2.1 NiTi-based SMA actuator

A NiTi SMA spring exhibiting a one-way shape memory effect was utilised as the primary actuation element of the mechanism. NiTi-based SMAs are widely employed in aerospace and precision actuation systems due to their high energy density, relatively superior corrosion resistance compared to other SMA compositions. These materials are capable of generating significant mechanical force through thermally induced phase transformations between the martensite and austenite phases, making them particularly suitable for compact actuator systems operating in spatially constrained environments. The selected SMA spring has an A _f of 45 °C, and its key properties are presented in Table 2.

Table 2.

Physical properties of the SMA actuator

Figure 3.

Landing gear door lock mechanism.

The force exerted by a conventional coil spring, f, when deformed by a displacement δ, is described by Hooke’s Law, as expressed in Equation (1),

(1) \begin{align}f = K\delta \end{align}

where K is the spring stiffness. For a helical spring, the stiffness can be expressed as a function of the material and geometric properties, as shown in Equation (2) [Reference Kim, Lee, Park and Park34]. Specifically, G is the shear modulus of the material, d is the wire diameter, D is the mean coil diameter and N is the number of active coils.

(2) \begin{align}f = \frac{{G{d^4}\delta }}{{8{D^3}N}}\end{align}

2.2 Joule heating

The actuation of the NiTi-based SMA spring in the proposed mechanism is achieved through resistive (Joule) heating. Electrical current supplied via terminals at both ends of the spring generates heat internally due to the alloy’s electrical resistance. As temperature rises, the SMA undergoes a phase transformation from martensite to austenite, resulting in contraction and mechanical output [Reference Mavroidis35]. This process follows the electrical power dissipation principle, as expressed in Equation (3).

(3) \begin{align}P = \;IV\; = \frac{{{V^2}}}{R}\;\end{align}

While the electrical power dissipation is described by this fundamental principle, the resulting thermal response and phase transformation of the SMA spring are governed by a more detailed energy balance. This balance accounts for the heat generated internally by the current, the latent heat absorbed during the martensitic phase transformation and the convective heat loss to the ambient environment. This thermodynamic process can be modeled using an adaptation of the one-dimensional heat transfer equation proposed in the literature [Reference Hadi, Yousefi-Koma, Elahinia, Moghaddam and Ghazavi36], as shown in Equation (4).

(4) \begin{align} m{C_p}\frac{dT}{dt} + mL\dot \xi &= \frac{{{V^2}}}{R} - hA\left( {T - {T_\infty }} \right) \end{align}

(5) \begin{align} R &= {R_A} + \xi \left( {{R_M} - {R_A}} \right) \end{align}

Here, m represents the mass of the SMA spring, C _p is the specific heat capacity, T is the temperature, L is the latent heat of transformation and ξ˙ is the rate of change of the martensite fraction. The term V ²/R corresponds to the electrical power input, while h is the convection heat transfer coefficient, A is the surface area of the spring and T _∞ is the ambient temperature. In addition, the term R is expressed in terms of ${R_A}$ and ${R_M}$ , which correspond to the electrical resistance in the austenite and martensite phases, as presented in Equation (5). This model provides a comprehensive framework for understanding the thermal dynamics that drive the mechanical output of the actuator.

Joule heating is preferred for SMA activation because of its simplicity, precise controllability and compatibility with embedded systems. It removes the need for external heat sources or complex thermal management, making it especially suitable for aerospace applications with severe limits on weight, volume and system complexity. However, careful regulation of current is essential to avoid overheating, which may degrade the SMA’s functional qualities and reduce operational lifespan. For aerospace applications, where actuator systems are required to operate under ambient temperatures ranging from −55 °C to +80 °C [Reference Sciascera, Giangrande, Brunson, Galea and Gerada37], a closed-loop control strategy is adopted as a design approach, in which thermal and electrical feedback obtained from the actuator system is used to estimate the SMA state. Based on this feedback, the activation energy supplied to the SMA can be adjusted to accommodate varying environmental conditions and to support stable actuation performance.

Future iterations will incorporate a closed-loop control system to guarantee operational consistency under various environmental circumstances, even if the current implementation depends on a fixed-power open-loop technique. Activation precision can be improved, and functional failure can be avoided by a feedback loop that uses temperature sensors or position sensors (e.g. Hall effect or linear variable differential transformer (LVDT)). In particular, the controller can dynamically increase the voltage and current to provide quick thermal compensation if the mechanism is unable to achieve the necessary 7 mm stroke because of extreme ambient cold during flight. This will guarantee that the unlocking time stays within the expected duration. On the other hand, the feedback system can control active cooling in high-temperature settings to preserve quick cycle times. Moreover, closed-loop control is essential for reducing the effects of SMA functional fatigue; the control algorithm can modify power parameters to compensate for changes in transformation temperatures and preserve the necessary mechanical output as the NiTi material degrades over thousands of cycles. By combining proportional-integral-derivative (PID) or fuzzy logic controllers with pulse width modulation (PWM), accurate energy management would be possible, preventing overheating and drastically lowering power usage during steady-state periods [Reference Sohn, Ruth, Yuk and Choi32].

2.3 Self-locking mechanism

The mechanism features a passive self-locking behaviour that maintains its locked position without the requirement for constant electrical power or additional actuation forces. This feature is achieved by properly locating a mechanical critical moment point, which creates a transition threshold between the locked and unlocked states. When in its passive condition, an external load supplied to the hook provides a clockwise torque that sustains the locked configuration of the system, as indicated in Fig. 4a.

When activated, the SMA spring is electrically heated beyond its activation temperature, inducing contraction and generating a stroke of approximately 7 mm, which retracts the piston and enables the release of the lock, as depicted in Fig. 4b.

A significant feature of the design comes in the mechanism’s capacity to maintain its current condition, either active or passive, without any external interference. This characteristic is ensured by the precise positioning of the critical moment point and the integration of tension springs under preload, which maintain the mechanism in its actuated state and prevent inadvertent resetting without the introduction of an external force. As the SMA material cools, the memory effect dissipates; however, in the presence of an external force applied to the hook, the preloaded tension spring is re-engaged, automatically restoring the mechanism to its original locked configuration, as demonstrated in Fig. 4c.

2.4 Secondary safety lock

The secondary safety lock is incorporated as an essential redundant element to guarantee the operational integrity of the mechanism in failure scenarios. A principal challenge in SMA applications is the potential for inadvertent activation caused by uncontrolled Joule heating, which may arise from electrical short circuits or severe ambient thermal conditions. The secondary lock functions as a mechanical barrier that physically inhibits the hook’s rotational trajectory, regardless of whether the primary SMA spring attains its transformation temperature prematurely, as illustrated in Fig. 5.

The operational sequence adheres to a defined safety protocol: the secondary clamp must be removed by a designated low-current instruction prior to the primary actuator commencing the unlocking stroke. This dual-stage activation guarantees that the system stays in a ‘fail-safe’ condition over extended flying phases. A preliminary failure modes and effects analysis (FMEA) was conducted to thoroughly assess the reliability of this architecture, as summarised in Table 3.

Figure 4.

Self-locking mechanism. (a) Locked state (b) unlocked state (c) self-locking state.

Figure 5.

Secondary safety lock.

This safety analysis reveals that the secondary lock, despite increasing the mechanism’s complexity, is crucial for limiting dangers associated with heat or electronic breakdowns. The design attains the high reliability criteria necessary for aircraft conditions by separating the safety restraint from the principal actuation force.

2.5 Cooling system design

Effective temperature management is a critical design consideration for ensuring the functional reliability and cycle repeatability of SMA-actuated systems. In the proposed mechanism, rapid cooling of the NiTi-based SMA compression spring is essential to minimise the residual force associated with the shape memory effect during the austenitic phase. As the SMA cools and transitions back to the martensitic phase, the generated actuation force progressively decreases, allowing the preloaded tension springs to regain mechanical dominance and return the system to its self-locking configuration.

Table 3.

Preliminary FMEA for the locking mechanism

Figure 6.

Overall view of the cooling system.

For aerospace applications, actuator systems are generally required to operate reliably under repeated thermal cycles across a wide ambient temperature range, typically from −55 °C to +80 °C. Accordingly, the SMA material must be selected based on its martensite–austenite phase transformation temperatures (A _s , A _f , M _s , M _f ) to ensure proper phase transition and actuation performance within this operational envelope. To maintain consistent recovery and thermal stability under varying ambient conditions, a closed-loop cooling approach is considered as a potential design strategy, in which the application of coolant would be adjusted based on the SMA temperature. The overall configuration of the cooling system is illustrated in Fig. 6. This approach aims to support repeatable actuation and minimise residual forces without relying solely on ambient temperature, while remaining at the conceptual stage in the present study.

Experimental observations indicate that the use of a coolant-assisted cooling method provides effective and repeatable recovery behaviour, contributing to improved thermal stability and reduced response time. To facilitate the practical implementation of this thermal management strategy and ensure efficient heat rejection from the internal environment to the ambient atmosphere, the physical architecture of the mechanism is optimised. Ventilation grills located on the exterior casing are retained to support passive heat dissipation and to facilitate the removal of residual heated air from the SMA chamber, as illustrated in Fig. 7. The combined use of controlled cooling and passive heat dissipation enhances the overall thermal robustness of the system without increasing structural complexity, making the proposed cooling design suitable for aerospace applications subject to fluctuating environmental conditions.

Figure 7.

Ventilation grills.

Figure 8.

Locked state.

3.1 Kinematic analysis

2D analyses were performed using CATIA V5 to evaluate the motion characteristics of the mechanism. The critical parameters identified were: (i) the moment threshold point governing the transition between locked and unlocked states; (ii) the minimum required stroke; and (iii) the placement of the biased tension spring pair, as illustrated in Figs. 8 and 9. The mutual interaction among these factors directly affects both the energy demand on the actuator and the overall dimensions of the design. A collective consideration during the optimisation process is essential to ensure reliable actuator performance and compliance with dimensional requirements.

Figure 9.

Unlocked state.

Figure 10.

Free body diagram of the lock mechanism.

Based on iterative 2D kinematic evaluations, the mechanism was systematically refined in two dimensions: the moment threshold and spring placement were precisely defined, and the necessary stroke length was reduced to 7 mm. The optimised stroke length reduced the force requirement on the SMA spring, improving energy efficiency and facilitating a more compact and mechanically robust mechanism capable of reliable locking and unlocking, which is essential for lightweight, high-performance systems.

In parallel, the configuration of the preloaded tension spring was refined to ensure autonomous return to the locked position after actuation. The spring was engineered to achieve its maximum extension during the unlocking motion, thereby generating sufficient restoring force and ensuring mechanical stability in both operational states without requiring external resetting. Consequently, the conducted 2D kinematic analyses established the fundamental framework of the mechanism and provided a solid foundation for the upcoming 3D modelling and dynamic simulation stages.

3.2 Static model

The static model can be derived from the force and moment equilibrium equations, which are obtained from Newton’s laws of motion under static equilibrium conditions. As shown in Fig. 10, there are four primary elements and ten state variables in the hook lock mechanism. The main objective of these formulations is to determine the unlocking force, f _unlock , required for the actuation of the hook mechanism. Here, ${f_{load}}$ is the load force; $f_{ij}^x$ is the force between the ith component and the jth component inthe x direction; $f_{ij}^y$ is the force between the ith component and the jth component in the y direction; ${f_{preload}}\;$ is the spring force; $f_{preload}^x$ is the spring force in the x direction; $f_{preload}^y$ is the spring force in the y direction; $f_{g4}^x$ is the force from the lock bracket to the jth component in the x direction; and $f_{g4}^y$ is the force from the lock bracket to the jth component in the y direction.

The static equilibrium of the mechanism is established by formulating the force and moment balance of the four primary elements. These relations are expressed in Equations (6)–(14). Ultimately, the unlocking force of the mechanism $,\;{f_{nlock\;}}$ , is determined from the force equilibrium given in Equation (15).

(6) \begin{align} \sum {F_x}\;:\;f_{34}^x - f_{g4}^x &= 0 \end{align}

(7) \begin{align} \sum {F_y}\;:\;f_{g4}^y - f_{34}^y - {f_{load}} &= 0 \end{align}

(8) \begin{align} \sum {M_{{O^7}}}\;:\;{f_{load}} \cdot \left| {l_{{O^7}{O^8}}^x} \right| - f_{34}^y \cdot \left| {l_{{O^6}{O^7}}^x} \right| - f_{34}^x \cdot \left| {l_{{O^6}{O^7}}^y} \right| &= 0 \end{align}

(9) \begin{align} \sum {M_{{O^6}}}\;:\;{f_{load}} \cdot \left| {l_{{O^6}{O^8}}^x} \right| - f_{g4}^y \cdot \left| {l_{{O^6}{O^7}}^x} \right| - f_{g4}^x \cdot \left| {l_{{O^6}{O^7}}^y} \right| &= 0 \end{align}

(10) \begin{align} \sum {F_x}\;:\;f_{23}^x - f_{34}^x &= 0 \end{align}

(11) \begin{align} \sum {F_y}\;:\;f_{34}^y - f_{23}^y &= 0 \end{align}

(12) \begin{align} \sum {F_x}\;:\;f_{12}^x + f_{g2}^x + f_{preload}^x - f_{23}^x &= 0 \end{align}

(13) \begin{align} \sum {F_y}\;:\;f_{23}^y - f_{preload}^y + f_{g2}^y &= 0 \end{align}

(14) \begin{align} \sum {M_{{O^3}}}\;:\;f_{12}^x \cdot \left| {l_{{O^2}{O^3}}^y} \right| + f_{preload}^y \cdot \left| {l_{{O^3}{O^4}}^x} \right| - f_{preload}^x \cdot \left| {l_{{O^3}{O^4}}^y} \right| + f_{23}^y \cdot \left| {l_{{O^3}{O^5}}^x} \right| - f_{23}^x \cdot \left| {l_{{O^3}{O^5}}^y} \right| &= 0 \end{align}

(15) \begin{align} \sum {F_x}\;:\;{f_{unlock}} - f_{12}^x &= 0 \end{align}

3.3 Dynamic analysis

Iterative dynamic analyses were conducted using MSC Adams software to evaluate the force requirements of the NiTi-based SMA spring and the preloaded tension spring under an external load of 20 N. This configuration corresponds to the per-lock requirement for a dual-lock architecture, yielding a total holding force of 40 N, representative of lightweight closure mechanisms implemented in small-scale UAV platforms. The selection of a thermoplastic material system served as a primary determinant in establishing the 20 N load threshold. In this prototype-scale investigation, polyethylene terephthalate glycol (PETG) was selected for laboratory validation owing to its adequate impact strength and thermal resistance, thereby ensuring operational integrity without permanent structural deformation. Furthermore, the underlying mechanism remains inherently scalable; by integrating SMA actuators with distinct thermomechanical and physical profiles, the system can be adapted to meet higher load requirements in various UAV platforms.

These analyses identified the minimum SMA activation force (austenite phase) and the maximum resistive force (martensite phase), and a robust linear relationship between these forces is revealed in Fig. 11. In Fig. 12, the mechanism’s stroke throughout its operational cycle is illustrated. The force profile under a 36 N activation force with a 16 N resistive force (the system remains functional for resistive forces below 19 N) is presented in Fig. 13. These results were validated through laboratory testing, highlighting the predictive accuracy of the dynamic model.

Figure 11.

Dynamic analysis results indicating stable system operation based on minimum activation and maximum resistance force.

Figure 12.

SMA stroke analysis.

Figure 13.

SMA force analysis.

5.1 Fatigue life and thermal cycle stability

NiTi-based shape memory alloys (SMAs) are widely employed in spring-type actuators due to their high recoverable strain capability (up to 8%), compact form factor and high actuation energy density, which make them suitable for force-generation applications in constrained volumes. In the literature, the fatigue behaviour of shape memory alloys is commonly categorised into structural fatigue, which is associated with crack initiation and fracture, and functional fatigue, which refers to the degradation of functional properties, including a reduction in shape memory or superelastic effects and an increase in the percentage of residual martensite with increasing cycle numbers [Reference Eggeler, Hornbogen, Yawny, Heckmann and Wagner41, Reference Yan, Wei, Chu, Wang, He, Ren and Sun42]. Owing to the inherently cyclic nature of such actuator applications, NiTi SMAs are commonly subjected to repeated thermomechanical loading. Under these conditions, the cyclic behaviour of NiTi SMAs is primarily discussed in terms of functional fatigue, which refers to the progressive degradation of functional characteristics such as actuator stroke, recoverable strain, plateau stresses, hysteresis width and transformation temperatures, rather than catastrophic structural failure. This degradation is attributed to the accumulation of transformation-induced defects in the microstructure [Reference Frenzel43]. For actuator-oriented applications, functional fatigue is often the dominant factor limiting long-term performance rather than structural fracture.

Reported experimental studies indicate that the functional fatigue behaviour of NiTi shape memory alloys is dominated by a pronounced early-cycle degradation stage, followed by a more stabilised cyclic response. As shown in Fig. 18, which reproduces strain-controlled low-cycle fatigue results obtained on NiTi wires by Lima et al., the functional characteristics of the NiTi wire were investigated through tensile loading–unloading tests conducted up to a maximum strain level of 6%. The results indicate that a substantial portion of the functional degradation occurs within the first 50–150 cycles. Specifically, Fig. 18(a) shows that the residual strain increases rapidly from approximately 0.2% to nearly 1.0%, while Fig. 18(b) indicates that the recoverable strain decreases from about 6.0% to approximately 5.0%, demonstrating that nearly 1% of the total transformation strain capacity becomes irreversible during early cycling. Concurrently, Fig. 18(c) illustrates a pronounced reduction in the direct transformation stress from approximately 360 MPa to about 230 MPa, and Fig. 18(d) shows a decrease in the dissipated energy per cycle from roughly 140 J/m³ to 80 J/m³, highlighting a significant loss in actuation capability despite the absence of macroscopic fracture [Reference Lima, Rodrigues, Ramos, Costa, Braz Fernandes and Vieira44].

Figure 18.

Mechanical behaviour of NiTi wire. (a) Residual strain; (b) recovered strain; (c) direct transformation stress and (d) dissipated energy versus function of the number of cycles [Reference Lima, Rodrigues, Ramos, Costa, Braz Fernandes and Vieira44].

Beyond this initial transient regime, Fig. 18 further demonstrates that the evolution of residual strain, recoverable strain, transformation stress and dissipated energy proceeds at markedly reduced rates, indicating the establishment of a stabilised cyclic response [Reference Lima, Rodrigues, Ramos, Costa, Braz Fernandes and Vieira44]. Previous studies have demonstrated that the fatigue behaviour of SMAs is strongly governed by microstructural evolution and its influence on the stress–strain hysteresis during cyclic loading. Appropriate thermomechanical treatments can stabilise the cyclic response and mitigate strain localisation, thereby enhancing functional stability. In pseudo-elastic operation, maintaining the transformation plateau stress below the yield stress of the austenitic phase effectively limits dislocation generation during cyclic loading, leading to reduced accumulation of residual strain. Nevertheless, even under optimised conditions, irreversible strain accumulation cannot be completely avoided, as continued cyclic loading promotes progressive increases in dislocation density and the stabilisation of martensitic variants that do not fully transform back to the austenitic phase upon unloading [Reference Eggeler, Hornbogen, Yawny, Heckmann and Wagner41].

From a strain-based perspective, the degree of functional fatigue can be quantitatively estimated using the reduction in recoverable strain. As shown in Fig. 18(b), the recoverable strain decreases from approximately 6.0% in the initial cycles to about 5.0% after 100–150 cycles, corresponding to a relative functional degradation of approximately 16–17% with respect to the initial recoverable strain. This strain-based fatigue metric provides a physically meaningful functional performance loss under cyclic thermomechanical loading.

In the present study, the NiTi-based SMA spring actuator was subjected to more than 100 thermomechanical activation cycles, with cooling predominantly achieved via two distinct approaches, namely forced convection and refrigerant-based cooling. The actuator, featuring a three-loop configuration and a solid-state length of 6 mm, was extended to approximately 60 mm during operation, corresponding to a nearly tenfold displacement relative to its initial length. When interpreted in terms of the reduction in force generation capability, the functional fatigue of the SMA spring actuator is estimated to be approximately 10%.

Although the results presented in Fig. 18 were obtained from uniaxial strain-controlled fatigue tests on NiTi wires, they provide valuable insight into the intrinsic material-level mechanisms governing functional fatigue in NiTi alloys. Consequently, while the absolute stress and strain values may differ due to geometric effects, the trends observed in Fig. 18, namely rapid functional degradation at low cycle numbers followed by a stabilised fatigue regime, are expected to remain qualitatively valid for spring-type SMA actuators.

Regarding the thermal cycle stability, the mechanism was activated for 12 cycles under controlled electrical parameters of 20 V and 12 A. The activation duration yielded a mean value of 16 s. Statistical analysis revealed that the deviation remained within a narrow margin, characterised by a 95% confidence interval (CI) of ±0.8 s. This high level of precision and the negligible variance across thermal cycles demonstrate the exceptional repeatability and operational stability of the mechanism under the specified conditions.