I.

ANALYSIS OF SYSTEM NEEDS

The target of this analysis is to find out the possible signals which can be replaced by fiber optical sensing. First of all, the technology aspect must be taken into account which signal can be replaced and further which signals make sense to be replaced in an economic way. A summary of the signals and the possibility of replacement is given in Table 1.

Table 1:

Possible signals which could be replaced by fiber-optical signaling

In addition also the number of signal lines is important to validate if a replacement makes sense or not. The rough number of signal lines for a telecommunication satellite is given in Table 2 taking into account only the signal types which can be replaced by fiber optical technology.

Table 2:

Number of signal lines for a telecommunication platform, estimation

II.

FUNCTIONAL REQUIREMENTS TO PANELS

The overall target of the panel demonstrator is to prove the concept of the integration of FBG sensors inside the panel for temperature monitoring and additional fibers for communication purposes.

The goal of the breadboard demonstrator is divided into different sections:

III.

COMPONENT SELECTION

For the project here some points where initially clear such as the usage of FBG sensors for temperature monitoring [1] and the need for a mechanical transducer for strain decoupling, see also section IV.B.

The identification of a suitable connector for the target application was not an easy task, here a compromise between cost and functionality must be made. On the one hand a space qualified connector is necessary for which until the current date no released specification is available, on the other hand a cost efficient multi-pole connector is necessary to lower integration time and cost per connection. For the demonstrator design the connectors from Glenair (GFR connectors) are selected, they are available in a DSUB 25 shape and offer up to eight single-mode connections. Some pictures of the used connector are given in Figure 1.

Figure 1:

GFR fiber-optical connectors from Glenair, pictures taken from datasheet.

The selected components for the project here are summarized in Table 3. The TRL levels are assumed due to the lack of qualified components or detailed test date. Nevertheless the components have been well analysed and cross checked if the can survive the thermal and mechanical loads for the demonstrator. Due to this reason also lots of commercial ‘cheaper’ components are not taken for the final design.

Table 3:

Selected components for the different subsystem functions.

IV.

FIBER-OPTICAL SENSING

A.

Temperature Sensing with FBG sensors

The selected technology for fiber-optical temperature sensing is based on so called Fiber-Bragg Gratings (FBG) sensors. The preferred wavelength of the gratings lies in the telecommunication C-band between 1528 nm and 1568 nm. In this wavelength range lots of commercial components are available not only for sensing but also for data transmission. For this wavelength also radiation hard fibers are available. The radiation induced loss (RIA) is a very important parameter for the use of fiber optics in space for long term mission such as for telecommunication satellites.

By knowing the Bragg wavelength, the temperature at the FBG can be determined. The Bragg wavelength is located where the highest reflected light intensity is assumed. Therefore a peak find algorithm, e.g. a fit function, is applied to the measured FBG spectrum. In general the reflected spectral response is used to evaluate the Bragg wavelength hence the spectrum shows a high intensity and so a high Signal-to-Noise (SNR) ratio in the area of the Bragg wavelength.

As mentioned before, the Bragg wavelength depends on the temperature at the FBG sensor. In Figure 2 on the left an FBG spectrum at 26 °C and 46 °C is measured by a interrogator system based on a tunable laser system. Hereby the spectrum moves toward higher wavelengths at ascending temperatures. In the case the characteristic tuning curve and the Bragg wavelength (λ_B1/λ_B2) of the FBG is known the FBGs temperature (T₁/T₂) can be estimated. The slope of the tuning curve depends on the FBG material and lies in general within a range of (10±4) pm/°C.

Figure 2:

Measurement of an FBG spectrum at two different temperatures (left) and a scheme of a characteristic Bragg Wavelength vs. Temperature tuning curve (right).

The effective refractive index and the grating period are affected by changes in temperature. The Bragg wavelength regarding to temperature changes ΔT can be written as:

Hereby it is assumed that no strain is applied to the grating. However in real FBGs it is difficult to decouple strain and temperature. To maintain strainless temperature measurements, the FBG is located within a mechanical package. The shift in Bragg wavelength is defined as follows (without strain):

Where α_Λ is the thermal expansion coefficient for the fiber and α_n is the thermo-optic coefficient. The thermooptic coefficient at room temperature for a fused silica fiber amounts to 8.5·10^-5 K^-1 and for the thermal expansion coefficient to 0.55·10^-5 K^-1. It is obvious that the refractive index change shows the strongest impact regarding to the Bragg wavelength shift.

B.

FBG Transducer Design

The aim of the transducer is the decoupling of strain and temperature which both affect the Bragg wavelength shift of the grating. In the project here temperatures only shall be measured, so the influence of the strain shall be cancelled out. This shall be done by an intelligent mechanical design. No electrical compensation signals or sensors shall be used. The measurement of FBGs in polarization maintaining fibers is noted here for reference only [2] but it is not considered as option due to the needed complex interrogation system. Basically four different topologies are possible which are described in more detail in the following section:

• • Pure grating glued to the panel

• • Feedthrough based transducer

• • Capillary based transducer

The different transducer topologies are graphically illustrated in Figure 3.

Figure 3:

Left: pure FBG glued onto panel plate. Middle: Feedthrough based transducer. Right: Capillary based transducer.

The pure FBG sensor glued onto the plate needs the smallest amount of integration time whereas the feedthrough transducer needs more time due to the different pre-assembly steps needed for this option. The capillary based transducer, shown on the right side of Figure 3, is a more experimental solution. The idea was to combine the effects from low integration time of the pure FBG sensor with the strain decoupling of the feedthrough based solution.

A four point bending test was done with the designed feedthrough transducer (red curve, rectangular markers) and the pure sensor (blue curve, karo markers) glued to the aluminum plate, the results are given in Figure 4. As can be seen the transducer is necessary to compensate for the strain-induced Bragg wavelength shift. The transducer has a sensitivity to plate deflections of 0.8 pm/mm, in reality deflections larger than 1mm will not occur for honeycomb satellite panels.

Figure 4:

Measurement results of the four-point bending test for the pure FBG glued onto the plate (blue curve) and for the FBG mounted inside the special designed transducer (red curve).

V.

DEMONSTRATOR DESIGN

After the functional requirements for the panel were identified, the possible components were selected a concept was established. The concept here foresees the setup of three different panels, a large one (900 mm x 800 mm x 25 mm) and two smaller ones (280 mm x 280 mm x 25mm). The interconnection scheme is illustrated in Figure 5. Each panel holds at least two strings equipped with FBG sensors (blue channels) and two strings for data communication (yellow channels).

Figure 5:

Illustration of the planned harness shown for the panel internal wiring and the patchcord connections.

For the satellite design a maximum sensor density of 41 sensors/m² resulted, so for the demonstrator the small panels are equipped with up to 10 sensors resulting in a total density of 127 sensors/m² which is much larger than in reality needed. However the aim of the project is here to demonstrate the possibilities of the technology. For the large panel 20 FBG sensors are foreseen, 10 sensors for each string. So in total the demonstrator is able to implement 40 FBG sensors for temperature monitoring.

As an additional demonstration also the interconnection between different panels with special manufactured patchcords is shown. For the demonstrator patchcord lengths of 0.5m are defined, the fiber itself is a single mode fiber protected with LCP buffer with an outer diameter of 1.65mm.

The 3D CAD model of the panel demonstrator is illustrated in Figure 6. The two smaller panels are mounted onto the large panel in an angle of 90° and fixed together with L-profiles. The fiber-optical connectors are mounted with an angle of 45° with respect to the panels, the corresponding plugs are connected afterwards. Also indicated in the 3D model is the internal routing of the fibers and the dummy heaters. The dummy heaters on one side of the panels are comparable in dimensions with the traveling wave tube amplifiers (TWTA) used in telecommunication satellites.

Figure 6:

CAD model of the panel demonstrator, illustrating the L-profiles for mounting the panels, the three different panels and the interconnections (schematically).

VI.

ESTIMATED SAVINGS IN TIME AND MASS

For the estimation of the mass and time savings first some constrains must be set. Here a telecommunication platform was analysed with the numbers of interfaces as given in Table 2. The time needed for each sensor is determined by the use of data from the OHB manufacturing personal and includes time for documentation and time for hardening of the glue. For the FBG sensor it is assumed that the sensor comes pre-assembled to the AIT personal, the time for manufacturing a single FBG transducer was tested to be approx. 15min. The final overview of the mass and time savings is illustrated in Table 4.

Table 4:

Mass and time effort for the classical electrical system and for a replacement with fiber-optical systems.

For the HPC system in the current architecture a lot of harness mass could be saves. This implies some modifications in the units, because of a change in physical signal property from current/voltage to light. The additional advantage which comes with the fiber-optical HPC system is the inherent galvanic isolation by the fibers.