Open AccessArticle

Cycle Time-Based Fault Detection and Localization in Pneumatic Drive Systems

Vladimir Boyko

and

Jürgen Weber

Chair of Fluid-Mechatronic Systems (Fluidtronics), TUD Dresden University of Technology, Helmholtzstr. 7a, 01069 Dresden, Germany

Author to whom correspondence should be addressed.

Actuators 2024, 13(11), 447; https://doi.org/10.3390/act13110447

Submission received: 28 September 2024 / Revised: 25 October 2024 / Accepted: 5 November 2024 / Published: 7 November 2024

(This article belongs to the Section High Torque/Power Density Actuators)

Download

Browse Figures

Figure 1
Potential fault locations within a pneumatic drive (symbolic depiction): (1) external leakage in piston-side chamber A; (2) internal (interchamber) leakage; (3) external leakage in rod-side chamber B; (4 and 5) external leakages between the directional valve and the throttle valves; (6) increased friction. "> Figure 2
Test bench for investigating cycle time-based fault detection with corresponding fault locations. "> Figure 3
Influence of external leakage Qext = 20 L/min between cylinder and throttle check valve with meter-out throttling on cylinder pressure and position: (a) piston side A (fault location 1); (b) rod side B (fault location 3). "> Figure 4
Influence of external leakage Qext = 20 L/min between directional control valve and throttle check valve with meter-out throttling on cylinder pressure and position: (a) piston side A (fault location 4); (b) rod side B (fault location 5). "> Figure 5
Influence of internal leakage Qint = 40 L/min in the cylinder (fault location 3) on the meter-out throttled cylinder’s pressure and position. "> Figure 6
Influence of increased piston friction Fadd,fr = 20 N (fault location 6) on the meter-out throttled cylinder’s pressure and position. "> Figure 7
Equal delay in the extension and retraction time in case of internal leakage Qint = 40 L/min in the rodless cylinder Festo DGC-18-200-G-PPV-A with meter-out throttling. "> Figure 8
Influence of external leakage Qext = 20 L/min between cylinder and throttle check valve with meter-in throttling on cylinder pressure and position: (a) piston side A (fault location 1); (b) rod side B (fault location 3). "> Figure 9
Influence of external leakage between directional control valve and throttle check valve with meter-in throttling on cylinder pressure and position: (a) piston side A (fault location 4); (b) rod side B (fault location 5). "> Figure 10
Influence of (a) internal leakage Qint = 30 L/min in the cylinder (fault location 3) and (b) increased piston friction Fadd,fr = 20 N (fault location 6) on the meter-in throttled cylinder’s pressure and position. "> Figure 11
Influence of the pneumatic frequency ratio Ω on changes in the movement time of the Ø25 × 50 Hoerbiger R6025/50 differential pneumatic cylinder for different fault locations and at constant fault values. Meter-out throttling: (a) fault location 1; (b) fault location 3; (c) fault location 4; (d) fault location 5; (e) fault location 2; (f) fault location 6. Meter-in throttling: (g) fault location 1; (h) fault location 3; (i) fault location 4; (j) fault location 5; (k) fault location 2; (l) fault location 6. "> Figure 12
(a) Influence of internal leakage on the extension time change of different pneumatic differential cylinders with meter-out throttling as a function of the pneumatic frequency ratio Ω, changes in pressure and position profiles, and interchamber flow direction in (b) well-sized cylinder with Ω < 1.5; (c) well-sized cylinder with Ω = 1.5; (d) oversized cylinder with Ω > 1.5. "> Figure 13
Influence of internal leakage on the extension time change of different pneumatic differential cylinders with meter-out throttling as a function of the mean piston velocity <math display="inline"><semantics> <mrow> <mover accent="true"> <mrow> <mi>x</mi> </mrow> <mo>˙</mo> </mover> </mrow> </semantics></math>. "> Figure 14
Change in the movement time of the Ø25 × 50 Hoerbiger R6025/50 differential pneumatic cylinder at Ω = 1.3 as a function of leakage rate. Meter-out throttling: (a) fault location 1; (b) fault location 3; (c) fault location 4; (d) fault location 5; (e) fault location 2. Meter-in throttling: (f) fault location 1; (g) fault location 3; (h) fault location 4; (i) fault location 5; (j) fault location 2. "> Figure 15
Algorithm for time-based fault detection and localization in pneumatic drives with well-sized, double-acting differential cylinders (PFR Ω ≤ 1.5) and meter-out throttling. "> Figure 16
Algorithm for time-based fault detection and localization in pneumatic drives with oversized, double-acting differential cylinders (PFR Ω > 1.5) and meter-out throttling. "> Figure 17
Algorithm for time-based fault detection and localization in pneumatic drives with double-acting differential cylinders (all PFR values) and meter-in throttling. "> Figure 18
Handling system (a) and its motion sequence (b). "> Figure 19
Changes in cycle time of pneumatic actuators and supply pressure of the handling system as well as room temperature during fault-free operation. "> Figure 20
(a) Pneumatic system for fully automated assembly of a fuel cell stack by XENON Automatisierungstechnik GmbH [<a href="#B39-actuators-13-00447" class="html-bibr">39</a>]; (b) external leakage generation at cylinder 1 (flap actuator). "> Figure 21
Above: Changes in cycle time of pneumatic actuators of the assembly machine resulting from external piston-side leakage in chamber A before and after the throttle valve; below: corresponding fault recognition paths of the algorithms. ">

Versions Notes

Abstract

Compressed air ranks among the most expensive forms of energy. In recent decades, increased efforts have been made to enhance the overall energy efficiency of pneumatic actuator systems and develop reliable fault detection methods for preventing energy losses. However, most of the methods developed so far require additional sensors, resulting in extra costs, and/or are not applicable during machine operation, which leads to their limited use in the industry. This article introduces a cycle time-based method for detecting faults in pneumatic actuators through the use of proximity switches, enabling cost-effective monitoring in real time without the necessity of further sensors. A systematic analysis is conducted, expanding the current state of knowledge by detailing the influence of all potential leakage points on the movement times of a pneumatic drive and taking into account the different velocity control strategies (meter-out and meter-in) and operating points expressed via the pneumatic frequency ratio. Previously unassessed specifics of internal leakage, including the impact of pressure profiles and differences between differential cylinders and cylinder with equal piston areas, are also presented. The applicability of the proposed method and its detection limits in an industrial environment are examined using pneumatic assembly machines.

Keywords:

energy efficiency; compressed air systems; pneumatics; pneumatic actuators; fault detection; condition monitoring; energy loss; leakage; friction

1. Introduction

As an easy-to-handle energy source, compressed air is used worldwide in a broad range of industrial sectors. Pneumatics is considered an inexpensive, robust, and environmentally friendly drive technology. However, its higher energy consumption compared to electromechanics has intensified research into energy efficiency and fault detection in compressed air systems. To exploit the vast energy saving potential of ca. 40% (see, e.g., [1,2]) in pneumatics, various energy-saving measures such as demand-oriented dimensioning methods for actuators, piping, and compressors as well as energy-saving circuit concepts have been developed in recent years [2,3,4,5,6,7,8,9,10,11]. There are also other concerns regarding the development of condition monitoring and fault detection strategies, which are addressed in this article.

1.1. Fault Detection in Compressed Air Systems

According to [12,13], leakage and increased friction are the most significant energy losses in compressed air systems. The robustness of pneumatic systems makes fault detection challenging, as they often remain functional even with severe faults such as major leaks. Leaks can occur anywhere due to factors such as component aging or wear. Since locating and eliminating leaks in a pneumatic system can save up to 40% of compressed air consumption [1,14], numerous manufacturers offer various tools for leakage detection.

Increased friction also indicates problems in the system, primarily due to seal wear or transverse forces causing asymmetrical piston loads [15]. This type of energy loss does not correlate with higher compressed air consumption, making ongoing detection more difficult compared to leaks.

Comprehensive reviews [16,17,18] have highlighted the growing focus on anomaly detection in compressed air systems over the recent years. However, although many proposed methods allow reliable fault detection, their usage mostly involve additional sensors, making them economically unviable for pneumatic machines with a large number of actuators. The summary below highlights the special features of current monitoring systems, with some methods facing technical limits such as requiring operation interruptions, which explains their low adoption in industry [19].

1.1.1. Leakage Detection

Most leakage detection methods have been presented in [16,17]; the following aims to focus on market-ready applications.

One example of a signal-based leakage detection and localization is Festo’s Motion Terminal (VTEM), which includes an extra “Leakage diagnosis” application that performs a diagnostic cycle to determine leakage volume flow from the cylinder volume, hose length, and pressure drop at a valve, interrupting normal operation [20]. Furthermore, the Festo MSE6-E2M maintenance device offers measuring, control, and diagnostic functions [21]. It detects leaks as a continuous pressure reduction in the blocked state over time or by exceeding a set flow limit. This maintenance device requires a predefined limit value, which is reported if exceeded. Similar volume flow-based systems from other manufacturers exist, e.g., [22], but they cannot localize faults without additional system information, such as actuation signals of directional control valves.

Another widespread signal-based method for leak detection is ultrasonic detection (see, e.g., [23,24,25,26]). Compressed air escaping from a leak generates turbulences that are detected using portable test devices without direct intervention in the pneumatic system. However, a manual ultrasonic inspection can be limited, especially in extensive or inaccessible pipe networks. Factors such as other ultrasonic sources and reflections from surrounding walls can complicate detection, requiring special sound shielding. Additionally, this method cannot be used during operation due to inherent venting noises. In [27,28], such noises were partially filtered out using a neural network, improving acoustic leak detection.

1.1.2. Increased Friction Detection

Determining the actual friction force typically requires pressure sensors, though pressure changes can also indicate leaks. Alternatively, the friction can be determined on a specialized test rig, making detection during operation impossible [29]. The article [30] presented a signal-based method for friction detection in pneumatic cylinders using high-frequency vibration analysis. It measures vibrations on the cylinder housing and evaluates their frequency responses. Increased vibration intensity correlates with seal wear. This method could theoretically be applied during operation, but it requires specialized technology and vibration protection.

1.1.3. Current Challenges: Summary

Approximately 60% of the fault detection strategies investigated in [17] require the use of multiple sensors (primarily pressure and volume flow) to identify and localize faults. For example, the authors of [18,31,32] developed fault detection strategies based on pressure and volume flow signals. While this certainly increases the informative value and also enables localization, it might be not economically viable for most companies.

The described state of the art demonstrates that there is still demand for monitoring systems that would enable fault localization during machine operation and without significant extra costs.

1.2. Cycle Time-Based Fault Detection

One possible solution to counter the issues described above is to exploit the information already available in pneumatic systems, such as proximity (reed) and valve switches signals, without the need for additional sensors. In this case, deviations in the cycle time can serve as the basis for fault detection. However, at the actuator level, this possibility has been scarcely addressed in the literature so far.

The review [17] mentioned only one publication by Gauchel et al. [33] on fault detection using proximity and valve switches, where external leakage and increased friction in a pneumatic clamp were distinguished by analyzing shifts in travel time and the reaction time between valve signals and clamp movement.

The studies [34,35] investigated the movement time of a pneumatic cylinder under external leakage conditions along with the volume flow and pressure signals, and observed a measurable deviation in cycle time. However, only a single leakage point in the exhaust air line was considered at the actuator level.

Another source addressing the issue of cycle time as a fault detector is the dissertation by González [31], in which a condition monitoring system for pneumatic drives based on pressure, volume flow, and switching valve signals was developed. The dissertation investigated the effects of faults such as external/internal leakage and increased friction on travel time in a pneumatic drive with meter-out throttling but did not discuss using proximity switches for condition monitoring. The author concluded that standard sensors, like proximity switches and valve signals, are insufficient for fault diagnosis.

In the dissertation by Fritz [13], cycle times were not used in the context of fault detection, although the author mentioned the effect of exemplary placed air leakages and increased friction on the movement time of the considered pneumatic cylinder.

It can be concluded that there has been a lack of systematic investigation in the area of time-based fault detection for pneumatic actuators until now. The known studies [31,33,34,35] have concentrated on double-acting differential cylinders with meter-out throttling and only considered some of the possible leakage locations in the drive system. The motivation for this article is therefore to expand the current state of knowledge and introduce a cost-effective and easy-to-implement fault detection and localization method by analyzing the influence of all potential leakage points on the travel times of a pneumatic drive, taking into account different velocity control strategies (meter-out and meter-in) and the degree of oversizing expressed via the pneumatic frequency ratio (PFR) Ω in accordance with [36]. The specifics of internal leakage, such as the influence of pressure profiles in cylinder chambers and differences between a differential cylinder and a cylinder with equal piston areas, which were not previously assessed, are also presented here. Finally, the applicability of the proposed method and its detection limits in an industrial environment are herein examined using the example of a pneumatic assembly machine.

2. Fault Types and Fault Impact on Cycle Times

The typical faults in a pneumatic drive can be classified as shown in Table 1 [13,18,31]:

Figure 1 symbolically shows potential fault locations within a standard pneumatic drive with meter-out throttling. The influence of fault locations 1 and 6 was investigated in [33] with the example of a pneumatic clamp system with meter-out throttling; fault locations 1, 2, 3, and 6 were considered in [31] for a cylinder with meter-out throttling and in [13] for a pneumatic welding gun without throttling. The study [34] examined leakage point 3, whereas in [35], leakage point 5 was considered. All cases involved differential cylinders only.

In this work, the influence of all the fault locations shown in Figure 1 on drives with meter-in and meter-out throttling is investigated. In the first step, the fundamental impact of fault cases shown in Figure 1 on the cylinder performance was analyzed on a test bench using a representative cylinder drive.

Figure 2 illustrates the test bench setup and shows the generation of various leakage points. The friction increase was simulated using the Festo DSBC-L-50-200-PPV-A pneumatic cylinder without compressed air supply mechanically connected to the guide of the cylinder under investigation.

The examined drive task was realized using the Festo DSBC-32-200-PPV-A double-acting differential cylinder and corresponds to a well-dimensioned cylinder with a PFR value of Ω = 1.3 … 1.7. According to [36,37], the PFR compares an actuator’s theoretical dynamics (frequency ω₀) to the system dynamics required by the task (frequency ω_f) using Equation (1):

Ω = \frac{ω_{0}}{ω_{f}} = \frac{t_{S}}{2 \cdot π} \cdot \sqrt{\frac{c}{m}},

(1)

where t_S is the stroke time, c the air spring stiffness, and m the moving mass. The cylinder is treated as a mass between two springs, with air in both chambers determining the spring stiffness. The stiffness for a single-rod pneumatic cylinder is given by the following:

c = \frac{A_{A}^{2} \cdot n \cdot p}{V_{A} + V_{d} + V_{t}} + \frac{A_{B}^{2} \cdot n \cdot p}{V_{B} + V_{d} + V_{t}},

(2)

where A_A and A_B are the piston area and the work area on the rod side, respectively;

n

the polytropic index; p is the supply pressure; V_A and V_B are the chamber volumes in the middle cylinder position for chamber A and B, respectively; V_d is the cylinder dead volume; and V_t is the tubing volume.

As stated in [36], properly dimensioned cylinders must have a PFR between 1.1 and 1.7. Cylinders with PFR values exceeding this range are considered as oversized, while those with values below 1.1 are under-dimensioned and require additional components, such as external hydraulic shock absorbers, to function effectively.

Table 2 summarizes the operating conditions of the cylinder in the reference state.

2.1. Meter-Out Throttling

The following figures show the pressure curves and the cylinder position of the chosen pneumatic cylinder with meter-out throttling in the fault conditions according to Figure 1 compared to the fault-free reference condition. As can be seen in Figure 3, external leaks located directly between a cylinder and a throttle check valve (fault locations 1 and 3 in Figure 1, e.g., a loosened push-in fitting) act as quick exhaust valves, allowing the respective cylinder chamber to be vented more rapidly. In the event of an external leak in chamber A, this leads to a reduction in the retraction time. As the specified leakage is relatively small (Q_ext = 20 L/min) for the given cylinder, it has minimal impact on the extension time, as the pressure build-up in chamber A is not affected, showing only a slight shift of around 800 Pa. Figure 3a, showing extension, shows that the pressure profiles in the fault-free (solid line) and faulty (dashed line) states overlap, with differences occurring mostly after the extension. Larger leaks, however, can affect the pressure profile and also delay the extension, as shown later in Section 3.2.

Similarly, an external leak in chamber B leads to a reduced extension time but has a barely detectable effect on the extension time. In addition to increased compressed air consumption, such leaks can also lead to increased impact velocities in the end position, shortening the service life of the cylinder (red circle in Figure 3b, extension).

External leakages between directional control valves and throttle check valves (faults 4 and 5) cause minimal change in movement time, as the quick exhaust effect is barely present in this case, and the pressure profiles are not affected during the cylinder movement due to the small size of the leakage (Figure 4). The fact that they only become noticeable with larger leaks, as shown in Section 3.2, makes their detection more difficult.

Internal (interchamber) leakage and increased friction (fault locations 2 and 6) can each potentially increase both extension time and retraction time (Figure 5 and Figure 6). However, it can be demonstrated that in the case of internal leakage, this increase is pronounced differently due to the surface and thus pressure difference of a double-acting cylinder. The decisive factor affecting the internal leakage’s impact on the movement time is the leakage flow direction, which results from the pressure ratio in the cylinder chambers. When the pressure in chamber A exceeds that in chamber B, the leakage flow moves from A to B (shown as a red area in Figure 5) and vice versa (blue areas in Figure 5).

Figure 5 demonstrates that during extension, the pressure p_A mostly higher than p_B, causing an increase in back pressure due to additional leakage flow into chamber B. However, at the start and end of the movement, p_B can shortly exceed p_A, allowing chamber A to receive extra air through leakage (Leak B→A), briefly accelerating the cylinder.

During retraction, the smaller ring surface of the cylinder piston works against the larger piston surface, requiring p_B to always be greater than p_A to achieve the movement. This means that during retraction, internal leakage always flows from B to A, increasing back pressure in chamber A and significantly slowing the movement.

The changes in the internal leakage flow direction during extension (with multiple possible shifts between B→A and A→B) and the consistent flow direction during retraction (B→A) explain why the extension time may increase less significantly than the retraction time. However, depending on the operating conditions and thus pressure ratio of a differential cylinder, extension may not only be slowed by internal leakage but could also remain unaffected or even accelerate. Section 3.1 explores in greater detail how different pressure conditions influence the extension time.

In the event of an increase in friction, however, the extension and retraction times increase almost equally (Figure 6). This can be used to differentiate between internal leakage and an increase in friction. The measured pressure change in the tested cylinder and the manufacturer’s specification for the DSBC-L-50-200-PPV-A cylinder indicate approximately 20 N of additional friction in the stationary state.

The assumption that movement time variations from internal leakage are due to surface area differences was confirmed by measuring the rodless Festo DGC-18-200-G-PPV-A cylinder with internal leakage (Figure 7). Since both sides have the same working surface, the movement slowed equally due to an internal leakage of Q_int = 40 L/min. This means that for rodless and double-rod cylinders, internal leakage and friction would be indistinguishable if observed solely by deploying the time-based fault detection method.

2.2. Meter-In Throttling

Though less common than meter-out throttling and mainly used for single-acting cylinders, meter-in throttling is important to consider, as it exhibits fault patterns different to those with meter-out throttling. An investigation of five industrial pneumatic machines built between 2019 and 2024 showed that 82.7% of 75 double-acting cylinders used meter-out, while 17.3% used meter-in throttling. The key features of this throttling type include fast cylinder movement but with increased stick-slip probability and load dependency.

Figure 8 reveals that external leakage between the cylinder and throttle check valve significantly slows the movement due to slower pressure build-up in the affected chamber, unlike the meter-out throttled cylinder in Figure 3. The lower back pressure speeds up the subsequent movement (retraction for fault location 1 and extension for fault location 3), though this effect is less pronounced than with meter-out throttling (under 2% vs. ca. 10% reduction).

Figure 9 shows that, similarly to Figure 4, external leakage located between the throttle and the directional control valve in a meter-in controlled system has a negligible effect on cylinder performance.

Internal leakage in a meter-in throttled cylinder occurs in the two phases shown in Figure 10a. During extension, air first flows from chamber B into chamber A, speeding up the pressure build-up in chamber A and allowing the cylinder to start moving sooner. However, once the pressure in chamber A p_A exceeds that in chamber B p_B, the movement slows down as the leakage shifts to A→B. A similar process happens during retraction: Initially, leakage from A to B causes faster pressure build-up in chamber B, followed by a delay when the leakage direction reverses from B to A.

As with exhaust air throttling, increased friction in a meter-in throttled cylinder leads to an increase in the extension and retraction time (Figure 10b).

3. Factors Influencing Cycle Time Deviations

Following the presentation of the influence of different fault locations on the cylinder performance, this section examines the impact of the operating point and leakage rate on movement time. The movement times presented here were measured using SMT-8M-A proximity switches on the Ø25 × 50 Hoerbiger R6025/50 cylinder with a 2 ms PLC sampling rate, matching the standard industry rate used in industrial pneumatic machines by XENON Automatisierungstechnik GmbH in Dresden, Germany (see Section 5).

3.1. Influence of the Operating Point (Pneumatic Frequency Ratio)

To describe the influence of the cylinder’s operating point on the travel times, the pneumatic frequency ratio (PFR) was used. The PFR of the cylinder was varied in two ways: by varying the target movement time t_S at the same load mass m and by varying the load mass

m

at the same time t_S. Figure 11 shows the influence of the PFR on movement times at fault locations from Figure 1, with meter-out and meter-in throttling along with the times for a specific PFR value at a constant load mass as well as the load masses needed to maintain a constant time.

It is apparent that the effect of a leakage placed between the cylinder and the throttle increases with increasing PFR, whereas a leakage between the throttle and the directional control valve shows no operating point dependence for both meter-in and meter-out throttling. External leakage in the locations 1 and 3 impacts the meter-in circuit more significantly (up to 162% slow-down) compared to meter-out (42% speed-up with meter-out at the same leakage rate). Internal leakage also affects meter-in circuits slightly more than meter-out ones.

In terms of increased friction, the movement times of the meter-out circuit remain comparatively the same, while the movement times of the meter-in circuit increase. This can be attributed to the load-independent operating principle of meter-out circuits as opposed to the load-dependence of a meter-in circuit.

As already introduced in Section 2, another important feature of internal leakage in meter-out throttled systems is that the flow direction of the leakage can change depending on the current frequency ratio, which can lead to both a reduction and an increase in the extension time. Figure 12a illustrates the correlation between the extension time change and the PFR, which defines the pressure profiles in cylinder chambers A and B and the flow direction of internal leakage. Three possible pressure profiles are shown on the right-hand side of Figure 12, corresponding to an increase (Figure 12b), unchanged operation despite leakage (Figure 12c), and a reduction in the extension time (Figure 12d).

Using the example of several standard industrial cylinders with different diameters, strokes, and internal leakage rates, it was observed that for lower PFR values below 1.2 … 1.5, the leakage flow direction A→B dominates, increasing the extension time, as demonstrated in Figure 12b. For higher PFR values (typically Ω > 1.2 … 1.5; for the Ø16 × 50 cylinder, Ω > 2), most of the internal leakage flows from B to A, accelerating the cylinder in accordance with Figure 12d. The range between 1 and 1.5, whose pressure profile is shown in Figure 12c, appears to be the balance point where the leakage volumes in both directions are roughly equal (volume ratio below 2), resulting in no change in extension time. These statements are valid for both methods of PFR variation shown with squares and circles in Figure 12a (constant time with variable mass and vice versa).

However, predicting the exact value of the balance point for a specific cylinder without prior simulations might be challenging, as pressure profiles are influenced by many factors, including friction. For example, the transition from reduced to increased extension time occurred at Ω = 1.2 for a low-friction Ø50 × 200 Festo DSBC-L-50-PPV-A cylinder, whereas for a standard-friction Ø50 × 200 Festo DSBC-50-PPV-A cylinder under the same load and leakage conditions, the transition point was at Ω = 1.65.

Another method to predict the impact of internal leakage on the extension time is by considering its average piston speed, which is defined in Figure 13 as the cylinder stroke divided by the target extension time. It was shown that internal leakage has no effect on the extension time when the cylinder operates near its maximum speed. The maximum achievable speed is primarily related to the cylinder stroke, as the measurement results in Figure 13 illustrate. In addition, for cylinders with adjustable end-position cushioning (e.g., Festo’s PPV), the cushioning settings affect pressure profiles and movement time. Testing the Ø20 × 100 Hoerbiger R5020/100 cylinder with and without cushioning showed a shift in the extension time–speed characteristic, as indicated by dashed arrows in Figure 13. However, the general conclusion remains unchanged: As piston speed increases, the impact of internal leakage on extension time decreases.

Overall, the PFR value of a cylinder seems to be a suitable indicator of potential leakage effects on movement time, with minimal differences between mass- and time-based PFR variations. These differences may result from changes in friction at different speeds and loads. The PFR approach was developed for meter-out throttling under the assumption of equal piston areas and non-existing external forces and may cause deviations when applied to meter-in throttling. For internal leakage, movement speed proves to be a more accurate predictor of extension time deviations.

3.2. Influence of Leakage Rate

In a further step, the influence of the leakage rate at different leakage points on the movement time of the cylinder was examined. The selected drive task corresponded to a PFR value of Ω = 1.3 (well-dimensioned cylinder), i.e., a range where the cycle time deviation is less pronounced than with oversized cylinders.

Figure 14 demonstrates that the effects shown so far (e.g., shortened movement time in the meter-out throttled cylinder due to the quick exhaust effect) remain valid across the entire leakage range investigated. With the selected PFR value, both meter-out and meter-in throttling respond to the external leakage between the cylinder and throttle valve. External leakage between the throttle and directional valve in meter-in throttling has no effect on movement time, while exhaust air throttling behaves similarly to leakage between the cylinder and throttle, though less noticeably.

Particularly noteworthy is the retraction time change due to internal leakage. As shown in Figure 5, the delay in retraction with meter-out throttling is associated with the significantly higher pressure p_B: The greater the leakage, the more air flows from B to A, which slows down the movement. With meter-in throttling, on the other hand, the pressures p_A and p_B remain close to each other during the cylinder movement (Figure 9a) so that increased internal leakage has less impact.

Regarding the detection limit, a significant deviation in cycle time (around 5%) due to external leakage between the throttle and cylinder can be detected at approximately 10 L/min for both meter-out and meter-in throttling. However, detecting external leakage between the directional control valve and the throttle based on cycle time is not feasible in a meter-in throttled system. A clear deviation of the movement time during internal leakage occurs quite late with meter-in throttling (at 50 L/min) and earlier with meter-out throttling (25 L/min detectable). It should also be noted that according to Figure 11, cycle time deviations are more pronounced in oversized cylinders (Ω > 2), which are commonly used in industrial machines. In comparison to Figure 14, which illustrates deviations in a well-sized cylinder, this would further simplify fault detection and lower the detection limit.

4. Time-Based Fault Classification Algorithms for Meter-In and Meter-Out Throttling

It can be concluded that the characteristics of the fault locations presented in this paper, which are reflected in the movement time changes, apply equally to different drives; only their severity can vary depending on the operating point and leakage rate. This allows the development of a fault detection and classification strategy based purely on movement times.

The figures in this chapter present the algorithmic structure for implementing such a strategy. Since the error characteristics for meter-out throttling are highly dependent on the PFR, these can be divided into two different algorithms: for well-dimensioned cylinders with Ω ≤ 1.5 (Figure 15) and oversized cylinders with Ω > 1.5 (Figure 16). Figure 17 shows the error detection algorithm for meter-in throttling for all PFR values.

As shown in Figure 15, when operating a well-dimensioned cylinder in the event of external leakage, at least the affected line (piston or rod side) can be clearly localized. In the event of interchamber leakage, cases may occur where both the extension and retraction times increase, especially at Ω ≈ 1, making this fault indistinguishable from increased friction without the use of additional sensors.

For oversized cylinders in Figure 16, not only can the sides affected by external leakage be localized but also their position in the line (before or after the throttle valve). This is possible due to the strong discrepancy between the change in extension time and retraction time in the case of external leakage between the cylinder and the throttle and the absence of such in the case of external leakage between the throttle and the valve (see e.g., Figure 11a–d). The experimental investigations on cylinders at different leakage rates and operating points have shown that this discrepancy is at least 10%, which could be used as an empirical threshold value to differentiate between the states “leakage before throttle” or “leakage after throttle” in drives with oversized cylinders.

Additionally, internal leakage in oversized cylinders has a clear pattern (increase in retraction time and decrease in extension time) that cannot be confused with increased friction, where both travel times increase equally.

With meter-in throttling (Figure 17), only external leaks between the cylinder and the throttle can be detected and localized, as external leaks between the throttle and the directional control valve have no effect on the travel times. These can be localized by excluding other fault locations. Internal leakage and friction show similar patterns (increase in extension time and retraction time) and also cannot be distinguished from each other.

To further improve the time-based method, a volume flow sensor can be used in cases where no differentiation is possible (e.g., internal leakage and increased friction with meter-in throttling). The additional consumption-based differentiation is included in the algorithms at the necessary points.

To evaluate the applicability of the algorithms in an industrial environment, the following section presents the results of investigations into the detection limit and exemplary testing on industrial systems.

5. Field Trial

5.1. Cycle Time Fluctuations in Fault-Free Operation

The question of setting a detection limit and the extent to which the travel times of individual actuators in a pneumatic system fluctuate during fault-free operation has so far remained unanswered. To address this, a 24 h test was conducted on the exemplary pneumatic handling system shown in Figure 18a, placed in a standard industrial machine hall. This system moves a work piece according to the sequence in Figure 18b and consists of a horizontal (Festo SLG-12-200-YSR-A) and a vertical (Festo DGSL-10-100-P1A) pneumatic slide actuators with meter-out throttling, a swivel drive, a vacuum suction gripper and an electromechanically operated slide.

The travel times of the slide actuators were recorded directly in the PLC CECC-D using CODESYS v3.5 programming environment and the supply pressure and room temperature using the OMB-DAQ-2416 measurement module.

The measurement results in Figure 19 demonstrated the following:

Cycle time fluctuations remained within ±2% throughout the measurement. A tolerance band can therefore be established to prevent false-error alarms. The deviations of ca. 5% shown in Figure 11c,d, corresponding to external leaks between the throttle and directional control valve, could therefore have been detected under these conditions;
The supply pressure of the machine p_S remained unchanged over the entire duration of the test. Slight pressure fluctuations might have potentially caused the variations in cycle time but do not appear to be the main reason for these;
A stronger correlation was observed between cycle times and room temperature. As the temperature increased, the actuators moved faster; when the temperature dropped, the movement slowed. This aligns with research by Pham et al. [38] showing that the friction force of pneumatic actuators decreases noticeably as ambient temperature rises. The observed cycle time fluctuations can therefore be attributed to the variation of the friction force.

The correlation with temperature is less pronounced during the first 6 h, likely due to the run-in behavior of the actuators. It is therefore advisable to wait for a run-in period before assessing movement time reliably when starting from a standstill. The horizontal actuator’s hydraulic shock absorbers (YSRG), with temperature-dependent damping, might have also affected movement time in the end position.

In order to make a reliable statement regarding the potential width of the tolerance band for movement times, further long-term measurements under various boundary conditions, such as seasonal temperature and humidity fluctuations, hall vibrations, and pressure fluctuations in the distribution network, are necessary. It was not possible to map this diversity within the scope of this study; rather, the results provide a representative statement for the fault-free operation of a conventional pneumatic system on one day. It is to be expected that the tolerance band must be set higher in an industrial environment to cover a wide range of temperature and thus friction force and other fluctuations.

5.2. Test on Industrial Machine

Continuous monitoring of the movement times of drives using PLC signals was implemented and tested on a machine by XENON Automatisierungstechnik GmbH for the fully automated assembly of a fuel cell stack, shown in Figure 20a. The machine consists of several pneumatic cylinders which, for example, ensure the vertical feeding of the stacks or actuate various mechanisms such as gas flaps. Two exemplary pneumatic cylinders with meter-out throttling were selected for testing, one of which is shown in Figure 20b. Table 3 summarizes the special features of the investigated cylinders.

For each cylinder, external leakages were generated at the piston side between the cylinder and throttle (point 1 from Figure 1; marked with circles in Figure 21) and between the throttle and directional valve (point 4 from Figure 1; triangles in Figure 21). Fault detection algorithms were then applied: Figure 16’s algorithm for the oversized cylinder 1 and Figure 15’s for the well-sized cylinder 2. In both cases, the affected supply air line was identified by reduced retraction time, but only for cylinder 1 could the leakage location be distinguished (between the cylinder and throttle or the valve and throttle).

6. Discussion

The study demonstrated that a purely time-based method can detect and localize various fault locations in a pneumatic drive system without need for additional sensors. Fault detection occurs during normal operation without interruptions or special measurement equipment. Signals can be evaluated directly in the PLC with a standard sampling rate (2 ms in the examples), enabling real-time detection of performance deviations. This cost-free implementation makes the method highly applicable in any industrial setting that involves PLC-based pneumatic machines comprising actuators with proximity (reed) switches, e.g., for handling technology, assembly automation, packaging, printing, food industry, etc. Using this method can reduce condition monitoring costs, as it does not require additional sensors at various points (e.g., for each valve terminal). In this way, the availability of systems can be enhanced, preventing costly failures and increasing overall productivity.

The method can be implemented as an algorithm on a PC or other device with the following steps:

Categorizing of the drives within the system (type of throttling; for meter-out: determination of the PFR value) in order to select the appropriate fault identification algorithm according to Figure 15, Figure 16 and Figure 17;
Establishment of an error-free reference by measuring movement times, preferably over several hours;
Application of a tolerance band to avoid false alarms. The test in Section 5.1 showed temperature-related fluctuations of ±1.5–2% in movement times, but a wider tolerance (e.g., ±10%) is recommended to account for seasonal or other variations;
Continuous or periodic monitoring of movement times and algorithm activation upon deviations exceeding the tolerance band, as shown in Figure 15, Figure 16 and Figure 17, for fault identification;
If fault locations are unclear, additional volume flow measurements are advisable. Internal leakage and increased friction in a meter-in throttled system both cause increased extension and retraction times, making fault differentiation difficult. Similarly, with meter-out throttling, distinguishing between these two faults can be challenging depending on the operating point.

The fault features presented here strongly depend on the operating point and throttling type, which was not considered in the previous studies. A generally applicable detection limit and the accuracy of the method are therefore difficult to define. The sensitivity of the method depends on the tolerance band defined by the user, with its size influenced by various factors such as pressure and ambient temperature fluctuations. Further long-term studies are needed to refine the accuracy, so the values presented here serve only as an initial recommendation. However, it could be demonstrated that most leaks can be detected from a leakage size of 20 L/min, which, according to [40], corresponds to a medium-sized leakage with moderate need for action. Even with well-dimensioned cylinders (Ω < 1.5), an external leakage of 20 L/min between the cylinder and the throttle was found to cause an easily detectable 20–50% deviation, depending on the pneumatic frequency ratio PFR and throttling type. Smaller deviations occur with leakages between the directional control valve and throttle, detectable only at larger leakage rates.

In terms of changed friction, the proposed method was able to detect a friction force increase of ca. 20 N. Moreover, movement time deviations due to faults become more significant with larger cylinder oversizing (expressed via PFR Ω), particularly when Ω > 2. Since industrial systems often feature oversized cylinders (due to safety factors or stock availability), time-based fault detection and localization become more feasible.

The simultaneous occurrence of several faults limits the applicability of the algorithm. However, it is at least possible to recognize the condition that several faults are present at the same time.

The influence of the cylinder’s alignment and external forces such as weight or pressing forces on fault characteristics also requires further investigation, as this study only considered horizontally mounted cylinders and a limited number of vertical ones. Additionally, the applicability of the investigated fault characteristics to rotary drives such as pneumatic motors and swivel drives can also be evaluated.

7. Conclusions

In this paper, the effects of five potential leakage points, which represent different potential fault areas within a pneumatic drive, and increased friction on cylinder performance were investigated experimentally for both meter-in and meter-out throttling at different operating points. For each fault location considered, unique movement time deviations were identified, forming the basis for a cycle time-based fault detection and localization method in the algorithmic form. It was demonstrated that it is possible to localize errors using the developed algorithms.

In contrast to the previous publications [13,31,33,34,35], this research provides a comprehensive overview of various fault locations within a pneumatic actuator and considers both types of throttling. Most tests were conducted on a test rig using the Ø32 × 200 Festo DSBC-32-200-PPV-A and Ø25 × 50 Hoerbiger R6025/50 differential pneumatic cylinders, representative of industry-standard sizes. The features of internal leakage with meter-out throttling as well its effect on cylinders with identical piston surfaces, which had not been previously investigated, are also presented. In addition, the method was field-tested on an industrial pneumatic machine.

The proposed method can be easily implemented in any pneumatic system without additional hardware costs. The fault detection and localization can be performed directly in the PLC (e.g., through continuous cycle recording) or via other means (e.g., monitoring limit switch signals with external measurement tools). To enhance its effectiveness, the method can be combined with other technologies (e.g., total volume flow measurement), enabling the detection of failures that do not affect the cylinder’s movement time. Simultaneous faults at multiple points may also limit time-based localization, potentially requiring additional sensors. Nonetheless, movement time proves to be a valuable indicator of system status with significant potential for broader application, especially when combined with other signals.

Author Contributions

Conceptualization, formal analysis, investigation, validation, visualization, and writing—original draft preparation, V.B.; project administration, supervision, and writing—review and editing, J.W. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by the German Federal Ministry for Economic Affairs and Climate Action (BMWK), grant number 03EN4050A (OekEffDatA: Data-driven System Monitoring and Design for Ecological Efficiency in Operation).

Data Availability Statement

The original contributions presented in the study are included in the article; further inquiries can be directed to the corresponding author.

Conflicts of Interest

The authors declare no conflicts of interest.

References

Radermacher, T.; Merx, M.; Sitte, A.; Boyko, V.; Unger, M. Potenzialstudie Energie-/Kosteneinsparung in der Fluidtechnik. Final Report No. 37EV181030. TU Dresden, University of Stuttgart, Germany. 2021. Available online: https://www.umweltbundesamt.de/publikationen/potenzialstudie-energie-kosteneinsparung-in-der (accessed on 27 September 2024).
Energy Efficiency in Production in the Drive and Handling Technology Field (EnEffAH). Final Report No. 0327484A-E. Available online: https://www.eneffah.de/EnEffAH_Broschuere_engl.pdf (accessed on 27 September 2024).
Gauchel, W. Energiesparende Pneumatik. Konstruktive sowie schaltungs- und regelungstechnische Ansätze. O+P Fluidtechnik 2006, 1, 33–39. [Google Scholar]
Saidur, R.; Rahim, N.A.; Hasanuzzaman, M. A review on compressed-air energy use and energy savings. Renew. Sustain. Energy Rev. 2010, 14, 1135–1153. [Google Scholar] [CrossRef]
Šešlija, D.D.; Milenković, I.M.; Dudić, S.P.; Šulc, J.I. Improving Energy Efficiency in Compressed Air Systems. Practical Examples. Therm. Sci. 2016, 20 (Suppl. S2), 335–370. [Google Scholar] [CrossRef]
Hepke, J. Energetische Untersuchung und Verbesserung der Antriebstechnik pneumatischer Handhabungssysteme. Ph.D. Thesis, TU Dresden, Dresden, Germany, 2016. [Google Scholar]
Gryboś, D.; Leszczyński, J.S. A Review of Energy Overconsumption Reduction Methods in the Utilization Stage in Compressed Air Systems. Energies 2024, 17, 1495. [Google Scholar] [CrossRef]
Boyko, V.; Nazarov, F.; Gauchel, W.; Neumann, R.; Doll, M.; Weber, J. Comprehensive Application-Based Analysis of Energy-Saving Measures in Pneumatics. Int. J. Fluid Power 2024, 1, 27–58. [Google Scholar] [CrossRef]
Dindorf, R. Estimating Potential Energy Savings in Compressed Air Systems. Procedia Eng. 2012, 39, 204–211. [Google Scholar] [CrossRef]
McKane, A.; Hasanbeigi, A. Motor systems energy efficiency supply curves: A methodology for assessing the energy efficiency potential of industrial motor systems. Energy Policy 2011, 39, 6595–6607. [Google Scholar] [CrossRef]
Harris, P.; O’Donnell, G.; Whelan, T. Energy Efficiency in Pneumatic Production Systems: State of the Art and Future Directions. In Proceedings of the 19th CIRP International Conference on Life Cycle Engineering, Berkeley, CA, USA, 23–25 May 2012. [Google Scholar]
Meindorf, T.; Stammen, C.; Murrenhoff, H. Online Condition Monitoring (OCM) in Fluid Power Technology. In Proceedings of the 4th International Fluid Power Conference (4. IFK), Dresden, Germany, 25–26 March 2004. [Google Scholar]
Fritz, J. Verfahren zur Erkennung sowie Diagnose von Fehlern in pneumatischen Systemen und Komponenten. Ph.D. Thesis, RWTH Aachen, Aachen, Germany, 2011. [Google Scholar]
Festo SE & Co. KG. Leckageortung und -Beseitigung. Energy Saving Services von Festo; Festo: Esslingen, Germany, 2018. [Google Scholar]
Belforte, G.; Mazza, L.; Raparelli, T. Analysis of Typical Failure Situations in Pneumatic Cylinders under Load. Lubrictation Eng. 1992, 48, 840–845. [Google Scholar]
Mallia, J.; Francalanza, E.; Xuereb, P.; Refalo, P. Intelligent Approaches for Anomaly Detection in Compressed Air Systems: A Systematic Review. Machines 2023, 11, 750. [Google Scholar] [CrossRef]
Borg, M.; Refalo, P.; Francalanza, E. Failure Detection Techniques on the Demand Side of Smart and Sustainable Compressed Air Systems: A Systematic Review. Energies 2023, 16, 3188. [Google Scholar] [CrossRef]
Wang, Z.; Yang, B.; Ma, Q.; Wang, H.; Carriveau, R.; Ting, D.S.-K.; Xiong, W. Facilitating Energy Monitoring and Fault Diagnosis of Pneumatic Cylinders with Exergy and Machine Learning. Int. J. Fluid Power 2023, 24, 643–682. [Google Scholar] [CrossRef]
Hülsmann, S.; Volk, R.; Boyko, V.; Raisch, A.; Weber, J.; Sawodny, O. Energieeffiziente Antriebstechnik in der Produktion—Herausforderungen bei der Umsetzung. O+P Fluidtechnik 2020, 9, 38–41. [Google Scholar]
Festo SE & Co. KG. Weltneuheit Festo Motion Terminal VTEM; Festo: Esslingen, Germany, 2017. [Google Scholar]
Energieeffizienzmodul MSE6. 2024. Available online: https://www.festo.com/media/pim/728/D15000100122728.PDF (accessed on 27 September 2024).
AVENTICS. Smart Pneumatics Opportunities for Industry Through Intelligent Pneumatics. Available online: https://www.cmafh.com/enewsletter/PDFs/Smart-Pneumatics_IoT_R500000017.pdf (accessed on 27 September 2024).
Digital Ultrasonic Testing Device SONAPHONE for Preventive Maintenance by Sonotec GmbH. Available online: https://www.sonotec.eu/fileadmin/user_upload/business_units/2-preventive-maintenance/products/sonaphone/flyer-sonaphone-en-sonotec_01.pdf (accessed on 27 September 2024).
Czopek, D.; Gryboś, D.; Leszczyński, J.; Wiciak, J. Identification of energy wastes through sound analysis in compressed air systems. Energy 2021, 239, 122122. [Google Scholar] [CrossRef]
Mahmoud, H.; Vlašic, F.; Mazal, P.; Jana, M. Leakage analysis of pneumatic cylinders using acoustic emission. Insight-Non-Destr. Test. Cond. Monit. 2017, 59, 500–505. [Google Scholar] [CrossRef]
Mahmoud, H. Diagnosis of Pneumatic Cylinders Using Acoustic Emission Methods. Ph.D. Thesis, Brno University of Technology, Brno, Czech Republic, 2017. [Google Scholar]
Zhang, S.; Asakura, T.; Hayashi, S. Gas Leakage Fault Detection of Pneumatic Pipe System Using Neural Networks. JSME Int. J. Ser. C Mech. Syst. Mach. Elem. Manuf. 2004, 47, 568–573. [Google Scholar] [CrossRef]
Johnson, D.; Kirner, J.; Grollmisch, S.; Liebetrau, J. Compressed Air Leakage Detection Using Acoustic Emissions with Neural Networks. In Proceedings of the INTER-NOISE and NOISE-CON Congress and Conference Proceedings, InterNoise20, Seoul, Republic of Korea, 23–26 August 2020; pp. 5662–5673. [Google Scholar]
Muth, A. Reibkraftermittlung an pneumatischen Ventilen und Zylindern. Ph.D. Thesis, TU Dresden, Dresden, Germany, 1997. [Google Scholar]
Vygantas, A.; Saunoris, M. Investigation of high frequency vibrations of pneumatic cylinders. Ultragarsas 2004, 51, 21–26. [Google Scholar]
González, R.G. Development of a Diagnostic Concept for Pneumatics Systems and Components. Ph.D. Thesis, RWTH Aachen, Aachen, Germany, 2012. [Google Scholar]
Kosturkov, R.; Nachev, V.; Titova, T. Diagnosis of Pneumatic Systems on Basis of Time Series and Generalized Feature for Comparison with Standards for Normal Working Condition. TEM J. 2021, 10, 183–191. [Google Scholar] [CrossRef]
Gauchel, W.; Streichert, T.; Wilhelm, Y. Predictive maintenance with a minimum of sensors using pneumatic clamps as an example. In Proceedings of the 12th International Fluid Power Conference, Dresden, Germany, 9–11 March 2020; pp. 175–183. [Google Scholar] [CrossRef]
Borg, M.; Refalo, P.; Francalanza, E. Pneumatic Fault Monitoring and Control for Sustainable Compressed Air Systems. Procedia CIRP 2024, 122, 217–222. [Google Scholar] [CrossRef]
Borg, M.; Refalo, P.; Francalanza, E. Autonomous Fault Monitoring for Efficient Multi-Actuator Compressed Air Systems:Data Analytics of Demand-Oriented Parameters. Procedia Comput. Sci. 2024, 232, 783–793. [Google Scholar] [CrossRef]
Doll, M.; Neumann, R.; Sawodny, O. Dimensioning of pneumatic cylinders for motion tasks. Int. J. Fluid Power 2015, 16, 11–24. [Google Scholar] [CrossRef]
Doll, M. Optimierungsbasierte Strategien zur Steigerung der Energieeffizienz pneumatischer Antriebe. Ph.D. Thesis, University of Stuttgart, Stuttgart, Germany, 2015. [Google Scholar]
Pham, V.-H.; Nguyen, T.-D.; Bui, T.-A. Behavior of Friction in Pneumatic Cylinders with Different Relative Humidity. Tribol. Ind. 2020, 42, 400–406. [Google Scholar] [CrossRef]
Assembly Automation for Sustainable Energy Solutions by XENON Automatisierungstechnik GmbH. Available online: https://www.xenon-automation.com/products/energy/fuel-cells-electrolyzers (accessed on 27 September 2024).
Unger, M. Systematische Analyse von Druckluftleckagen. Ph.D. Thesis, University of Stuttgart, Stuttgart, Germany, 2021. [Google Scholar] [CrossRef]

Figure 1. Potential fault locations within a pneumatic drive (symbolic depiction): (1) external leakage in piston-side chamber A; (2) internal (interchamber) leakage; (3) external leakage in rod-side chamber B; (4 and 5) external leakages between the directional valve and the throttle valves; (6) increased friction.

Figure 2. Test bench for investigating cycle time-based fault detection with corresponding fault locations.

Figure 3. Influence of external leakage Q_ext = 20 L/min between cylinder and throttle check valve with meter-out throttling on cylinder pressure and position: (a) piston side A (fault location 1); (b) rod side B (fault location 3).

Figure 4. Influence of external leakage Q_ext = 20 L/min between directional control valve and throttle check valve with meter-out throttling on cylinder pressure and position: (a) piston side A (fault location 4); (b) rod side B (fault location 5).

Figure 5. Influence of internal leakage Q_int = 40 L/min in the cylinder (fault location 3) on the meter-out throttled cylinder’s pressure and position.

Figure 6. Influence of increased piston friction F_add,fr = 20 N (fault location 6) on the meter-out throttled cylinder’s pressure and position.

Figure 7. Equal delay in the extension and retraction time in case of internal leakage Q_int = 40 L/min in the rodless cylinder Festo DGC-18-200-G-PPV-A with meter-out throttling.

Figure 8. Influence of external leakage Q_ext = 20 L/min between cylinder and throttle check valve with meter-in throttling on cylinder pressure and position: (a) piston side A (fault location 1); (b) rod side B (fault location 3).

Figure 9. Influence of external leakage between directional control valve and throttle check valve with meter-in throttling on cylinder pressure and position: (a) piston side A (fault location 4); (b) rod side B (fault location 5).

Figure 10. Influence of (a) internal leakage Q_int = 30 L/min in the cylinder (fault location 3) and (b) increased piston friction F_add,fr = 20 N (fault location 6) on the meter-in throttled cylinder’s pressure and position.

Figure 11. Influence of the pneumatic frequency ratio Ω on changes in the movement time of the Ø25 × 50 Hoerbiger R6025/50 differential pneumatic cylinder for different fault locations and at constant fault values. Meter-out throttling: (a) fault location 1; (b) fault location 3; (c) fault location 4; (d) fault location 5; (e) fault location 2; (f) fault location 6. Meter-in throttling: (g) fault location 1; (h) fault location 3; (i) fault location 4; (j) fault location 5; (k) fault location 2; (l) fault location 6.

Figure 12. (a) Influence of internal leakage on the extension time change of different pneumatic differential cylinders with meter-out throttling as a function of the pneumatic frequency ratio Ω, changes in pressure and position profiles, and interchamber flow direction in (b) well-sized cylinder with Ω < 1.5; (c) well-sized cylinder with Ω = 1.5; (d) oversized cylinder with Ω > 1.5.

Figure 13. Influence of internal leakage on the extension time change of different pneumatic differential cylinders with meter-out throttling as a function of the mean piston velocity

\dot{x}

Figure 13. Influence of internal leakage on the extension time change of different pneumatic differential cylinders with meter-out throttling as a function of the mean piston velocity

\dot{x}

Figure 14. Change in the movement time of the Ø25 × 50 Hoerbiger R6025/50 differential pneumatic cylinder at Ω = 1.3 as a function of leakage rate. Meter-out throttling: (a) fault location 1; (b) fault location 3; (c) fault location 4; (d) fault location 5; (e) fault location 2. Meter-in throttling: (f) fault location 1; (g) fault location 3; (h) fault location 4; (i) fault location 5; (j) fault location 2.

Figure 15. Algorithm for time-based fault detection and localization in pneumatic drives with well-sized, double-acting differential cylinders (PFR Ω ≤ 1.5) and meter-out throttling.

Figure 16. Algorithm for time-based fault detection and localization in pneumatic drives with oversized, double-acting differential cylinders (PFR Ω > 1.5) and meter-out throttling.

Figure 17. Algorithm for time-based fault detection and localization in pneumatic drives with double-acting differential cylinders (all PFR values) and meter-in throttling.

Figure 18. Handling system (a) and its motion sequence (b).

Figure 19. Changes in cycle time of pneumatic actuators and supply pressure of the handling system as well as room temperature during fault-free operation.

Figure 20. (a) Pneumatic system for fully automated assembly of a fuel cell stack by XENON Automatisierungstechnik GmbH [39]; (b) external leakage generation at cylinder 1 (flap actuator).

Figure 21. Above: Changes in cycle time of pneumatic actuators of the assembly machine resulting from external piston-side leakage in chamber A before and after the throttle valve; below: corresponding fault recognition paths of the algorithms.

Table 1. Typical faults in a linear pneumatic drive.

Fault	Definition and Examples
External leakage	Air leakage from the actuator into the environment. Possible causes: Mechanical damage to the hose or hose aging; Damaged/worn cylinder wiper; Loosened/damaged push-in fittings.
Internal leakage	Interchamber leakage. Possible causes: Damaged/worn cylinder piston sealing.
Increased friction	Additional external force. Possible causes: Damaged/worn cylinder piston sealing and/or wiper; Changed friction conditions on mechanical components (e.g., guide wear or contamination).

Table 2. Operating conditions of the cylinder.

Parameter	Value
Supply pressure p_S (bar abs.)	7
Stroke h (mm)	200
Valve–cylinder distance l_vc (m), PUN-8 tubing	0.5
Extension time t_e (s)	0.290
Retraction time t_r (s)	0.380
Moving mass m (kg)	11
PFR, extension: Ω_e (-)	1.3
PFR, retraction: Ω_r (-)	1.7

Table 3. Selected cylinders of the industrial demonstrator.

Parameter	Cylinder 1 (Oversized)	Cylinder 2 (Well-Sized)
Type	Festo DSNU-10-50-P-A	Festo DFM-16-20-P-A-KF
Task	Flap actuation	Fuel stack feeding
Throttling	Meter-out
Alignment	Vertical, rod upwards
Supply pressure p_S (bar abs.)	6
Valve–cylinder distance l_vc (m)	2 (PUN-4 tubing)	9 (PUN-4 tubing)
Extension time t_e (ms)	267	99
Retraction time t_r (ms)	53	129
Moving mass m (kg)	0.025	0.23
PFR, extension: Ω_e (-)	6.4	1.0
PFR, retraction: Ω_r (-)	1.3	1.3

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

© 2024 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).

Share and Cite

MDPI and ACS Style

Boyko, V.; Weber, J. Cycle Time-Based Fault Detection and Localization in Pneumatic Drive Systems. Actuators 2024, 13, 447. https://doi.org/10.3390/act13110447

AMA Style

Boyko V, Weber J. Cycle Time-Based Fault Detection and Localization in Pneumatic Drive Systems. Actuators. 2024; 13(11):447. https://doi.org/10.3390/act13110447

Chicago/Turabian Style

Boyko, Vladimir, and Jürgen Weber. 2024. "Cycle Time-Based Fault Detection and Localization in Pneumatic Drive Systems" Actuators 13, no. 11: 447. https://doi.org/10.3390/act13110447

APA Style

Boyko, V., & Weber, J. (2024). Cycle Time-Based Fault Detection and Localization in Pneumatic Drive Systems. Actuators, 13(11), 447. https://doi.org/10.3390/act13110447

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Menu