Robotic Device for Cleaning of Photovoltaic Arrays V2

Transcription

1 Robotic Device for Cleaning of Photovoltaic Arrays V2 Design Team Greg Belogolovsky, Steve Bennett, Istvan Hauer, Salome Morales, Leonid Nemiro Design Advisor Constantinos Mavroidis, Ph.D. Richard Ranky, M.S. Abstract Solar energy and its popularity has increased in the past decade due to heightened electricity costs and an ever growing public concern for environmental responsibility. In order for solar panels to produce the maximum amount of energy, it is vital that the panel surface be free of dust and debris. The collection of debris on the panel surface can decrease efficiency output by 20-30%. The decrease in power efficiency is magnified at a solar farm, which can house thousands of solar panels, leading to a considerable annual monetary loss. A calculation for the Edison Technology Park in California which houses over 2300 panels shows that when solar panels are cleaned, $73,000 of energy is saved; when the panels are left dirty only $51,000 is saved. This results in an annual loss of $21,000 of energy. It is the group s goal to eliminate this loss of efficiency for all solar farms using a state of the art autonomous cleaning system. For the solar panels to achieve maximum efficiency, the panel surface must be cleaned between 1-3 times per year depending on geographical location. Current cleaning methods of solar panels include manual labor, which is not cost effective when cleaning a solar farm with 120,000 panels such as the one in Jumillia, Spain. Alternatively a sprayer system can be housed to the frame of the solar panel that washes the panel down with water. This system becomes complex and unprofitable when having to embed thousands of these installations at a solar farm. The goal of the Robotic Device for Cleaning of Photovoltaic Arrays V2 is to evaluate the power efficiency of the existing V1 system and modify key assemblies to improve overall cleaning speed and effectiveness. Primarily, the V2 robot must clean a solar panel faster than a human in order to show its effective efficiency. Figure 1 shows the V2 robot on a solar panel, highlighting the key functional components of the system. Top Trolley Cleaning Head Figure 1: V2 Robot Linear Track

2 The Need for Project Most current cleaning methods for solar panels are conducted through manual labor using brushes and water hoses. An automated solution to cleaning panels can be achieved to cut down the expenses of manual labor. Figure 2: Dirty Solar Panels The proposed method of cleaning by the V2 team is unique in its application. An automated solution currently does not exist in industry and, previous to the V1 team, had not been attempted at this scale. Figure 2 shows what a dirty solar panel looks like. A dirty solar panel like this puts out a decreased power output efficiency. The V2 PV cleaner would operate at a cost effective price which is further discussed in the Financial Issues of this document and also the report (Ref ). Applying a small business plan to the V2 robot, the group calculated it would take 1.5 years for a cleaning company to see profit using the system at a solar farm. This figure was found applying the cleaning robot at the Edison Technology Park. With a V2 robot being built specifically for the New England area, there would be considerable growth potential until the market is saturated. Profits can be maintained by rentals, repairs, and selling of new systems to replace outdated ones. This robot would fill the needs of the original V1 sponsor, Green Project, as well as other companies in countries with solar cell initiatives such as Spain and the UAE. The Design Project Objectives and Requirements The mission of the V2 team is to Design Objectives address the cleaning efficiency The specifications for this project were broken down into two of solar panels by the robot. The separate categories. The first category envelopes the research robot will be cut down in size conducted by the V2 team contacting solar farms and solar cleaning and weight to make the system companies in the U.S., while the second takes some of the original more user-friendly compared to specifications given to the V1 team by their European industrial V1. sponsor: Green Project. The specifications taken from the V1 team were as follows: the robot must climb a 45 incline, weight must be less than 50 kg, robot must clean in the forward and reverse direction, less than 40 L of water per hour can be used, and it must operate in arid environments up to 35 C (95 F). Next are the specifications that the V2 team added: any part of the robot contacting the surface will induce less than 30 lbs/square foot of pressure, cleaning speed must be faster than 2 minutes/panel, and drive train resistance should be cut from V1 by 50% (Ref. Table 5 & 6).

3 Design Requirements Based on the friction testing conducted by V2 on the V1 design, it was found that the V1 system current draw was oscillating between 10 to 20 amps for both motors on the power supply. Each of the V1 trolley motors is only specified to operate at 2 amps. V2 plans to achieve 50% of the force required to move robot across the panel surface. Design Concepts Considered The V2 group developed four design concepts and compared them against the specifications of the project. Figure 3: Design Concept The first priority of the V2 team was to evaluate the V1 prototype. It was found that the prototype was not performing to the specifications set for V2. The main issues discovered were in the areas of size, weight, and function. Using the positive and negative design aspects of the prototype, brush manufacturers, solar farms and solar panel cleaning companies, several design concepts were iterated to improve performance. All of the redesigns would be rigid, with all rotating parts supported on bearings that would be half the size of V1 to overcome the weight and friction issues they faced. However, each redesign had a different implementation to achieve proper cleaning function. The first concept was a redesign of the V1 that would achieve the cleaning motion in the same way as the V1 prototype; vertical motion for cleaning, and lateral motion to move the assembly down the solar panel. A different version of this concept was proposed that would not use rotating brushes to clean, but an oscillating sponge. The second concept had the cleaning components aligned vertically to achieve cleaning of the solar panel array in one horizontal pass. The brushes, sprayers, and squeegees would be segmented and modular to facilitate cleaning of any sized commercial solar panel. The third concept was unique in that it was a small crawling robot that would raster on the surface of the solar panel itself. One drawback was operating at the extremes of the specified solar panel angel of 45 degrees. To overcome this, a ratcheting tether attached to one horizontal unit riding on the top rail was considered to overcome the tangential forces and prevent slipping. The fourth concept was an iteration of concepts 1 and 3. It consisted of a robot that would ride on the solar panel, but would be supported by a rigid frame resting on the frame of the panel. It was

4 novel in that it achieves vertical motion and rastering with fewer, multipurpose components (Ref. 4.1). This design can be seen in Figure 3. Recommended Design Concept The design that was chosen satisfied the following criteria that were important to the functionality of the robot: ease of manufacture, least amount of friction between solar panel surface and brush head, friction in drive train, and drive train efficiency. Figure 4: Brush Head All of the designs were evaluated based on how well they met the design specifications. From this a decision matrix was created to evaluate the concepts against one another (Ref. Table 8). It was determined that several design concepts ranked equally high and the most positive aspects were combined to formulate the final design. The final concept relied on a rigid frame that would clean in a horizontal path. It would be much lighter and would contain fewer components than V1. The cleaning head would tilt depending on motion to the left or right. This tilting would also allow for a controlled force on the surface of the solar panel. Two water delivery pipes will be connected in front of both brushes to an external water supply spraying water onto the panel surface. One of the greatest difficulties that the V1 design experienced was the friction from moving laterally across the solar panel. For the wheel design of V2, slippage is of major concern to the autonomous operation of the cleaning system. To maximize contact surface, the wheels have a 90 degree V-shape cutaway that enables contact on the solar panel with the highest normal force. The wheel design can be seen in Figure 4. Large amounts of slippage on any one of the driven points of contact can result in putting the load onto one single motor. To avoid this, it was determined that a custom wheel was to be designed to provide for a high coefficient of friction at points on the panel that will exhibit the highest normal force during operation. In order for the V2 robot to be adjusted to different sized solar panels, it is important for the team to demonstrate adjustability of the top trolley mounting with the linear rail. The adjustability of the prototype system should be specified to fit any solar panel that is no more than 82 in height. This measurement is the current height of the solar panel experimental array that V2 will be testing on. The system can be adjusted by the slotted brackets that hold the trolleys to the linear track. For all dimensions of solar panels down to 5, the trolley can be adjusted via the bolt holes in the linear track and the slots in the

5 bracket. This was the simplest adjustment system and requires minimal amount of manufacturing because the track comes stock with the aligned bolt holes in its base. The adjustable mounting bracket seen in Figure 4 allows the top trolley to be lowered to the necessary height of the solar panel. Financial Issues The total cost for the V2 robot was $2500. This includes all hardware components along with raw materials Figure 5: Manual vs. Robotic Cleaning Cost From a financial standpoint it is important to see how using a robot would be beneficial in cleaning a solar panel compared to a human. Figure 5 compares the cost to clean solar panels in three different states. Based on surveys conducted with commercial solar farms, cleaning occurs up to three times a year by humans at the cost of $1 to $4 per panel, taking the average solar panel size to be 4 feet x 2 feet x 2 inches. The blue column in the graphs shows the higher end of cost to clean a panel by a human per the specified states. The red columns show the lower end of the cost to clean. The green columns show a robot cleaning the panels with a human operating the system. Comparing the three columns shows that using a robot is much cheaper solution than hiring a human crew. These calculations were done taking into account minimum wage per human, labor cost, and price of water per liter for each state. The total cost of the V2 robot was $2500, which is 31% of the V1 project cost from the previous year. Breaking the V2 system by components, the following was spent: seven DC motors and gear boxes were $925, hardware/raw materials were $840, and the linear rail cost $735. Fabrication for 90% of the hardware components were completed by the team in-house. Recommended Improvements The system can be improved by implementing advanced closed loop operational system. Further advancements can be examined for implementing automated inspection and data logging devices to the robot. Inspection of solar panel with an infrared camera would give the robot the ability to both clean the surface and inspect the grid in the panel for damaged photocells. It could also give feedback as to whether cooling the surface of the panels with water could improve the solar panels performance in especially hot desert environments.