Engineering design specifications and testing protocols for orthosis joint designs are available through national and international standards committees. 1–3 Understandably, these standards were developed with respect to the then state-of-the-art orthosis joint components. Testing typically has been conducted with the knee joint locked at full extension. This made sense because the devices are generally locked and loaded during ambulation and unlocked only for transfers and seating. Stance phase control knee brace designs have been developed recently that promise greater functionality. 4 These advanced designs promise dynamic control of the locked and unlocked states, and some provide a multiplicity of locking positions. Current structural testing standards do not address the dynamic nature of this new generation of knee joint designs. 2–4 This is of particular concern with respect to durability because of the increased number of potential failure modes. A new testing protocol needs to be developed to investigate the fatigue and wear properties of stance phase control knee joint mechanisms. Therefore, the purpose of this project is to develop a mechanical fatigue test for the stance phase control knee joints now entering the commercial market. Such a system of hardware and methodology should be scalable according to target population anthropometrics and should simulate anticipated loading patterns.
To design a realistic protocol, normal knee kinetics and kinematics were used as a test parameter selection guide. Although knee brace users are not able to reproduce normal knee kinetics or kinematics, we assumed that users would naturally adopt similar patterns based upon esthetics and inherent energy efficiencies. Normal sagittal knee motion can be divided into stance and swing phases (Figure 1). Stance phase begins at foot strike (0 percent of gait cycle) and ends at foot off (approximately 62 percent of gait cycle). Although it is assumed that a locking knee brace will provide stability during stance phase, normal kinematics require approximately 40° of knee flexion at the end of stance. Swing phase begins at foot off and ends at foot strike or 100 percent of the gait cycle. Beyond foot off, the knee flexes another 10° to 20°, then extends to 0° in preparation for the subsequent foot strike and stance phase. Normal sagittal knee kinetics provided an estimate of the character and magnitude of moments about the knee. These data typically are normalized by the product of weight (N) and height (m) of the individual (Figure 2). Peak internal knee extensor moments during stance equal approximately 3 percent of the weight-height product. Of particular note is the second peak that occurs during double limb support and equals approximately 1 percent of the weight-height product. Thus, there exists an internal knee extensor moment when the joint needs to be released and allowed to flex in preparation for swing. Based on this information, the following four principal requirements are set forth for stance phase control devices: 1) lock reliably; 2) unlock reliably under physiologic loads; 3) provide adequate durability; and 4) allow adequate range of motion during swing phase of gait.
The testing system consisted of a steel frame, dead weights, pneumatic circuit, electronic interface circuitry, and personal computer-based data acquisition (Figure 3). A steel frame measuring 1.6 m high, 0.89 m wide, and 0.99 m deep was constructed from 0.038 m square tubing. A vertical 6.3-mm steel flange provided support for the orthosis joint under test. A pneumatic cylinder (Numatics 1750D02–15A; Numatics, Highland, MI) was mounted vertically to the top of the frame with a pin joint. The cylinder rod was joined to the dead weight and the orthosis joint through a second pin joint. The moment arm distance from the rotation axis of the orthosis joint to the cylinder rod/dead weight pin joints was 40 cm. An 11.4-kg dead weight load was selected, resulting in a 44.7-Nm joint load. This represented the peak flexion moment during gait for a 50th percentile (height) male weighing 87 kg (76th percentile). A proportional pneumatic valve (Proportion-Air QB2TFEE100; Proportion-Air, McCordsville, IN) modulated compressed air supplied to the cylinder through a flow control valve (Numatics 2FPTN8). Compressed air was delivered to the system through a regulator (Wilkerson R08–02-F0G0; Wilkerson, Englewood, CO) with a relief valve (Airtrol RV-5300–100; Airtrol, New Berlin, WI) in series with the proportional valve (Figure 4).
An ‘S’ style load cell (INTERFACE 250) was mounted in line with the cylinder to measure force taken up by the cylinder. The orthosis joint flexion/extension was monitored using a 10KΩ single-turn potentiometer mounted coincident with the orthosis joint rotation axis. Temperatures of the electromechanical actuator and the orthosis joint were monitored using two separate programmable temperature controllers (Analog Devices TMP01; Norwood, MA).
Two separate 1,000,000-cycle tests were performed on identical joint assemblies. The joints used for testing were based upon a newly patented over-running wrap spring clutch designed for the orthopedic appliance industry. 5 When engaged, the device restricts joint flexion but allows joint extension (over-running). It is self-engaging and requires electromechanical input only to disengage. Both flexion and extension rotations are possible when the device is disengaged.
A separate electronics enclosure housed regulated power and signal conditioning components for the fatigue testing system. A DMD-460 series bridge amplifier (Omega Engineering, Stamford, CT) provided load cell conditioning. The computer interface for monitoring and control of the fatigue tester was a multifunctional data acquisition expansion card (National Instruments, PCI-MIO-16E-4; Austin, TX).
Computerized control and monitoring were performed by a custom designed program (National Instruments, LabView). The program was designed to monitor performance throughout each test cycle. Each cycle was subdivided into six distinct stages (Figure 5).
- Stage 1: The pressurized cylinder lifts the dead weight and moment arm until the orthosis joint rotates to a horizontal position. This corresponds to knee extension during the second half of swing phase of the gait cycle.
- Stage 2: The vented cylinder allows the orthosis joint to take up load. This corresponds to weight acceptance at the beginning of the stance phase of gait.
- Stage 3: The load is held for a specific time period. This corresponds to the middle portion of stance phase.
- Stage 4: The cylinder is again pressurized, reducing the load on the orthosis joint to a user-specified level. This stage represents the unloading portion of stance phase.
- Stage 5: The disengaged orthosis joint allows the dead weight to fall back to the starting position. This critical stage represents terminal stance phase and the first half of swing phase. Knee joint release is triggered when the moment load across the joint has been reduced to the user-specified level.
- Stage 6: Data from the preceding cycle are logged, and the system resets to prepare for the start of a new cycle.
Fault checking was conducted at each stage. When a fault condition was detected, all testing was stopped. No additional command signals were initiated until the user reset the system. The stage-specific criteria for stopping testing are:
- Stage 1: Rotary potentiometer fails to indicate that the orthosis joint attained hold position, indicating a failure to allow full range of motion.
- Stage 2: Rotary potentiometer indicates the orthosis joint failed to lock in the hold position, signifying a loss of reliability and/or mechanical strength.
- Stage 3: Rotary potentiometer indicates the orthosis joint allowed excessive slippage during the hold stage, which is the loss of reliability.
- Stage 4: Load cell indicates that the pneumatic cylinder failed to unload the orthosis joint to the user-specified level, evidence of a problem with test equipment.
- Stage 5: Rotary potentiometer fails to indicate that the orthosis joint reached maximum flexion position, indicating a loss of unlock reliability.
Fault checking at all stages:
- 1) No change in or failure to meet target parameter values within 6 seconds of command initiation.
- 2) Excessive heating of the electromechanical release or orthosis joint.
Data collection, display, and storage have been designed to provide evaluation of the test system and the knee joint performance without excessive storage requirements. The control panel provided a graph of the hold angle for the most recent 2,000 cycles, temperature, and control state indicators in real time. Data were stored for each of the first 10 cycles of a test and thereafter at 500-cycle intervals. Each data file included header information identifying the specific device, test condition, date, time, and number of test cycles completed. Data were sampled at 75 Hz for the load cell, the rotary potentiometer, and the temperature sensors.
The system successfully operated for more than 2,000,000 cycles (approximately 2,400 hours). A cycle period of 4.3 seconds is achieved when an applied moment of 44.7 Nm at the knee joint and an imposed joint motion of 50° are used. Joint load at the time of release was set at 2.4 nm to replicate the physiologic load present at the end of stance phase (Figure 2).
Representative data are shown in Figure 6. The test cycle starts with the joint in the flexed position. Control stage 1 pressurizes the cylinder, thereby increasing the force registered by the load cell and extending the joint, as indicated by the position trace. When the fully extended 50° position is reached, control stage 2 turns off the compressed air line. The pneumatic cylinder is vented to atmospheric pressure, eventually reducing the load cell reading to zero. The dead weight load is held in position by the knee joint under test as indicated by the position trace. Stage 3 defines the requisite hold time at full load before the joint is released. Compressed air is again added to the cylinder during stage 4 until the moment load on the joint is reduced to 2.4 Nm. Stage 5 begins at disengagement of the knee joint. The electronic pneumatic valve simultaneously vents the pressurized cylinder to the atmosphere. The joint begins its return to the 0° position under the influence of the dead weight load. The air entrapped within the pneumatic circuit moderates the descent of the weight and therefore the flexion of the joint. This is demonstrated by the relatively stable load cell output during the later half of stage 5. Stage 6 begins once the joint has reached the 0° position. The knee joint is re-engaged, data are recorded, and the system is reset for the subsequent test cycle.
Although orthosis design and testing share some similarities to prosthetic design and testing, they do not share similar representation in national or international standards. Both static and cyclic structural testing are well defined for lower-limb prostheses 6 and are referenced by orthoses specifications. 2 The same cannot be said for the orthoses standards in which the techniques are defined but loading specifics generally are left to the discretion of the supplier. 2,3 The exception is the Japanese Industrial Standard, which provides both static and dynamic cyclic load levels specific to each axis of the joint. 1 These specifications for bending a locked knee orthosis joint in the sagittal plane are 50 Nm of load at a frequency between 2.0 and 3.3 Hz for 100,000 cycles. For comparison, the test results presented in the current report were collected with a 44.7-Nm load at a frequency of 0.23 Hz for 1,000,000 cycles. None of the currently available standards addresses the performance of the release or locking mechanisms incorporated into the stance phase control designs. This new test methodology is designed to integrate dynamic joint motion with endurance testing of the locking and unlocking mechanisms. This methodology, in combination with established static test measures, provides a more complete proof of performance without endangering human subjects. The potential impact on the clinicians fitting and maintaining these knee orthoses will be improved hardware performance in actual field use. Fatigue and wear testing draw out the weaknesses of designs, leading to revisions before the designs advance to market. Such testing can be used to define inclusion/exclusion criteria for the population being served.
Thus, mechanical testing is the bridge between the designer and the user. In particular, fatigue testing provides designers and manufacturers evidence that the design, materials, and the manufacturing practices combine to meet performance expectations. The optimal stance phase control knee joint should meet four main requirements in an endurance test: 1) reliable locking, 2) reliable unlocking under physiologic loads, 3) durability in the field, and 4) adequate range of motion. Relevance of fatigue tests relies on application of loads representative of those encountered in the field. This test equipment provides quantitative data to document stance phase knee joint performance for each of these requirements. Results show successful locking and functional durability strength to 1,000,000 cycles, a 50° range of motion, and magnitude of applied load when the joint is unlocked.
This fatigue system has limitations. The first two are test speed and load control. The testing frequency is low because of the 50° arc of motion imposed on the joint. The dead weight must move through this same arc, which initiates undesirable inertial dynamics. As the pneumatic cylinder raises the beam, the dead weight is driven into oscillations, causing longitudinal and lateral stresses on the orthosis joint. Tethering the weight to the test frame diminishes this effect; however, million-cycle tests still require 1,200 hours. Future modifications may include the use of pulleys to remove the pendulum oscillations inherent to a hanging dead weight from the actual point of application. Also being considered is a complete replacement of the dead weight with bidirectional control of the pneumatic cylinder. Either option would provide opportunities to reduce cycle time and provide a more fully defined loading profile. The third limitation of this system is that it has been designed to apply only sagittal plane forces, even though longitudinal and transverse loads also are seen in actual knee-ankle-foot orthosis applications. These forces may be significant in some cases but are lower in magnitude than sagittal plane forces because of the moment arm increase as the knee joint moves into flexion. Thus, the focus of this methodology is to challenge the sagittal plane locking and unlocking in combination with the dynamic range of motion. It is not intended to supplant established protocols but to be an additional tool for designers and manufacturers.
The fatigue system has been assembled from standard industrial components, making it easy to replicate. Instrumentation for control and test monitoring includes a tension load cell, a rotary potentiometer, and two thermocouples. All of these components are readily available at reasonable cost. They are also mature and robust designs, lending themselves to extended test protocols. The control software has been constructed as a closed-loop system so that failures in any component stop the test. In this way, both the device under test and the testing system are protected from a runaway condition. The electronically controlled proportional control valve is the most important component in determining system performance. It provides exquisite control of the pneumatic cylinder and load during both stage 1 (raising load through range of motion) and stage 4 (partially unload the orthosis joint preparing for release).
This fatigue test system has been constructed as an open test bed for stance phase control knee orthosis components. However, it could be modified to test complete orthoses for upper and lower extremities or possibly prosthetic components. Mounting full orthoses for testing would entail significant challenges. Most rely on limb volume within the cuff to provide additional structural stability and to maintain proper alignment of orthosis and anatomical joint axes. Proxy limb segments would need to be developed and standardized for each test. The integrated programmability of this novel fatigue and wear test system would make it possible to accommodate the different dynamic ranges of motion and loading criteria for each application.
The authors thank Dan Stucky, Brent Adams, Jennifer Mroz, Andrew Anderson, Beth Galle, Jim Craighead, and Paul Kane for their valuable contributions.
1. Metallic knee joints for lower extremity orthoses. Japanese Industrial Standard, JIS T 9216–1991, Japanese Standards Association, Tokyo, Japan, 1991.
2. External limb prostheses and external orthoses: requirements and test methods. European Standard EN 12523, European Committee for Standardization, Brussels, Belgium, 1999.
3. Lower limb orthoses. British Standard BS 2574; Parts 1–3, BSI, London, UK, 1991.
4. Travolta RL. Stance control revolutionizes knee bracing. Biomechanics 2002; 10: 53–62.
5. Irby SE, Kaufman KR. Electromechanical joint control device with wrap spring clutch. United States patent 6 500 138; December 2002.
6. Prosthetics: structural testing of lower-limb prostheses. ISO 10328: Parts 1–8, International Organization for Standardization, Geneva, Switzerland, 1996.