Wind Power Light Information
The design of the wind-solar parking lot light was based upon the previous senior design project P05307 but focused on several design upgrades which would allow the wind powered light to be displayed prominently on the RIT campus as fully operational and durable. Some of these changes include: not using a solar panel, using RIT's existing 14 foot light posts for mounting and using a LED light to minimize the power requirements for illumination.
The walkway light was erected on April 10th and was dedicated to the RIT community on May 4th 2007. According to the results of an aesthetics survey, the general sentiment from RIT's students and faculty was positive with 76% of those polled answered that they really liked the project. The group has made several suggestions as to how the walkway light can be improved and hopes that future senior design teams continue to improve upon the design and integrate several more wind powered lights onto the RIT campus. This will hopefully contribute to an overall greening trend as RIT moves into the future.
Senior Design 1
MSDI Schedule/ProjectSDI Project Plan
Customer Needs and Design SpecificationsCustomer Needs and Engineering Specifications
Concepts ReviewWind-Light Concept Review Reading Packet
Design ReviewWind-Light Design Review Reading Packet
Prelimiary Test PlanTest Plan
Senior Design 2
MSD2 Schedule/ProjectSDII Project Plan
Detailed Design ReviewWind-Light Detailed Design Review Reading Packet
Function and Performance ReviewWind-Light Funcation and Performance Review Packet
Technical PaperWind-Light Technical Paper
MotivationThe senior design team has built the wind powered walkway light in an effort to show the feasibility of alternative energies from a pollution, cost and aesthetics standpoint:
- Cost: The team hopes to have created a design which will be cost competitive with the current system. A successful design will showcase the use of wind power as a financially feasible alternative to using typical power from the grid.
- Pollution: The design is powered solely by wind power, because of this there will be virtually no carbon footprint compared to the annual use of a regular light. A commitment of RIT towards carbon-free energies will help to promote its environmental image.
- Aesthetics: The design emulates the current lighting systems as closely as possible. The team wants to showcase wind power but is also determined to create a design which will not be deemed a campus eyesore. The goal is to prove that green energy does not need to be ugly or unwieldy.
Project ObjectiveThe wind powered walkway light was designed to be seamlessly integrated as a replacement for the current 150 watt metal halide walkway lights on the RIT campus. The light provides power off the grid and be will be completely self-sustainable and should remain relatively hands-off for facilities, with battery replacement being the only needed servicing over the life of the light. The group completed an overall analysis to ensure that the light met safety as well as campus lighting standards. The design team provided the following in detail to showcase the design and its proof of concept:
- Brainstorm possible alternatives, but attempt to use similar components of current light systems.
- Analyze wind direction and speed to size battery components
- Stress analysis to ensure survival of all structural components
- Design and build the system within the pre-approved budget
- Analyze the LED light to determine its ability to supply light equivalent to the current lighting systems.
- A survey to determine overall campus satisfaction of the completed project.
Fundamental Physical Concept
The goal of this concept was to produce a 2nd generation stand alone roadway light that utilizes hybrid technology power generation for best performance. According to the analysis done by the 1st generation of this project P05307 hybrid technology is the best way to produce enough power for a stand-alone roadway light.
February Design ChangeA dramatically more efficient turbine was chosen for the final design which ultimately led to the abandonment of the solar panel.
Average wind power is based the power curve and wind data provided by the National Climatic Data Center. Height difference between the anemometer at the airport and our turbine height were not accounted for in the calculation of average wind power. Other energy losses from battery/circuitry were compensated by deducting 10% of the power generated for the turbine and 20% for each of the solar panels.
The economics are simple. It is approximately $250 for a solar panel and $150 for another battery. The mechanics are simple as well and extra battery can be mounted to the base with almost no effect on the overall performance of the pole. The solar panel not only will add weight but may act as a sail during certain wind conditions. Both of which will add stress and affect the performance of the pole.
This design change that we abandon the hybrid design and focus solely on wind power was approved February 7, 2006 by Cathy Ahern, FMS Senior Mechanical Engineer.
Pursued SolutionAccording to Gary Prokop it costs approximately $3600 to install a 30-ft tall roadway light which uses a 400W HPS bulb. From a light survey done by Fisher Marantz Stone for RIT in 2002, a roadway or parking lot light needs to produce 0.6 foot candles of luminance. Using an average cost of energy equaling $.08 KWhr, the cost to run this light would be about $140/year.
Further Design ChangesAfter meeting with Dave Harris, the Director of Training, Utilities, & Environmental Management, on Jan 26th we turned our focus towards the 14' lights on campus. These have a 4" square pole with constant cross sectional area unlike the 30' pole which are round and tapered. There are several reasons to choose the 14' versus the 30'. One being the issue of the feasibility of us working with any design at 30' and another being our concern that the LED fixture that has been chosen may not produce the luminance need at 30'.
We analyzed weather data for Rochester for the past 10 years to estimate the performance of our chosen 400w turbine. Based on this data we found that there is only a 5% chance there will be 48 consecutive hours without any wind power generation. 48 hours is significant because we begin to hit a threshold of low charge on our battery at that time. The idea originally was that the solar panel would provide the extra power need during this time. But we also found that on average we would be producing 800 Whrs extra every month which would be dumped to ground due to lack of energy storage capacity.
10-year of Historical Wind Data taken from the Rochester Airport by the NOAA (National Oceanic and Atmospheric Administration) was used to analyze the wind behaviors on RIT's campus. The data was also utilized to predict the normal operation performance, worst case battery analysis (percentage of uptime vs. downtime), number of consistent hours without wind in the Rochester region, and feasible locations for the wind-powered light. Due to the complexity of the analysis and the amount of data available, all related data files are zipped and linked under Battery Analysis below for future use.
Types of Turbines
|The chart on the right shows an in-depth search of wind turbines which were suitable for the application. This shows that the Air X surpasses its competitors when it comes to production for value.|
|The chart to the left compares the Air X and the Air 403 wind turbines developed by ETA Engineering. It was later found out that the 403 is no longer in production.|
Supplementary Wind Turbine Files
|AirX Manual||This is a PDF file that is the Air X Owners Manual. This file includes a variety of useful information about the wind turbine.|
|AirX Specs||This is PDF file includes information on the wind turbine performance charts as well as its dimensions, voltage, rated power and surival wind speeds.|
|Hours of Wind Distribution||This is Microsoft Office Document file that provides more detailed information on history of the winds at the Rochester Institute of Technology|
LEDThe light selected to be used for the wind walkway light was a warm white, 12-14VDC 20 Watt LED. This is a key component to making the system work, Supplying power for a 20 W light is much more feasible than for a 150 W light. According to the manufacturer's recommendations it is suitable for the replacement of 70, 100 and 150 watt lamps. The LED light was purchased for a cost of $725. This high price is due to the infancy of LED lighting technology at this point in time. Because both the turbine and the light operate using 12 volts DC, it could easily be integrated with the 12 volt turbine. This also allowed the group to avoid using an AC/DC inverter, thus eliminating inefficiencies of 5-15% for power conditioning. (Please note that later testing of the product revealed that the light actually operates at 30 watts.)
A test was performed upon receiving the LED fixture to assess the quality of light that it provides. Field data was gathered using a light meter, readings were taken at ground level from directly beneath the light, to distances 25 feet away. A comparison of the results with the those of the 150W metal halide light that RIT currently uses is shown below.
The minimum illuminescence requirements for outdoor campus lighting are .5 foot candles at 25ft from the source. The LED lamp exceeds this requirement up to 5 feet away from the light. However, there is a sharp dropoff at this point as anything beyond 10 feet does not meet the lighting standard. Note that the LED light is 1.8 times as bright as the standard metal halide fixture directly beneath the source. There is a lot of light available, but it is not disbursed well. The metal halide light has a much more even distribution of light.
Supplementary LED Files
Battery AnalysisAn energy storage analysis to find a suitable battery was also conducted. The battery needed to be able to store enough power, based on the turbine output, for the light to remain on without any power interruptions. Because of the nature of the application a deep-cycle battery was desired. The reasoning behind this choice is that the life of a normal battery is significantly reduced the deeper the battery is cycled. The selection of a deep cycle battery allows a more liberal range of discharge depth without sacrificing battery life. Based upon the manufacturer's specification the threshold voltage for the battery is 10.2V, approximately a 30% depth of discharge which allows for longer periods of low wind (no power production).
The selected battery was the Optima Yellow Top DEEP Cycle D850U 55ahr (660 Wh) which cost $160 each. Based upon the wind analysis a graph was created which shows the state of charge of the battery based upon each months 10 years of historical wind data. The figure below shows that based upon the historical data and the capacity of the battery that the lowest state of charge which the battery will experience is around 89% during the month of September. This is well above the maximum depth of discharge (70%) quoted by the manufacturer.
|The graph on the left shows a 10 year historical data compared with the purchased battery and shows the lowest depth of discharge the battery should experience, this is around 11%|
Battery Endurance TestA battery endurance test was completed to analyze the nature of discharging the battery. This was done to gain some insight into the ability of the batteries to power the light for the maximum required amount of time. The test was setup by first charging the batteries to 12.9 volts. They were then discharged by making a connection to the LED light. A shunt resistor was placed in series with the light so that the power being supplied to the light could be determined. Readings were then taken over time of the battery and shunt voltages. These were then used to find the current and then power being supplied to the light. The two graphs shown below represent two important findings that came from this test. First, the battery voltage did not drop below the threshold voltage of 10.2 until 30 hours after the test had been started. This means that the batteries can supply the light with enough power to run for more than two nights without any wind. Another noteworthy observation is the power draw of the light. It operates at 30 watts, not 20 watts as specified by the manufacturer. Fortunately, the batteries were still able to supply this power for the required amount of time. Also note that at a certain voltage (approximately 10.6 V) the power supplied to the light drops dramatically and stays below 5 watts. This suggests that there is a threshold voltage that must be supplied to the light for it to operate properly. The light still operates below 10.2 V, however it is much dimmer.
CircuitryThe wiring circuit was designed tailoring to the specifications within the Air-X manual. A kill switch, 50 amp fuse and a battery disconnect switch are all features which were added to the circuit to guarantee the safety of anyone interacting with the walkway lighting system. Also, several ground wires were used in order to prevent any shorting of the system and included a light pole ground, battery ground, LED light ground and a turbine ground. In addition to the built in safety precautions, a volt meter, amp meter and Doc Watson meter (which measures battery voltage, watt-hours drawn by the light and max amperage drawn by the light) were installed into the system which allowed for real-time monitoring of the current state of the system.
StructureStress analysis was done in order to determine if using the current RIT facilities light posts would be a viable option for mounting all of the system components. The design allowed for the turbine to be mounted atop a 3 foot section of 2" O.D. schedule 40 pipe connected to the pole via a metal top cap. The pipe and top cap were welded together and the top cap was bolted into the top of the light pole. This allows for the integration of the aluminum light pole with the steel turbine mounting post. The turbine manufacturer specified that the turbine must be mounted to a structure that can withstand 150 pounds of horizontal force applied at the turbine. Indications are that this 150 pounds of force is equivalent to approximately a 110 mph wind. The highest wind speed recorded at the airport in the last 10 years was only 70 mph.
The FEA analysis was completed using ANSYS with the post being restrained at its base in the x,y and z axis where the 4 inch square aluminum light post meets the cast aluminum foot of the pole. The load was then placed on top of the structure where the turbine is mounted. The results show that even at the maximum loading force the pole will not experience failure. Aluminums yield strength is 31 ksi and the loading which is seen with a 150 lbf being applied at the 17 foot apex of the pole is only 12.6 ksi which leads to a factor of safety of 2.46. The top steel pole experiences 17.4 ksi with a yield strength of 45 ksi leading to a factor of safety for the upper pole being 2.59. The simulation also showed that under maximum loading conditions the pole will see a 9 inch deflection where the turbine is mounted.
Although FEA analysis had shown that the structure was secure even under the worst possible loading situation, (150 lbf at the top of the structure), an extra safety measure was taken. A safety cable assembly was connected between the concrete base below the pole and the steel pipe used to mount the turbine. In the event of structural failure at any point on the pole this will help to ensure that the turbine or pieces of the post will not fly away.
Supplementary Structure Files
FEA AnalysisFirst FEA Analysis
ProE and Finite Element AnalysisProE Modeling Files
System ModelingOne of the major concerns with this type of system was the battery. Depending on the battery size and the environmental conditions, there was no way of knowing if there would be enough energy production to power the light for the length of any given night. The solution here is simulation. A simulation was done to see if there was chance for a period of time where insufficient wind and insufficient battery capacity would lead to a non functioning light. In addition, if this scenario were to happen, we wanted to see for how long. Analysis was conducted in order to determine the ability for the LED light to be powered based upon the 10 years of historical wind data supplied by the NOAA Satellite and Information Service. It was determined that based upon the use of our batteries coupled with the 20 watt light, that over the course of the year the light would have a 99% operational up time. This figure was amended when it was discovered that the LED light actually draws 30 watts as opposed to the manufacturer spec of 20 watts. This changes the theoretical uptime to 95% which is still seen as acceptable.
|The graph on the right shows the frequency of hours with insufficient energy to power the light.|
ResultsAfter running the simulation for 1 year and 6 replications, to keep randomizing the numbers generated, the following results were concluded. On average, the battery was kept at ~84% capacity. However, at one point in time the capacity reached extremely close to zero. Although the battery never died, at this point, there was still insufficient energy to power the light. During this period of time (6 years), the light could not function for a maximum of 23 hours. The forecasted uptime will now be closer to 95.2 %. Although the probability of this happening is fairly low, it is still a good idea to further analyzing the risks associated with this scenario.
Once the light was fully operational, data was gathered from the system to assess it's performance. This was done by physically looking at the meters in the circuit and recording their values a few times each day. The meters allowed for record keeping of battery voltage, amp-hours supplied, watt-hours supplied, and amps being drawn (if the light was on while recording). This data is tabulated below in a spreadsheet. Analysis was done to see what the up time of the light was every night. This was done by subtracting the watt-hours recorded one day from that recorded the next day. Assuming that the light operates through the night at 30 W, a 10 hour night should have recorded 300 watt-hours (.3 kW-hr). The watt-hours recorded every night divided by this number gives us a ratio of the amount of time the light actually was on. it should be noted that this is a conservative estimate of the up time. It has been previously noted that the light will operate below 30 W at lower voltages by dimming itself. So theoretically, the light could be powered all night at less than 300 watt-hours.
Daily Performance Summary Table
The results of this analysis were troubling, as it appears that the light has an up time much lower than anticipated. It should be noted however, that at no time did the turbine charge the batteries any higher than 12 volts. The system had been viewed at times, charged to 12 volts, with the turbine mechanically braking itself as if the batteries were fully charged. This suggested that the turbine was regulating to a lower voltage than what was desired. An adjustment was made to the regulator to fix this, however at this time the results of this adjustment are not available. The same type of analysis will be done once the appropriate data has been gathered.
Life-Cycle AssesmentA 30 year life cycle analysis was done comparing initial costs, maintenence costs, and electricity costs over th etime period. Costs that would be equivalent for both systems were not considered. Simple pay back is calculated as the difference between the initial investment of both systems divided by the annual energy cost savings.
Cost Summary TableTabulated below is a breakdown of the contributing costs to the system. This does not include parts and labor that were donated to the project. Note that the LED fixture was a disproportionately large cost compared to the other components. A budget of $3,500 was given to the team upon project approval. The final cost of the prototype $2,646 was significantly less than the intitial approved budgeted. This was due in large part by generous donations of parts and supplies from RIT FMS.
Survey ResultsA survey study was conducted after the working prototype was put in place. Below are the survey results:
RecommendationsMoving forward there is plenty of room to improve the design of the wind-powered light system.
The LED light is one area that could be changed in the future. The light was specified to operate at 20 W at the time of purchase. However, testing has shown that it actually operates at 30 W. Finding a light that has a lower draw would allow for more turbine down time. Additionally, the light does not provide the same quality of light that the metal halide does. It has been suggested that the individual LED's could be oriented to shine at different angles rather than all of them shining straight down. A lens could also be used to better disperse the light away from right beneth the structure. Another suggestion was to mount the lamp so that it is angled up a little from the post. This would correct the problem of light being lost behind the post rather then shining on the walkway. Finally, the cost of the compnent was disproportionately large when compared to the rest of the system. Since we are currently on the front edge of this technology it would make sense that the price and performance will improve as it moves towards mass production. It has been suggested that a future senior design team could take on the task of designing a light meant specifically for this task.
If this product was to be installed on campus in large numbers it would make sense to couple multiple light posts for each power generation system. For example, 3-4 lights could be powered by a single turbine and more battery storage. A study could be done, using the current system, to better understand the capacity of the system to power more than one light.
A better data acquisition system could also be very useful in the future. Gathering more data on the turbine side of the circuit with a watt hour meter would help provide insight as to how much power the turbine is producing. A system that automatically logs data at set intervals would be ideal. The system currently employed is somewhat cumbersome when it comes to analyzing the data for trends in the system performance.