Tools › Structures › Loads and Dynamics › Mass Acceleration Curve (MAC) Loads › Method

Mass Acceleration Curve (MAC) Loads: Method

Open the Mass Acceleration Curve (MAC) Loads tool

Mass Acceleration Curve Loads: Method, Sources, Assumptions and Limits

What the tool does

The tool gives preliminary design limit-load factors for spacecraft equipment and secondary structure before a coupled loads analysis exists. For each item in a table it reads a load factor from a mass acceleration curve at the item's mass, estimates the random-vibration load factor with Miles' equation, and compares or combines the two by the rule chosen. It lists the governing load factor and the design load (load factor times weight) of every item and draws the curve with the items on it. The results are limit-load factors: the project's factors of safety are applied afterwards.

The mass acceleration curve

A mass acceleration curve (MAC) plots a bound of acceleration in g against mass on log-log axes. JPL D-5882 (Trubert, 1989), the report that documents the method as developed at the Jet Propulsion Laboratory, puts it this way: the accelerations of the physical masses of a spacecraft are bounded by a curve, the lighter the mass the higher the acceleration, the concept is semi-empirical, and the curve is a bound and not a simulation of the actual response. The report distinguishes two curves.

The physical curve is entered with the mass of a hardware item. The report (section 5) limits it strictly: one lumped mass for the subsystem, a mass below 500 kg, an appendage-like structure with a statically determinate attachment to the spacecraft core or to a larger subsystem, and use for preliminary design when no finite element model exists. It gives the acceleration bound of the lumped mass in each of the three directions. The steady thrust acceleration is not in the curve: the report adds it in the axial direction (in its example a 100 kg instrument on Titan 4 / IUS takes 15.0 g laterally and 15.0 + 3.9 = 18.9 g axially). One static load case is run for each direction and the largest load is the preliminary design load of the support; the report says the three cases generally need not be combined, because the curve keeps the largest acceleration of any direction, and that combining them only adds conservatism.

The modal curve (the report's MDMAC) is lower and is entered with the effective mass of each mode of a finite element model of the payload; the model may be crude, but its modes should account for about 90 % of the rigid mass. The curve gives the purely dynamic acceleration of each mode, and the report combines the modes by root-sum-square with the interface accelerations and adds the thrust separately (its equations 11 and 12). This tool reads either kind of curve at the masses in the table; it does not carry out the modal recovery of member loads.

The report's curves envelop the bounds found for Galileo effective masses from liftoff, stage 1 depletion and an equivalent random vibration analysis (its Figure 7). JPL Preferred Reliability Practice PD-ED-1211 says the same of its typical curve: it is limited to appendage masses up to 500 kg with frequencies up to about 100 Hz and already combines transient and mechanically transmitted random vibration. NASA-STD-5002B requires preliminary design load factors for components to come from a mass acceleration curve or a table (section 4.3.2), describes the curve as a combined factor for the low-frequency and random vibration loads, and says one curve can usually serve a launch vehicle configuration and a broad class of payloads (Appendix C.5). NASA-HDBK-7005 (section 8.1.2.2) adds that a curve can be sensitive to the configuration.

The curve as an equation

JPL D-5882 began from Bamford's original curve A = K / √m and replaced it with a form that carries the damping (its equations 2 and 3):

A = G0 / √[m/M + (ξs + ξl)²] × e−α / tan α, with α = tan−1[√(m/M) / (ξs + ξl)]

Here m is the mass, ξs and ξl are the damping ratios of a spacecraft mode and a launch vehicle mode, and G0 and M are coefficients chosen so that the curve envelops the data. The form comes from a two-mass model of a spacecraft mode tuned to a launch vehicle mode under an impulse; for a vanishing mass it tends to G0 / [e (ξs + ξl)]. The report says any curve that envelops the data would do; the equation is a convenience.

Built-in curves

STS / IUS modal (JPL D-5882 Figure 1): G0 = 1.60 g, M = 6,000 kg, damping 0.010 + 0.010, as printed on page 10 of the report. Titan 4 / IUS modal (Figure 2): G0 = 1.45 g, M = 12,000 kg, damping 0.010 + 0.010. Both are computed from the equation; readings of the printed curves at their grid lines agree within 2 %.

STS / IUS physical (Figure 11): the report prints the parameters only in the figure heading, G0 = 3.98 and SW = 2553. The tool reads SW as the mass coefficient in pounds (1,158 kg), as it is in the modal figure, with damping 0.010 + 0.010; the equation then agrees with readings of the printed curve within 2 % from 1 to 400 kg and within 6 % at its 500 kg end. That reading of the heading is this tool's, checked against the chart.

Titan 4 / IUS physical (Figure 12): the report prints no complete parameter set, so the curve is a set of readings of the printed chart at its grid lines, about ±3 %. The report itself takes 15.0 g at 100 kg from this figure; the reading here is 14.6 g.

PD-ED-1211, Figure 1, "Typical Mass Acceleration Curve": readings of the printed chart, about ±3 %. The practice names no launch vehicle. The readings lie 1 to 7 % below the STS / IUS physical curve up to 500 kg, more than the reading accuracy at 2 to 5 kg, so it is a similar curve and not the same one. NASA-HDBK-7005 Figure 8.2 reprints the STS / IUS physical and modal curves of Galileo.

Each physical and modal curve is supplied as breakpoints close enough that the straight-line interpolation stays within 0.5 % of the equation. All of these curves were derived for the Galileo spacecraft on launch vehicles that no longer fly. They are documented examples. For a project, enter the curve issued by the launch vehicle organization or the project loads document.

Interpolation

Straight lines on log-log axes join the breakpoints. Between (m1, N1) and (m2, N2):

N = N1 (m / m1)b, with b = ln(N2 / N1) / ln(m2 / m1).

Outside the curve the tool flags the item. A mass lighter than the first breakpoint follows the first segment's line beyond the end of the curve, which gives a higher factor when the curve falls; a mass heavier than the last breakpoint keeps the last load factor. Both rules are choices of this tool, made to stay on the high side; neither is in the sources.

Random-vibration load factor

Miles' equation gives the rms acceleration response of a single-degree-of-freedom system of natural frequency fn and amplification Q = 1/(2ζ) to a base acceleration PSD W, taken at fn (NASA-HDBK-7005 eq. 5.27):

Grms = √[(π/2) fn Q W]

The random-vibration limit load factor is R = k Grms. NASA-STD-5002B (Appendix C.4.3) says random vibration limit loads are typically taken as the 3-sigma load, or are computed to the required statistical enclosure level, so k = 3 is the opening value and may be changed. The standard also says that Miles' equation is a conservative estimate when the item can be represented as a single-degree-of-freedom oscillator and the maximum of the input PSD is at its natural frequency. An item with a blank or zero PSD level has no random-vibration load in this tool.

Combination

Four choices of governing factor are offered.

S + MAC (opening choice). The method of JPL D-5882 section 5 for a physical curve: the curve value in each direction, plus the steady thrust acceleration S in the axial direction (1.5 g for STS / IUS and 3.9 g for Titan 4 / IUS in the report's Tables 2 and 1; zero laterally). The curve is not combined with the Miles random factor, because the curve is built to envelop the transient and the equivalent random vibration. The Miles factor is listed for comparison.

Larger of S + MAC and S + √(L² + R²). GSFC-STD-7000B (section 2.4.1) combines the launch loads of an axis as the steady-state load factor plus the root-sum-square of the low-frequency dynamic factor L and the high-frequency random factor R, N = S ± √(L² + R²), and says that where the steady-state and low-frequency factors are given as one, the steady-state value must be separated out first. NASA-STD-5002B (4.8.5, 4.8.6) accepts the root-sum-square of the maximum low-frequency and maximum random vibration load factors for liftoff and applies the root-sum-squared values in all axes simultaneously; NASA-HDBK-7005 (eq. 5.34) gives the same root-sum-square. Taking the larger of this and the curve value is a choice of this tool.

S + √(L² + R²) alone. For a project without a curve.

S + √(MAC² + R²). The curve is read as the low-frequency dynamic factor only. This is a choice of this tool, for a project curve that is stated to cover the low-frequency transient alone; with a curve that already includes random vibration it counts that load twice.

The tool computes one axis at a time with the plus sign. PD-ED-1211 adds that a non-negligible direct acoustic load on an item may be combined with the curve load by root-sum-square; the tool does not compute acoustic loads.

Design load

P = N m g0, with g0 = 9.80665 m/s². JPL D-5882 computes its example with 9.807 m/s². JPL D-5882 and PD-ED-1211 limit a physical curve to masses below 500 kg, and NASA-STD-5002B (4.8.5) lets the random-vibration load of a component heavier than 500 kg be neglected; the tool notes items above 500 kg and leaves the decision to the analyst.

What it assumes

The curve applies to the launch vehicle, the spacecraft class and the kind of mass (physical or modal) of the items entered. Each item is a rigid mass on one spring, with one mode in the axis considered, on a mounting that is not affected by the item. The PSD is flat near the natural frequency and the response is Gaussian and stationary. The low-frequency and random maxima are independent, so that the root-sum-square applies. One axis is treated at a time.

When it misleads

A curve taken from another launch vehicle or spacecraft can be wrong in either direction: NASA-HDBK-7005 warns that the curve can be configuration sensitive. The built-in curves were derived for Galileo on the Shuttle or Titan 4 with the Inertial Upper Stage, not for any current vehicle. A physical curve used for an item that is not a single lumped mass on a statically determinate, appendage-like support is outside its stated limits; JPL D-5882 then requires a finite element model and the modal curve. Miles' equation takes no credit for the item loading its support (the impedance effect), which NASA-STD-5002B names as a conservatism of base-drive analysis, most pronounced for modes of high effective mass; for heavy items it can exceed the true load several times over. It is low when the PSD rises away from the natural frequency or when several modes respond. A PSD specified for a test (a qualification level) is not the flight environment at the time of the peak transient. Items with a natural frequency inside the range of the coupled loads analysis respond to the launch transient in ways no curve bounds reliably. The numbers here are for first sizing; the coupled loads analysis and the project loads document supersede them.

Verification

The equation is checked against the report's own numbers: the modal force of 6.10 × 104 N for an 1,800 kg effective mass on the Titan 4 / IUS modal curve (JPL D-5882 Table 7, page 24) and the inertia loads 14,706 N and 18,535 N of its 100 kg instrument example (page 31; the first of these corresponds to 9.804 rather than the 9.807 m/s² the report states, a rounding in the source). The equation curves are checked against readings of the printed figures. The interpolation, Miles' equation and the combinations are checked against closed forms and an independent numerical computation.