Monitoring in situ in real time of resin infusion for thermoset composite structures

The presented work investigates changes in electrical resistance of embedded sensory yarns as a method to monitor the resin flow front position and curing degree of resin during manufacturing of composite structures by vacuum infusion technology. The sensor concept is based on Piezo-resistive sensors integrated to the flax fabric, having almost identical propriety and dimensions as the flax threads used for the production of reinforcements. In the first time sensors have been characterized and first measures of the resin infusion have been realized in order to demonstrate the feasibility of the proposed approach. Then, the measures in real time were realized with fibrous sensors added to the flax fabric (green composites) to monitor the flow front of resin. A large amount of data recorded, filtered, examined, analysed and processed in order to understand and to optimize the infusion and polymerization process.


Introduction
The composites are more and more used in various fields such as aerospace, aeronautics, defense, sport and biomechanics. They are used due to their remarkable mechanical properties such as: density, rigidity mechanical resistances, lightweight, but also lifetime duration, impenetrability, electrical insulation, resistance against shocks [1] etc. A robust and reliable infusion process is vital for the quality of composite parts. Therefore, monitoring of resin flow and cure development in liquid composite molding processes was the subject of extensive researches in the past. A variety of sensor concepts based on thermocouples [2], pressure transducers [3], optical fibers [4][5], and electric resistance sensors [6][7] was investigated. In situ monitoring in real time of the resin diffusion inside the composite is the main objective of this article. The understanding of resin behavior during the infusion process of composite materials in press molding is a major concern, enabling their optimal mechanical properties.
A disadvantage of methods using vision systems or camera is the impossibility to see inside composites materials and the difficulty of adapting this process to industrial production. The use of thermocouples or optical fibres within composites could also damage their structures.

Measuring system
In situ monitoring of the resin infusion has been realized using piezo-resistive sensors embedded to composite reinforcement. A protocol has been set up in order to follow the resin movement supposed to modify sensors electrical resistance. To achieve this objective, conductive multi filament yarns (polyamide silver coated) were sewn onto flax fabric used as composite reinforcement, with electrical resistance ranging from 30 to 100 Ω/m. The flax fabric used as reinforcement is twill 2/2 with the reference WD55018113, 13X13 105 DTEX (305 g/m²). The stitching of sensory yarns has been realized on the X axis (same direction as the resin infusion), with a distance of 5 cm between two consecutive sensors. The flax fabric with high damping factor [8] is used in regards the specification of "green composite" with the specific rigidity superior compared to those of glass fibres ( Figure 1). For the electrical resistance measurements, Keithley 3706A was used, able of over 14 000 measurements per second, storing data to the memory with optional high performance resolution. Camera has been used to follow from outside the resin advance. Measurements in real time were realized by sensory yarns [9] embedded to the flax based reinforcements of the composite. Methods for production, placement, and calibration of the sensors adapted to our measurement protocol have been developed and data filtering algorithms set up. Moreover, experiments and manipulations have been established in order to characterize and verify the sensitivity, the rapidity, the response time and repeatability of measures of our sensors.

Result and Discussion
The computation of the relative resistance variations is realized using the following expression: (1) First experiment of resin flow monitoring has been conducted in our lab in order to verify the feasibility of the concept and to test fibrous sensors embedded to the flax fabric in the direction of the resin advance. Figure 2 shows the outputs of 5 embedded sensors. It could be considered that the feasibility has been proven and that the selected polyamide silver coated yarns composed of a number of multifilaments may be used as sensors for detection of resin flow. As the resin used in infusion process is non-conductive, it reduces the conductive surface of the yarn (multifilaments) and implies an increase of its electrical resistance. It is possible to observe in Fig. 2 the increase of the sensory yarns resistances (5 sensors) and to conclude consequently, that the resin has progressively covered all the sensory yarns. The infusion process has been conducted over 10 minutes approximately and during this time the electrical resistance of sensors increased. It may also be observed that the increase in the electrical resistance is faster in the beginning of the infusion process and that it slowed down after approximately 5 minutes. It means that the resin front flow was faster in the beginning of the infusion that has also been observed visually during the experiment and that is quite normal for this kind of processes. It could also be noticed that the fifth sensor (violet colour) generated lower output that other sensors because it has been located close to the border of the flax fabric and the resin flow was delayed at that position comparing to other sensory yarns. A decrease of electrical resistance in the beginning of the infusion process for this sensor may be explained by the change in the pressure of plastic sheets because the resin has arrived close to this sensor, but it has not yet been "coated" by the nonconductive resin.

Sensors' Characteristics
In this chapter, static and dynamic characteristics of sensory yarns will be presented. Their importance will be highlighted and their influence on the operation of sensing systems will be described.

Measurement Range
The maximum and minimum values of the measurand that can be measured with a sensing system are called the measurement range, also defined as the dynamic range or span. This range results in a meaningful and accurate output for the sensing system.
All sensing systems are designed to perform over a specified range. Signals outside of this range may be unintelligible, causing unacceptably large inaccuracies, and may even result in irreversible damage to the sensor [10].

Experiment
A protocol has been set up in order to check the measurement range of sensors. The conductive multi filament yarns (polyamide silver coated) were sewn onto flax fabric used as composite reinforcement, with electrical resistance ranging from 30 to 100 Ω/m. The sensor has been located in the middle of the reinforcement. The resin used, is Epoxy SR 8200 mixed with the epoxy hardener SD7201 (37 g of hardener for 100 g of resin). The density ⍴ = 1.2 g/cm3 and the volume part VResin = 0.6 VTot.
A multimeter was used to measure the electrical resistance before the beginning of infusion process, then the pump was switched on and the infusion started. The process of liquid resin infusion lasted 2 minutes and 23 seconds. The dimensions of the sample were: 20×17cm², and the sensory yarn length: 17 cm. Measures of electrical resistance were realized with the same multimeter at different time intervals. The following table summarizes the evolution of electrical resistance as a function of time.

Results
According to the results obtained in our tests of infusion phase, it may be deduced that the range of sensor varies between 15.21 Ω (0 mm of resin) and 18.51 Ω (150 mm of polymerized resin).

Sensitivity
The sensitivity of the sensor is defined as the minimum input of physical parameter that will create a detectable output change. In some sensors, the sensitivity is defined as the input parameter change required producing a standardized output change. In other terms, it is defined as an output change for a given change in input parameter [11].

Experiment
In order to check the sensitivity of our sensors, the following manipulation has been set up: On a reinforcement made of one-layer flax fabric with dimensions (20×20 cm²), using the sewing machine, four polyamide silver coated conductor yarns have been integrated, with electrical resistance ranging from 30 to 100 Ω/m. A distance of 5 cm between the yarns has been left. The resin used in this process was Epoxy SR 8200 mixed with the epoxy hardener SD7201. For the electrical resistance measurement, ohmmeter was used. 1 ml of resin on each sensor was dropped using a pipette, and the dose was increased progressively (1 ml by 1 ml) to reach the value of 10 ml. A time interval of 90 seconds has been left between the drops, 1 ml each.

Results
In the previous section, it is mentioned that the sensitivity is defined as the ratio between the output signal and measured property. The following graph represents the change of electrical resistance as a function of time, the dose used and the covered distance by resin on the flax one layer reinforcement.

Repeatability
This is the ability of a sensor to repeat a measurement when put back in the same environment. It is often directly related to accuracy, but a sensor can be inaccurate, yet be repeatable in making observations [12].

Experiment
Flax fabric twill 2/2 was applied with dimensions of (20×17) cm². Two sensors were integrated to the flax fabric with a distance of 7 cm between the yarns. The objective of this trial is to check that the two sensors give the same responses with regard to the repeatability.
For this experiment two ohmmeters have been used, the process of liquid resin infusion lasted 2 minutes and 30 seconds. The electrical resistance is measured for 30 minutes. The following table summarizes the evolution of electrical resistance for the first and the second sensory wire as a function of time.

Results
As stated before, repeatability is defined as the range of actual positions the system takes while being repeatedly commanded to the same location under identical conditions (The same dimension and weight of flax fabric, the same the sensory yarn length, the same pumping speed, the same resin and the same temperature of the test room). Even if the LRI (Liquid Resin Infusion) process is not uniform at sensory yarns 1 and 2 location, it is considered that it has been identical. From the data obtained from the first and the second sensors, it may be deduced that the output of two sensors is almost the same.

Response and Recovery Time
Generally, Recovery Time is defined as the time for a sensor to return to baseline value after the step removal of the measured variable. Usually specified as time to fall to 10% of final value after step moval of measured variable. In real-time systems the response time of a task or thread is defined as the time elapsed between the dispatch (time when task is ready to execute) to the time when it finishes its job (one dispatch).

Experiment
As it was impossible to measure the recovery time of our sensors, because it is impossible to remove the resin from composite once it has been infused and polymerized, only a response time has been evaluated. To do that following trial was conducted: For this manipulation the same flax fabric and sensory yarns has been used with dimensions of 5×17cm². The length of our flax fabric was reduced to 5 cm to ensure that the resin arrives and coats the sensor all over 5 cm in the same time. It means that the step input (Heaviside function) has been applied to three sensors at different times, scheduled by the velocity of the resin flow. In this way three different step responses have been obtained from the infusion process applied on three sensors for statistics purposes.
The stitching has been realized on the X axis (parallel to the direction of the resin infusion in order to generate a step response), and a distance of 1 cm has been set up among the sensory yarns. To observe the change of the electrical resistance three ohmmeters were used.

Results
The time constant relative to the response time if the system could be modelled as a first order, or a second order well dumped system, in response to a step input as it is a case for our lateral infusion process.
First order system transfer function was used to model the sensory yarn (Equation 2), the input is defined as x(t), where x is the front resin position and ∆ (s) is the sensory yarn output. If the input is equal to the step input (Heaviside function), the output is presented in Figure 5. The transfer function identification gives the static gain K = 0.025 and the time constant τ= 5 second (simulated step response is given in Fig. 6). (1 st order system) and ∆ ( ) . .
(2 nd order system) Where K = 0.025, proper frequency n=0.4 and the damping factor =0.7 (see simulated step responses in Fig 6).