5.4 Auto Balance Results

Buoyancy offset can degrade vehicle performance. This is caused by a mismatch between the positions of the vehicle's center of buoyancy and its center of mass. These may be aligned by using the RBCs to shift the position of the center of mass. Manually balancing Ranger NBV for each test produces varying results. The automatic balancing algorithm helps to increase the repeatability of balancing accuracy between each test. The results of testing the automatic balancing algorithm from Section 4.8 are outlined in this subsection.
The algorithm works by averaging the commanded moment for 5 seconds. This average is assumed to be produced by buoyancy offset. The weighs in the appropriate RBCs are shifted proportionally to the commanded moment (up to a maximum per-step limit), and the cycle is repeated.
This test examined balancing behavior in the x-y plane. The vehicle was given initial buoyancy offset along the x-axis. The test was started with the vehicle holding attitude with the z-axis pointing down. The high gain PD controller was used for this test, however since the algorithm averages the moment over 5 seconds, the results of various controllers should produce similar results as this method isolates the bias moment.

Figure 5-25 Video control consoles in the control room

Initially, the buoyancy offset produces a 2.8° steady state error. When the auto balance algorithm is started, the weights move toward the neutral position with a response time of about 20 seconds and a settling time of about 80 seconds. The total shift to balance the vehicle was 9 inches along the x-axis, and about 1 inch along the y-axis. The x-RBC weight position overshoots by about 1.5 inches, or about 17%. The response of the Y-RBC is similar but smaller. Examination of the raw data stream from the vehicle reveals that initially the vehicle had to run the thrusters at an average of 47.5% of maximum velocity in order to hold attitude against the buoyancy offset. After balancing, the average thruster speed was reduced to about 8% of maximum.
It is an interesting exercise to compare the change in moment produced by the shift in RBC position with the initial average moment required by the vehicle required to counteract it. Initially, the thrusters required 47.5 % of the maximum velocity to counteract the buoyancy moment. From the thruster performance data in Figure 2-9, we see that this velocity produces a thrust of 1.4 lbf. By taking into account the position of the thrusters on the vehicle, and noticing that four thrusters contribute to the pitch moment, it can be shown that the commanded moment is about 84 inch-lb. The weight in the one x-RBC used in this test shifted a total of 9 inches. To produce an 84 inch-lb. moment on a 9 inch lever arm, the x RBC weight would have to be approximately 9.3 lb. Table 2-11 shows that the actual weight in each x-RBC is 10 lb. This level of correlation when taking into account the models and performance of several different vehicle systems is one example of the level of accuracy to which the vehicle physical and dynamic properties have been determined.