Вы находитесь на странице: 1из 13

Burgin Halim Quesnelle 1 Introduction Gravity is a very powerful force. On Earth, it accelerates at 9.8 m/s2.

2. Earths gravity is what causes everything to remain on the ground. Gravity is also what causes various objects, such as satellites and the moon, to remain in orbit around the Earth. Using the mass of these satellites and their average distance from the Earth, it could be determined how much acceleration due to gravity exists on them. If there appeared to be a linear relationship, it would be possible to run a linear regression test. This test would be run twice, once not including the moon and Galaxy 15 and once including the moon and Galaxy 15. This would be done to see if very massive and far away satellites would have an effect on the data. Data Table 1 Satellites and the Data Associated With Them
Satellite ISS MIR Hubble Skylab Sputnik Salyut 2 Salyut 5 NAVID RISAT-2 MOST Moon Galaxy 15 Mass (Kg) 450000 129700 11110 169950 83.6 267500 246000 50 300 53 7.3477E+22 2033000 Average Distance (m) 408000 364000 569000 272150 577000 267500 246000 312.5 550 412.75 383990000 35771500 Actual Radius (m) 6779000.00 6735000.00 6940000.00 6643150.00 6948000.00 6638500.00 6617000.00 6371312.50 6371550.00 6371412.75 390361000 42142500 Force of Gravity (N) 3901865.53 1139346.51 91914.95 1534490.14 690.05 2418660.54 2238741.19 490.80 2944.57 520.23 1.92136E+20 456129.16 Acceleration due to Gravity (m/s^2) 8.67 8.78 8.27 9.03 8.25 9.04 9.10 9.82 9.82 9.82 0.00 0.22 Percent of gravity on Earth (%) 88.48 89.64 84.42 92.13 84.23 92.26 92.86 100.16 100.16 100.16 0.03 2.29

Table 1 above gives a list of the satellites used to in calculations. All of these satellites are artificial with the exceptions of the moon and Galaxy 15, which are natural and heavier and farther away than any other satellite listed. The average distance of the satellite away from Earth is found by adding the apogee, or closest point, and perigee, or farthest point, and dividing the sum by two to find the average. The actual radius column used the average distance column and

Burgin Halim Quesnelle 2 added the radius of the Earth (6,371 kilometers) to find the distance the satellites were from the center of the Earth. A sample equation of how to determine the force of gravity is shown in Figure 1.

Figure 1. Example of How to Calculate Force of Gravity In order to find the acceleration due to gravity, the force of gravity was divided by the mass of the satellite. An example of this is shown in Figure 2.

Figure 2.Example of How to Calculate Acceleration Due to Gravity The percentage of gravity on the satellite as compared to Earth was then calculated by taking the acceleration due to gravity calculated, dividing it by 9.8, or the acceleration of gravity on earth, and multiplying it by 100 to get a percentage. An example is shown in Figure 3.

Figure 3. Example of How to Calculate Percent of Gravity on a Satellite

Burgin Halim Quesnelle 3

Figure 4. Distance vs. Acceleration Due to Gravity without Moon or Galaxy 15 Figure 4 above shows a graph showing distance vs. acceleration due to gravity not taking into account the moon or Galaxy 15. The data appears to have a strong linear relationship with a negative slope. Using technology, it was possible to find the correlation coefficient, r, and the coefficient of determination, r2. The correlation coefficient was -0.999 and the coefficient of determination was 0.998. The large r value is more evidence to suggest that there is a strong negative linear relationship between distance and acceleration due to gravity. The r2 value of 0.998 shows that 99.8% of the variation in acceleration due to gravity (y) is explained by least squares regression of acceleration due to gravity (y) on distance (x). Because the data appears to be linear, it would be appropriate to run a linear regression test to determine if there is truly a correlation between distance and acceleration due to gravity.

Burgin Halim Quesnelle 4

Figure 5. Distance vs. Acceleration Due to Gravity without Moon or Galaxy 15 Regression Because the data appeared to be linear, it was appropriate to create a linear regression line of best fit. The regression line shown in Figure 5 in = a + bx form is = 9.7986-0.000003x. This indicates that as the distance from Earth increases by one meter, the acceleration due to gravity decreases by 0.000003m/s2. This regression line appears to fit the data well, giving further evidence that linear regression will be a safe test to use. Before carrying out the linear regression test to see if there really is a correlation between distance and acceleration due to gravity, the residuals must be plotted to check for patterns that would imply a non-linear relationship in the data.

Figure 6. Plot of Residuals not Including the Moon or Galaxy 15

Burgin Halim Quesnelle 5 Figure 6 shows a plot of residuals found when the moon and Galaxy 15 were not included in the data. This residual plot corresponds with the scatter plot from Figure 5. There may be a slight pattern to the residuals, which may imply that the data is not really linear. However, the number of data points is very small, so they may not be an accurate representation of the relationship of the data. Despite this possible pattern, the regression line in Figure 5 fits very well, so the linear regression test will still be run.

Figure 7. Histogram of Residuals Calculated without the Moon or Galaxy 15 Figure 7 is a histogram used to determine if the response variable, acceleration due to gravity, is normally distributed. This histogram does not show normal distribution, but there also appears to be relatively low skewness. Because of this, it may suggest a nonlinear relationship, but the test will be carried out anyway because the regression line was a good fit.

Burgin Halim Quesnelle 6 Statistics Hypotheses: Ho: = 0 Ha: < 0 The null hypothesis states that the standardized slope, , is equal to zero. This means that there is no correlation. The alternative hypothesis stating that is less than zero means that there is in fact a correlation between distance and acceleration due to gravity and it has a negative slope. Assumptions: For any fixed value x, the response y varies according to normal distribution Repeated responses of y are independent of each other The mean response has a linear relationship with x. and are unknown parameters The standard deviation of y is the same for all values of x

Figure 8. Simple Equation to Find Test Statistic Figure 8 shows a simple equation to find the t-statistic. The t-statistic is a standardized form of b, the least squares regression slope. The t-statistic depends on the found value of b, as well as standard error about the slope. Figure 9. Equation Used to Find Standard Error About the Slope

Burgin Halim Quesnelle 7 Figure 9 above shows the equation used to find the standard error about the slope of the least squares regression line. The standard error is found using the sample standard deviation s, which is found in the numerator, and dividing it by the square root of the sum of residuals squared.

Figure 10. Equation Used to Find Test Statistic After finding the standard error, it was substituted into the equation from Figure 8. This is shown above in Figure 10. In order to calculate the t-statistic using this formula, the sum of residuals had to be found. Each individual residual was calculated by subtracting the predicted value from the observed y value. An example of this is shown below.

Figure 11. Calculation of an Individual Residual

Burgin Halim Quesnelle 8 After each residual was calculated, it was squared, because the equation uses the sum of the squared residuals. A table of all calculated residuals and squared residuals is shown in Figure 12.

Figure 12. Table of Residuals and Squared Residuals Results from the residuals squared column were then added together. This value was found to be 0.00578. Figure 13. Calculating the Standard Error About the Slope Now that the standard error had been calculated, it could be used to find the test statistic. The value of 4.09843*10-8 gets plugged in to the equation to find t. This is shown below.

Figure 14. Calculating the Test Statistic

Burgin Halim Quesnelle 9 After carrying out this linear regression test shown in Figure 14, the test statistic was found to be -66.9718. This t-statistic had a corresponding P-value of 1.374*10-12. This value was found using technology. Using this P-value, the null hypothesis was rejected at the 0.05 significance level. There is strong evidence to suggest that there is a negative linear relationship between the distance away from Earth and the acceleration due to gravity that a satellite experiences. The P-value of 1.374*10-12 shows that there is virtually a 0% chance of receiving results this extreme by chance alone if the null hypothesis is assumed to be true. The next step was to compute a confidence interval for the given data. A confidence interval can be calculated using the formula shown below in Figure 15. Figure 15. Equation Used in Finding Confidence Intervals. Using the equation shown in Figure 15 and technology, a 95% confidence interval was computed. The lower value of this interval was found to be -2.83933*10-6 and the upper value of this interval was found to be -2.65032*10-6. This means that there is 95% confidence that the true mean slope would fall between the values of -2.83933*10-6 and -2.65032*10-6, meaning that as the distance away from the center of Earth increases by one meter, acceleration due to gravity would decrease by a value within this interval. After it had been determined that the relationship between distance and acceleration due to gravity on all of the artificial satellites was in fact linear, the natural satellites could be put back in to calculations to see if huge, distant satellites would have an effect on the results. In order to see if this data would be linear, the two extra points were added back in and a new scatter plot was made.

Burgin Halim Quesnelle 10

Figure 16. Distance vs. Acceleration with Moon and Galaxy 15 This new scatterplot shown in Figure 16 was created after the data points corresponding to the moon and Galaxy 15 were added back in. This graph has a very large scale to account for the moon, which is the farthest point to the right. Galaxy 15, the point closest to the origin, appears to be far removed from any other data points. To see if this graph is linear, a regression line was drawn. This graph is shown in Figure 17 below. The correlation coefficient, r, and the coefficient of determination, r2, were found using technology. The correlation coefficient was -0.74 and the coefficient of determination was 0.547. The r value provides evidence to suggest that there could possibly be a linear relationship between distance and acceleration due to gravity. The r2 value of 0.547 shows that 54.7% of the variation in acceleration due to gravity (y) is explained by least squares regression of acceleration due to gravity (y) on distance (x).

Burgin Halim Quesnelle 11

Figure 17. Distance vs. Acceleration with Moon and Galaxy 15 Regression Figure 17 above shows a regression line drawn for the new set of data. The regression line shown in Figure 17 in = a + bx form is = 8.40149-2.36469*10-8x. This indicates that as the distance from Earth increases by one meter, the acceleration due to gravity decreases by 2.36469*10-8m/s2. This regression line appears to fit the majority of the data relatively well, but Galaxy 15 appears to be an outlier. This could suggest that the data is not totally linear. For more information used in determining if the data truly has a linear relationship, a new residual plot was created that included the points of the moon and Galaxy 15. If the residual plot did not appear to have a pattern, it may suggest that there is a linear correlation in the data.

Figure 18. Residual Plot Containing the Moon and Galaxy 15

Burgin Halim Quesnelle 12 Figure 18 above shows a residual plot corresponding to the scatterplot in Figure 17 which uses data containing the moon and Galaxy 15. There does not appear to be a pattern in the data. However, there is a cluster of data points about zero which include all of the artificial satellites. Because there are so few data points, this set of data may not be an accurate representation of the correlation between variables.

Figure 19. Histogram of Residuals Calculated with the Moon and Galaxy 15 Figure 19 is a histogram used to determine if the response variable, acceleration due to gravity, is normally distributed when far away satellites are included. This histogram does not show normal distribution, and there is skewness to the left. Because of this, it may suggest a nonlinear relationship, but the test will be carried out anyway in order to compare this set of data with the data that did not include the natural satellites. To compare results of the first test to the results found by adding the moon and Galaxy 15 back in, the same statistical formulas for linear regression were used. The formula and an example problem showing how to find the t-statistic for this test is shown in Figure 14. Once the test including the natural satellites had been carried out, the new t-statistic was found to be -3.48, drastically different than the first test. The corresponding P-value in this situation is 0.005917. Using this P-value, the null hypothesis was rejected at the 0.05 significance level. There is strong evidence to suggest that there is a negative linear relationship between the distance away from

Burgin Halim Quesnelle 13 Earth and the acceleration due to gravity that a satellite experiences, including the large natural satellites. The P-value of 0.005917 shows that there is about a 0.5917% chance of receiving results this extreme by chance alone if the null hypothesis is assumed to be true. The next step was to compute a confidence interval for the given data. A confidence interval can be calculated using the formula shown above in Figure 15. Using the equation shown in Figure 15 and technology, a 95% confidence interval was computed using all data points. The lower value of this interval was found to be -3.8789*10-8 and the upper value of this interval was found to be -8.5042*10-9. This means that there is 95% confidence that the true mean slope would fall between the values of -3.8789*10-8 and -8.5042*10-9 meaning that as the distance away from the center of Earth increases by one meter, acceleration due to gravity would decrease by a value within this interval. Conclusion After all calculations had been made, it was determined that there is in fact gravity in space. Although adding in the points of the Moon and Galaxy 15 changed the slope and test statistics, it still shows that there is a linear relationship between distance from the center of Earth and acceleration due to gravity. The graphs, data, residuals, and histograms all provide further evidence to explain this relationship. And finally, the answer to the ultimate question: gravity does exist in outer space, but it accelerates less than it does on Earth, which causes objects to float on the ISS.

Вам также может понравиться