is thus calculated by using the force of the wind, fluid density and the collector area.

(1)

From the above formula then there is a three step procedure for calculating the drag and lift coefficients using all the three approaches, at this point the tilt angle of the system is kept at zero degrees.

I. Analytical Approach

The following equation (equation 2) is used to obtain the Reynolds number of the fluid flow in which V is the velocity of the fluid v is the kinematic viscosity of the fluid and d is the panel length.

(2)

V = 89.4 ms-1,

d = 22.86m

v=1.57 x10-5

The photovoltaic collector was thus calculated to be 1.3 x 10x thus indicating that there was a fully turbulent flow. The following table is showing the drag forces subdivided into parts which are considered to influence the drag forces calculation.

To calculate the lift forces it is required that one solves the boundary layer differential equation as shown in the following equation

FL = (3)

On the other hand the calculated drag force could vary from positive to negative ten Kilo-newton’s (10kN) hence the final drag force could vary by 0.01 on both negative and positive margins.

Table 1: drag coefficient with reference to area and drag force for the STS parts at zero degrees

Drag and lift forces in the procedure are calculated with an increase in the angle of tilt by ten degrees per simulation, from zero to eighty degrees (0 -80 degrees).

Table 2: analytical drag and lift coefficients for different tilt angles on the panel

II. Wind Tunnel Testing

The reproduction process of wind drag on a full scale is costly especially at high speed winds and also time consuming when it comes to building the model. It is imperative to take note that using an appropriate scaling model it is possible to simulate all the forces under research on full scale just like in the wind tunnel. To get accurate results in the three procedures then scaling and an experimental set-up while an estimate in magnitude of the forces is obtained from the analytical approach. Thus using the first equation (equation 1) then the ratio obtained from this formula can then be used in designing the force probe fixtures and also in choosing the appropriate type of force sensors.

a. Scaling Methods

To determine the scaling factor then force ratio, Reynolds number and blockage factor.

a. I. Reynolds number:

It is difficult to keep the Reynolds number for the scaled consistent being that the Reynolds number along both the width and the length are considered turbulent hence the scaled model would be impractical to reproduce it is thus imperative to ensure that the final flow of the for the miniature prototype is turbulent enough.

a. II. Force ratio:

Both the drag and lift forces can be calculated using equation 2 using the analytical and simulated analysis to get both drag and moment experienced at the base of the model.

a. III. Blockage factor

Calculations related to blockage factor include the scaling factor and aspect ratio. The aspect ratio of the model surface area which is perpendicular to the wind flow thus having no tilt angle. With regards to the blockage correction the scaling factor was then obtained using the following formula.

(4)

When we take into consideration the standard dynamics of manufacturing then and also the wind tunnels setup then with a scaling factor of 90 and taking wind speeds of approximately forty five meters per second creates an optimum condition under which the Reynolds was still turbulent at 7.28x 105. Under these conditions the drag and lift reached a maximum of 65 and 70 Newton respectively.

Scale model design and fabrication

i. Reduced scale models

The reduced scale prototype was based mostly on the design model used for the analytical analysis. The first one was that of a rigid prototyped body with no degrees of freedom and with the panel in the horizontal position. The support beams were replaced by a flat sheet through which the board could be fixed by screws. Another support was added at the bottom of the base to secure the force probe. This is shown in the figure below.

Figure 6: End view of prototyped reduced scale model.

The next was made entirely from a combination of aluminum sheet for the panel, acrylic for frame support and rapid prototyping for more complex parts. The main aim of this second prototype was to vary the polar angle of the panel and find the corresponding drag and lift at the base. The prototype is shown below.

Figure 7: Second demonstration of the prototype with modified parts to allow rotation and reinforcement of the model.

ii. Wind tunnel

For the testing purposes, the WPI wind tunnel with cross-sectional dimensions of 0.61m by 0.61m and a maximum wind speed of 60m/s. The frequency of the blades at 62.5 Hz corresponded to a wind speed of approximately 55m/s.

iii. Force probe system

A force probe system was designed to find drag and lift on the prototype simultaneously. A method of moments measured by a force transducer was used to determine the drag force on the prototype. The original design was iterated from a shorter arm to a longer arm about 0.4m since the maximum force was exceeded on the force transducer. With a maximum force of 65 N, the force balance system would produce a maximum voltage value of 1.2V.

The lift force on the prototype was measured directly using a dual-range force transducer and data was collected using a Vernier Lab Pro device. Analytical study at all polar angles yielded lift direction pointing downwards so that the measured force will be compressive if the sensor is placed beneath the force probe.

Figure 8: A setup for the prototype to measure drag and lift in the wind tunnel.

METHODOLOGY

The prototype was fixed onto the lid and inserted into the wind tunnel with the force probe firmly screwed to the base. The position was such that the angle of attack of the wind was directed onto the top of the panel which made the force probe arm to be in contact with both force sensors as shown in the figure below.

Figure 9: A section of the prototype in the wind tunnel.

i. Calibration of the force probe

A combination of 100g masses from 100g to 500g were used I the calibration of the two forces transducers. The force probe system was set up in a configuration of drag and lift measurements as demonstrated in the below diagrams.

The horizontal configuration measured the resulting drag by measuring the moment produced by the drag. The pulley was placed on a stable surface to ensure that the surface was in parallel to the ground. An initial mass reading was recorded with no load while the other reading has been registered with varying loads. The weight and the force probe were attached to a string which was in turn passed over a frictionless pulley so that the mass pulling on the force probe in a direction perpendicular to the force probe stand to simulate the required drag. The corresponding voltage readings were recorded after which no load was retaken to confirm the presence of any hysteresis error.

Figure 10: A setup for the calibration of the force transducer with mass pulling perpendicular to the pivot so as to simulate the drag.

The vertical configuration shown below allowed for lift measurement alone since drag produced a moment and raised created a translational displacement which was assumed to pass through the pivot. By restricting the rotation of the arm while allowing the other one for translation, the lift component could be measured. Since the distances moved were minimal, the rotational motion could be neglected so that translational motion made possible.

Figure 11: A setup for the calibration of the force transducer with mass pulling perpendicular to the pivot so as to simulate the lift.

To calibrate the lift configuration, the masses were made to push parallel to the force probe pivot in a downward direction. The force probe was initially in contact with the force transducer while the axis was found slightly in the middle of the slot to allow for downward displacement. This meant that the initial no-load error reading on the force sensor for the lift was that of the weight of the prototype.

ii. Drag and lift measurements

After calibration, the polar angle of the panel was adjusted, and an initial no-load reading was recorded for both of the force sensors. The wind tunnel was tuned at a frequency of 52. 1 Hz corresponding to approximately 45 m/s based on the wind tunnel calibration. The wind flow, however, took about 1 minute to stabilize. The real drag and lift were found from the calibration curve drawn. By increasing the polar angle by 10 degrees from 0-80 degrees, the corresponding drag and lift were measured for each angle. Hysteresis error was checked by taking a final no-load voltage.

RESULTS AND DISCUSSION

Typical experimental drag and lift values for incremental polar angles were as shown in the table below.

Table 3: Corresponding experimental drag and lift coefficients as the polar angles increase

From the table, it can be deduced that the corresponding experimental drift increased as the angle increased. This relation was much expected because the surface area facing the wind increased hence resulting in increasing the resultant force acting on the panel. The corresponding lift, on the other hand, increased at a decreasing rate of angle to reach almost a constant value.

The setup had got various limitations such as a gap between the base and the bottom floor with spaces where the pivoting probe passed through which resulted to escape of air thereby leading to high forces bending in some of the features components causing the whole prototype system to be slightly tilted. The tilting produced an error in almost all the reading of the respective angles.

Simulated approach

The two steps involved in the simulation were:

i. Conduct simulation based on the reduced prototype with the same set-up and conditions as the wind tunnel experiment. The simulated data was validated with the experimental data.

ii. Conduct simulation on the full-scale STS 444 with real-life conditions and obtain realistic values of the forces acting on the system.

The resultant force acting on the prototype was simulated using ANSYS workbench while the wind flow was modelled using fluent with uni-directional wind speed in Z-direction. On the other hand, the flow type was modelled using a simple standard turbulent model. In both cases, the wind pressure was imported onto the structure, and a static structural analysis was carried out on the prototype. For time and computation considerations, all simulations were conducted on the model symmetry. Therefore, drag and lift forces obtained were half of their actual value.

The corrected simulation was then carried out using a wind flow of 89.4 m/s (200 mph) with modified Turbulence intensity and Intensity scale length characteristics. Correction made between the prototype and the full-scale STS 444 were based on location, type of grounds, weather conditions, materials selection and geometrical modifications as listed in the table below.

Table 4: Parameters for the full scale model and reduced prototype in the CFD modeling

Results

Wind speeds from 0m/s to 12.01m/s.

Table 5: simulation values.

Figure 12: flow simulation (fluid at 12m/s)

The graphs of drag coefficient with polar angle were plotted for the three approaches that are analytical, experimental and simulated. A linear increase with the increase in polar angle was shown in all the three methods. The endpoints for all three methods coincided with some deviation occurring in between depending on the way they were approximated. Graph of drag coefficient against polar angle for analytical (blue), experimental (red) and simulated (green) showing corresponding drag increased almost linearly with the polar angle. The graph is shown below.

Figure 13: A graph of drag coefficient against polar angle for the three methods that is analytical (blue), experimental (red) and simulated (green). The end points coincide since most of the approximation occurred in between the end points.

Figure 14: flow simulation with a considering the lower part of PV.

The graph of lift coefficient against the polar angle for the three methods was plotted as shown below. Comparison between the three approaches revealed that all the methods followed different trends about how they were approximated.

Reynolds number

Re = which is equivalent to

= 339378343.9

CD = = = 1.874 x10-3

Drag on both sides of the photovoltaic collector is given by the following equation

D = 2(CDV2A)

= 480N

Gear

Internal diameter = 36.6mm

Outer diameter = 296.56mm

CD 2.1 D = (CDV2A)

= 17204N

Total 17684N

Figure 15: simulation

Figure 13: A graph of lift coefficient against polar angle for analytical (blue), experimental (red) and simulated (green) showing all three approaches with different trend for different angles. Due to the different ways in which the values were approximated, they tend to converge at the endpoints but diverge in between.

A graph of corrected drag and lift against polar angle plotted showed no much significant discrepancy between the experimental drag and the adjusted drag. The trend was also found to be relatively linear. A similar result was observed between the simulated lift and the fixed simulated lift with a concave down shape, reaching the peak at around 40 – 45 degrees.

Figure 14: Graph of corrected drag and lift against polar angle for full scale STS444

Maximum Von Mises stresses were applied to the simulated system with the appropriate material assigned to the different parts: concrete to the base, steel to the frame and aluminum to the panel. Maximum stresses were observed to occur on the steel beams at the horizontal and vertical beam junctions. Maximum weights for different polar angles were listed in the table below, and they followed an increasing trend as the polar angle was increased. For a tensile strength of 250 MPa for steel, all the maximum stresses exceed this value by far which indicated failure of the beams at those locations of maximum stresses.

A contour plot of safety factor of 80 degrees was plotted with a safety factor of more than 1.5 being acceptable. The safety factor was zero at places where the maximum stresses exceeded the tensile strength of the respective material as it was expected. Based on this model, only the horizontal and vertical beams would be subject to failure at the junctions.

Discussion

According to the results, drag coefficient was directly proportional to polar angle as was expected. As the angle was increased, the surface in contact with the wind increased which increased the resistance of flow opposing the wind load. Corresponding drag is directly proportional to drag for constant variables. This expected trend was confirmed by all three approaches with deviation observed in between the endpoints. The surface areas of contact with the wind are clear-cut, and it is easier to approximate without introducing too much error at the endpoints.

For example, at 80 degrees, for instance, the panel consisted of mostly two rectangular shapes: the base and the panel as seen from the frontal area which introduced minimum error in the analytical approach. Whereas in between the endpoints, the systematic approach assumed the flow to be blocked entirely by the surfaces lying in the path of the horizontal wind flow. This resulted to an overestimation of the drag. In both the experimental and simulated, however, the flow continued past the panel as expected to happen in reality. The shapes of the parts are also very crucial in estimating the analytical drag coefficient for the system. This explains why the analytical drag coefficient was much higher than the simulated and experimental.

Consider the resultant force as the force perpendicular to the panel surface. The horizontal and

Vertical components would be taken as the drag and lift. When the angle is zero, the force perpendicular to the panel surface is at its minimum so that the lift which is in the same direction as the resultant force would be at its minimum. As the angle is increased, the resultant force acting on the panel would increase and so will the lift. When the angle is 45, the resultant force is at its maximum. As the angle is increased the vertical component will decrease to almost zero. The parabolic trend for the lift, therefore, would be expected to be concave down. The general trend yield by lifts in all the three approaches was an increase in CL as the angle is increased. The analytical approach, for example, was bound to introduce significant discrepancies since it was calculated from the sinusoidal component of the drag. It was expected to follow an inverse tangent curve as expected with an undefined value at zero angles.

To prove the effectiveness of the turbulence on the simulated values, the same set-up conditions were simulated one with laminar and the other with turbulent flow at zero angles. The pressure distribution for laminar flow resulted in the greater pressure difference between the top and bottom pressure on the panel than for turbulent flow.

It is hard to achieve the three approaches regarding geometry, surface roughness, and turbulence model and measurement standards. The beams of the model were modified to a flat plate to prevent the prototype from breaking.

The nature of flow was found to be affected significantly by surface roughness. The roughness of the walls was assumed to be negligible to keep the data as generalized as possible. Other sources of error included bending in the fixture and prototype frame support, human error and other measurement inaccuracies and limitation.

Snow load analysis

Although snow loads are not as dangerous as compared to the wind loads, they contribute a lot to factors affecting the linear actuator kinetics. Modelling of a snow sliding profile was crucial in finding the time taken for the snow to be completely stowed and the maximum polar angle required. The snow sliding pattern was also significant in determining the torque applied to the linear actuator shaft by the sliding snow.

A linear actuator enables the panel to rotate hence towing the snow. The mechanical speed and electrical power of the linear actuator depend on the maximum torque exerted by the snow load. Therefore, driving a snow torque profile to determine maximum torque exerted by the snow load while sliding off was significant in selecting the right linear actuator.

Model parts and assembly

The package required knowledge of the snow sliding profile needed for the simulation set-up. The biggest problem in this analysis was to simplify the snow sliding profile. Therefore, assumptions were required for the snow unloading simulation with detailed simulation parameters as given.

i. Snow was assumed to be a discrete load and was modelled as a part to capture both the change in load direction and load magnitude.

ii. Snow was taken as fresh new snow with the density of 50-70kg/m3.

iii. The static and dynamic friction of snow on PV panel was approximated as 0.1 and 0.03 respectively.

iv. All other connections and contact surfaces were assumed to be frictionless

v. Linear actuator was assigned a constant speed

The solar tracker system subassembly consisted of 4 parts: panel, linear actuator motor housing, linear actuator shaft and frame support. The frame was assembled with a rigid connection while the linear actuator motor housing was the constraint with a pin connection at the axis of rotation with the frame support.

Mechanism analysis

After assembling the various components, a defined linear servo motor was added to the linear actuator shaft slider connection at a constant speed to control the rotation of the panel as well as gravity. The angular position, velocity, acceleration and torque measurements were defined at the pin connection between the panel and the frame support while another torque measurement was defined on the linear actuator shaft.

Material for snow at the maximum weight was applied to the snow part and aluminum applied to the panel and linear actuator parts. The simulation was run for the servo motor acting downwards with constant velocity at zero degrees with gravity taken into consideration. The distance between the edge of the snow relative to the side of the panel was the only parameter needed for this simulation to obtain the distance slid by the snow along the panel.

Figure 16: Separation measurement taken between the edge of the snow and that of the panel to obtain step-wise value of the respective distance with time

CHAPTER IV

Conclusions and recommendations

According to the results, the snow load was not as crucial as compared to the wind pressure by at least a factor of 10. The wind pressure, however, can cause an enormous amount of mechanical damage if the frequency of the wind is equal to that of the mechanical system. This requires optimization of the beams supporting the panel and the base based on the wind force and stress analyses while the snow load analysis will be mostly geared towards selecting the right linear actuator.

Increase in wind velocity results to increase in the angular position of the solar panel as well as the stress. However, the simulations show that we mustn't only focus on the largest angles since some moderate angles can cause high stress which may lead to failure due to fatigue.

Based on the stress analysis of the solar tracker, it was recommended that the vertical beams supporting the panel should be increased and reinforced further to prevent failure. Weights of beams and panel parts need to be included in the selection of the linear actuator because it helps in predicting the maximum torque of the snow.

Both the wind and the snow load are considered critical to optimizing the design of the STS 444. Work will focus on optimization of structural design, and how parameters lie energy power to the driving actuators and gears affect the STS 444 depending on the speed of the cycles, the amount of tracking, type of monitoring and configurations depending on the required conditions.

CHAPTER VI

FUNCTIONAL MODEL

The functional model in this chapter is supposed to demonstrate to customers how the STS 444 which is a large-scale dual axis solar tracker works. And thus it is imperative to note that I do not take credit in any findings and design of the STS 444 as my project findings and is mainly for aiding in my research in the solar tracking system construction and workability and highlighting on the modelling and simulation process of design which will equip me with scientific skills of design.

There are two main constraints in the design process of this design namely that the design must have an optimal efficiency and have a rotation of eighty degrees around the horizontal axis. Secondly, the solar panel should be 2m by 2m.

On a light note most designs on the market have been patented, or in the process of acquiring patents, it was difficult to obtain detailed measurements and design charts to aid in this project. Design challenges were however reduced through referring to a similar project by four students in the Worcester Polytechnic Institute which detailed the structure and performance of STS 444 tracker in which they performed the relevant wind loading analyses through wind tunnel testing technique, simulation, and analytical approaches [1]. The design of this solar tracker is thus based on the STS 444 solar tracker design specification in which the panel I installed on a dual axis tracking structure equipped with a linear actuator fixed on brackets and a couple of sensors to aid with the tracking mechanism.

Power generated will increase by averagely thirty-five percent. Factoring in the lifetime of the system say it is twenty years with a reduction of five percent in efficiency every year then the solar panel will have to be replaced within ten to fifteen years for maximum output.

Support frame

The metallic frame of the dual axis tower consists of several rectangular cylinders which are welded and together form a compact structure made from tough steel metal for maximum stability. With the assumption that the complete system is made from steel then the total mass would be equal to 81 Kg

Concrete base

The reinforced concrete base is not included in the system though it forms the substructure upon which the system is installed and thus holding the total weight of the system. In this project reinforced concrete is to be used at the base for system support

Linear actuator

The motor controlling the panel’s rotation to cause the desired tilt from zero to eighty degrees angle forms the actuator. The movement required should be smooth and accurate with regards to the triggering circuit hence a sophisticated motor control system will be required. In this project rotary encoders that reduce noise through the smoothening of the sign wave of the voltage applied to the actuator will be used as the motor controller. With regards to installing the linear actuator are much easier to install as compared to hydraulic and pneumatic systems since they are easier to install and additionally take up less space. They also provide precise position feedback which is desired to keep the tracking system in real time, the linear actuator also controls the motion and with regards to acceleration and velocity.

Rotational motor

The rotational motor is installed at the base of the tracker and is thus used for the azimuthal angle adaption, and it thus rotates the panel around a vertical axis which runs through the middle of the motor consequentially allowing the panel to track the sun from the east to west direction in real-time. For increased efficiency, it is imperative to custom makes a motor or customizes an already existing motor. The motor rotary motion and response will thus act as per the user specification and thus more efficient though this approach is costly.

12 V battery

The solar tracking system is power consuming since the solar beams have to stay at 90 degrees with the CPV tracker. Hence power has to be available and stable throughout the time of operation of the system. This battery can, however, be recharged by the solar panel during the day hence self-running cost efficient.

Electronic Box

The electronic box of this system has a microprocessor and some sensors. The sophistication of this circuit is of different level as to the level of production. Similarly, the microcontrollers will also differ with this regard. During the modeling stage, it is possible to use a microcontroller such as ATMEGA 328 which is programmed using the AVR software or with a microcomputer such as the Raspberry pie. These systems are however not cost effective in the case of production hence the use of other chips such as PIC is highly recommended. Additionally, the pic chips are chip and also machine friendly hence it is recommended for the modeling stage. Additionally the circuit contains sensors that will bring the rotational requirement of the system into a reality. Again, in this case, the robustness of the system is important. At low levels one could use a simple photo sensor such as the TSL2561 though when it comes to the installation of large solar panel systems then more accurate sensors will be required since the hobby photo sensors often tend to be unstable. And at times might either be slow or fail to sense the light with regards to the point of high intensity. Another important sensor is the included in the design of an anemometer, it takes in the speed of wind and strength and thus gives a reading to the circuit to calculate the loading factor. If the conditions do not lie within the specification of operation of the system, then the system adjusts itself to a safe position until the required conditions are back.

The electric circuit design process is less time to consume and also easy since there are several simulation software available on the internet that one can use to measure all the necessary parameters before circuit production.

NPV Analysis

With regards to the size of the solar panel and power generation specifications then it can be assumed that the panel will produce 0.7kW of power. Hence total power per system will depend on the number of functioning solar panels installed [1]. The annual power generation of in the machine says the panel works efficiently for nine hours per day then it will produce a total of 2212kW when the efficiency of the tracker is added as a factor in the power production. CAD design

The Cad design of the tracker has a hydraulic provision for shock and vibration s from the wind loading and turbulence

The photovoltaic panels are mounted on

Figure 17:

Figure 17:

Design specifications

1. The total system weight should be thirty six kilograms or less

2. The panel degree of freedom should enable it to rotate at 360 degrees around the azimuth axis and also have a maximum tilt angle of eighty degrees with the polar axis.

3. It should also have a snow removal simulation and an accuracy of five degrees in solar tracking

4. The model should be based on a three piece architecture in which the panel, base and supporting frame can be disassembled, this assembly should also fit in an envelope of 62 by 32 by 12 inches for ease of transportation.

5. The assembly should also be easily be disassembled within two minutes.

6. This model should also contain a 12V battery that will be running the electrical circuit and the liner actuator

7. The parts for the project should also be easy to acquire and compatible and lastly easy to trouble shoot in the case of mechanical malfunction by a trained mechanic.

Design Description

The photovoltaic panel achieves a rotary motion from a system of linear actuator and a slew gear. The altitude angles are controlled by the linear actuator while the polar motion is controlled by the slew gear. The following assembles are an extension of the parts already listed above.

Linear actuator positioning

The linear actuator position has its attachment points defined by the stroke length and size of the actuator. The transmission angle and maximum angle of tilt of the panel also form the design considerations for the linear actuator. To find the location of the linear actuator graphing and detailed analysis is applied and by extension, holonomic constraints are considered.

In this design, a small crack is not alarming since it is a small prototype, and thus damage is least likely to occur. While in the large STS 444 structures a crack will lead to creeping of the structure and eventually cause it to experience damages that at times are irreparable hence it is imperative to compress the linear actuator.

Linear actuator selection

The election process of the linear actuator falls under the dilemma of equipment procurement in which cost plays a great role while availability, specifications, and performance weigh on the other side. Online electronic sites such as hobby king eBay and ServoCity offer a diverse range of actuators that fall within the budgetary allocations of this project. They have a maximum load capacity of five hundred and twelve Newton, and the maximum speed is 0.05 m/s.

Linear actuator Dimensional limitation

Dimensional specifications of the linear actuator create a critical limitation when considering the size of the solar panel and the loading effects on the assembly. Thus this parts were modeled with strict adherence to standard dimensions of the available materials and also with regards to the additional limitations explained in the following sections.

System kinematics

Kinematic systems are based on the number of link and constraints since they will determine the degrees of freedom of the system. The system in this design is a four – bar linkage with the links at the ends grounded and link three having two degrees of freedom since it’s a crankshaft. Additional, the latter is considered as the input link. Link 4 and link 1 are pivoted to ground hence when link 3 is assumed to be frictionless and to move at a constant velocity. Link 4 which is the output (panel) will move in a clockwise direction following an arc of up to eighty degrees along the polar axis. The mechanism should thus exhibit angular accuracy as the servo motor inputs rotary motion concerning the rotation of the earth about the position of the overhead sun.

Figure 18: four – linkage equivalent of the linear actuator

Design: first iteration

Design positioning iteration of the linear actuator is in this project assumed to have its bottom base psssing through the center line of the pin connection between the frame support at point B and the panel at the point O as shown in the following figure

Figure 19: First Iteration side view.

xy =

a graph of y (f(x), g(x), h(x)…….k(x)0 against the distance x from the center 0 where the functions are the y distances for the linear actuator 1- 6 respectively which are available from actuator manufacture gives the strokes length range and also the products of x and

Table 5: Product of XY and Stroke lengths.

The values of x and y entered for the analysis are chosen with reference to the rule of thumb such that low values of y are desirable since it reduces the amount of material size while on the other hand high values of x will produce lower torque coupled with greater sensitivity. However, the actuator will be protruding from the sides of the design in the case of high values of x and this might make the linear actuator design to be cumbersome. Additionally from the set values of x and y, the lower stroke lengths turned out to be shorter than expected thus requiring greater torque while the higher strokes lengths led to a heavier linear actuator

From the tabulations and results from the graphing the desirable actuator type was type 3.

Design limitations

Points of attachment of the linear actuator on the beam and panel are considered to be lying right on the mid- axis.

In reality the rotational point is found slightly offset along the y axis while in the modelling process it is assumed to be at the center.

Dynamic motions are not taken into consideration while calculating the dimensional since only the extreme positions are considered.

Second design iteration

When the constraint at point B is removed then this iterations addresses the maximum tilt angle which in this case is eighty degrees, this analysis is possible since the removed constraint was at the point of attachment between the bottom of the linear actuator and the reference frame in the design.

Figure 20: iteration 1 side view with an offset on O to O’

From the diagram we can then obtain the optimum value of δ was derived using both geometrical and analytical approaches. From the analytical approach, two possible solutions were obtained for each specific length the linear actuator that had been graphically checked and then their transmission angles were subsequently compared with both solutions. From the deductions made from the latter solutions, we observe that the linear actuator is more stable when its center is closer to the center of gravity of the support frame. Take note that as for the second solution the values of x and y result in a larger distance from the two centers of gravity. The large difference results in instability that is already being caused by the momentum of the moving parts.

Third iteration design

Figure 21: six bar linkage mechanism

In the first linear actuator diagram, a four-bar linkage was used to archive a rotary motion of eighty degrees. Similarly, this six-bar linkage can be broken down to two four-bar linkage mechanisms as shown in the above figure where the links 5, 7 and 8 form one mechanism and the other links 4, 3 and 2 form the second mechanism.

From the design, the linear actuator length was limited by the extension and retraction lengths. These lengths made it possible to optimize the system to the maximum polar angle. In this case, both transmission angles µ1 and µ2 each accounted for 400 degrees thus making up the desired 800 degrees, and this four-bar linkage optimized design was achieved using the linkages software. The size of the optimum mechanism design dimensions was then scaled to depending on the size of the chosen linear actuator’s stroking length which in this case is 15cm.

Power analysis

Linear actuator

The linear actuator power estimation process is of importance since it brings forth the feasibility estimate of linear actuator motor and the amount of power required for it to follow the assigned kinematic motion. During the simulation process to ensure the validity of the simulation additionally, it is preferable to execute the azimuth rotation for less than fifteen seconds which is the half cycle. It was also assumed the acceleration of the model only lasted for a quarter of a second only. Calculations involved include mass moment of inertia, power required to run the actuator was analytically found to be 0.16W. During the simulation process when the linear actuator was assigned a constant velocity a varying angular velocity of between 4.5rad/s and 6.4rad/s occurred. Then as the simulation power increased to a maximum of 0.7W there was an increase in the angular velocity while a deceleration of 0.9rad/s2 occurred.

Pla is the power that the linear actuator needs to rotate the panel initially.

Pla ≥ Iωp

Figure 22: Graph of power against polar angle.

From the graph the maximum power reached is approximately 0.73 at a polar angle of 80 degrees. In this analysis the half cycle lasted for sixteen seconds.

Gear

Using a spur gear and a motor it is possible to create the tracking motion of the functional model. This motion is about the azimuth axis. For optimal motion of the system then the motor needs to have a low rpm while the spur gear should have a high pitch for a smoother motion. Using the equation used to calculate power the power needed to power the tracker can be calculated, this is in consideration that the polar rotation should be less than thirty minutes and also that the acceleration time for the model was one second. The torque required by the motor to rotate the system is given by the formula:

Figure 23: spur gear

The torque required was 67.2 Nm although this torque was more than the torque required by the tracker, thus using PWM the signal to the motor would be modulated to achieve an effective torque that will saves on the battery power usage.

State equation for the system

To get the state equation of this system’s motion the Lagrangian method played an important role and through it the equations of motion of the linear actuator and gear driving motion were obtained. During the derivation process three assumptions were made.

1. All the parts in the assembly are rigid thus only changes in the CGs will change the system’s energy.

2. Pin jointed linages are assumed to be frictionless, this is because they have low friction bearings that have an efficiency of approximately ninety five percent.

3. The presence of dynamic forces in members moving at a moderate high speed often cancels out the effect of gravity on the system. However, on this assembly, motion is relatively low hence less dynamic forces and consequently a greater effect of gravity on the system.

Assembly

Figure 24: gear system assembly

Figure 25: beam and gear assembly

Figure 26: panel mount assembly

Figure 27: panel mount and beam assembly

Figure 28: panel assembly

Figure 29: final assembly

Conclusion

The functional model for this project met all the design requirements for a standard solar tracking system with a final weight of approximately 62lb with the three sensors integrated. It could easily be assembled and disassembled by one or two people at most. Finally as for the efficiency of the model the polar resolution was 0.4 degrees for the polar and axis.

References

[1]Two-axis tracker for solar panels and the like. Washington, D.C.: United States. Dept. of Energy., 2013.

[2]V. Poulek, A. Khudysh and M. Libra, "Self powered solar tracker for Low Concentration PV (LCPV) systems", Solar Energy, vol. 127, pp. 109-112, 2016.

[3]"Fuel Cells 5/2015", Fuel Cells, vol. 15, no. 5, pp. 648-649, 2015.

21