Railways Passenger Comfort/Discomfort: Objective Evaluation

Abstract

Railways are one of the most used public transportation modes. Due to the low environmental impact, governments are promoting their use as a solution for mass transportation for middle-distance. Critical parameters to satisfy passengers and attract new ones are related to comfort, safety, and user conditions. Comfort is classified by dynamic and static methods, ideally combining both. The former is based on the comfort evaluation in the presence of vibration, whilst the latter is applied to standstill situations. Vibration, derived from the train motion, will be presented throughout the trip; thus, it is a major concern. For the same reason, assessing vibration levels associated with comfort gives important information about the train’s condition and its major maintenance or repair needs. Vibration transmissibility allows the study of the vibration frequency that is being transmitted to the user. In this chapter, the complete seat structure and the passenger’s comfort are evaluated combined with the interface pressure. This chapter presents the commonly used methodologies and results from the tests conducted on railway environment whilst running passenger service. These tests report the seat comfort evaluation and a new methodology to identify train and rail maintenance needs based on comfort analysis.

1. Introduction

World population is increasing, and therefore the needs for mass transportation. Nowadays, railways are one of the most widely used public transportation systems, mainly because of their safety, superior transportation capacity, reduced boarding time, and the possibility of making better use of the journey time to work or enjoy train facilities. Additionally, trains have been revealed to be the most sustainable transportation system. In Europe, the predominant proportion of rail services are operated by electrified trains; thus, its CO₂ emissions are residual. The emissions per passenger-kilometer are much lower for the train than for air or car transportation [1, 2, 3].

Due to its low environmental impact, multiple governments are promoting rail use as the major mass transportation system, especially for middle-range distances. Based on this proposal, train passengers have continuously increased from 2013 until the 2020 COVID pandemic. Between 2015 and 2019, a historic 4000 billion passenger-kilometers were recorded worldwide. However, during the pandemic years, due to multiple lockdowns, train passenger numbers drastically decreased [4, 5, 6]. Currently, those numbers are even higher than the pre-pandemic ones [7].

The quality of public transportation services influences travelers’ choices. Passengers with previous good travel experiences will probably use the same travel transportation mode again. On the other hand, customers that experienced problems with the journey service may change to a different transportation mode for their next trip. Therefore, to keep increasing continuously the number of passengers, it is crucial to raise trains’ attractiveness and provide comfortable journeys [8, 9]. Those are defined based on safety, comfort, and user conditions [10, 11]. In particular, the seat and vibration are strongly linked with discomfort. Passengers spend most of their travel time seated; thus, vibration, derived from train motion and train-rail interaction, is transmitted to the user through this surface, being classified as whole-body vibration (WBV). A specific feature of a seat is its dimensions which must be able to accommodate people with different anthropometric characteristics, and simultaneously provide low fatigue levels and a general feeling of comfort [12]. Although affecting passengers’ comfort, vibration is also linked to safety. High safety levels are ensured through adequate maintenance. That will guarantee the reliability and longevity of the rail, which needs to provide a stable and safe platform for train operation. Corrective and preventive interventions are accomplished during track maintenance. Its main goal is to preserve the system’s functions and prevent its breakdown or failure [13, 14, 15, 16].

Comfort and discomfort are two concepts not consensually defined in the literature. Hence multiple interpretations can be found. However, it is well established that comfort should be evaluated based on objective and subjective methods, ideally a combination of both. Objective methods are derived from mechanical approaches; thus, these methods quantitatively define the physical variables of comfort [17]. In opposition, subjective evaluation methods rate users’ feelings and, this way, they quantify the psychological impact of comfort on the passengers based on questionnaires and rating scales [18].

Comfort parameters are divided into “Motion” and “Non-Motion” factors. The latter is characterized by issues such as noise, smell, illumination, humidity, and temperature. To properly classify “Motion” parameters, it is essential to conduct both dynamic (in the presence of vibration) and static (absence of vibration) tests. The former is evaluated by ride comfort, seat effective transmissibility (SEAT) and transmissibility tests. The latter is mainly defined by interface pressure measurements [10]. Seats may present good dynamic behavior but poor static performance. The ideal seat has a combination of optimum dynamic properties to minimize unwanted vibration and the best static behavior to equally distribute pressure at the seat surface and, this way, reduce the interface pressure.

The present chapter intends to introduce a new methodology to identify train and rail track infrastructure sections’ maintenance requirements based on ride comfort analysis. Moreover, transmissibility and interface pressure experiments are conducted as complementary comfort analysis methods. Vibration transmissibility allows the study of the vibration frequency that is being transmitted to the user. If associated in combination with the interface pressure (static analysis), it is possible to define the complete seat structure and conclude about passengers’ comfort. The chapter is organized into seven sections to report on the aforementioned goals. Multiple comfort and discomfort definitions will be presented, and the one used in this research will be detailed. Based on vibration analysis, the ride quality evaluation methods are defined, leading to the rail vehicle and track infrastructure maintenance needs identification. Finally, transmissibility and interface pressure experiments are described leading to the final passengers’ comfort evaluation and seat structure analysis.

2. Comfort and discomfort definition

Whether for pleasure or work, people want to be comfortable. Thus, when designing a seat, it is vital to start by analyzing the comfort that it will provide to the users. Therefore, understanding how comfort and discomfort are defined and evaluating them, is a major assignment. These two parameters do not have a consensus; hence multiple definitions and interpretations for comfort and discomfort can be found in the literature [19].

In 1958, Hertzberg defined comfort as the absence of discomfort. Hertzberg also postulated that comfort and discomfort could not coincide, so discomfort is not present when someone feels comfort [20]. Based on that statement, Shackel et al. and, later, Richards developed a concept where comfort and discomfort were defined as two states placed on opposite extremities of a linear scale [19, 21, 22].

However, in 1992 and 1996, Zhang and Helander and Zhang et al. developed a work that ostracized the previous linear concept and introduced the non-linear model of comfort and discomfort. Zhang et al. defined comfort and discomfort as independent, individual concepts associated with different underlying factors. Comfort is related to feelings of well-being and relaxation, and the esthetic impression of a product or environment influences it. In opposition, discomfort is associated with pain, soreness, numbness, and stiffness and is influenced by the product’s physical constraints [23, 24]. In 2003, as a complement to his previous work, Herlander stated that comfort is an emotional state, whilst discomfort is a physical state of being [25].

De Looze et al. built a theoretical model that illustrates the non-linear relationship between comfort and discomfort and its relationship with the physical product. This model also distinguishes the three categories influencing comfort/discomfort assessment: the human, the product, and the environment [26, 27, 28].

In this model, it is possible to observe the differences between discomfort (left side of the model) and comfort (right side of the model). Also, it concludes that discomfort has a dominant effect on comfort. De Looze indicated that discomfort levels are influenced by the interaction of the human’s physical capacity, like weight, physiological processes, muscle activation, body temperature, intradiscal pressure or nerve conduction. Relatively to the comfort side, it is highly influenced by the user’s emotions, expectations, and esthetic design. Regarding the rail context, the emotions and expectations of the user (human level), as well as the train temperature (environment level), can influence the passenger’s perception of comfort/discomfort (product level) [19, 26, 27].

In 2005, Moes introduced a different model concerning sitting discomfort. This model defines four initial parameters: the person, the seat, the purpose, and the usage. Then, five steps are taken before discomfort, interaction, effect on the internal body, perceived effects, appreciation of the effects and, finally, discomfort. The author also states that if a person is seated for a specific purpose, the interaction occurs.

For a train journey, the interaction can result in higher-pressure distribution in the seat interface. This pressure will lead to internal body effects, like nerve compression, which can change the user perceived effects and its appreciation and, consequently, lead to discomfort [26, 29].

The previous methods have different backgrounds and approaches but are unanimous in framing comfort as a combination of physical and psychological factors. Whilst the physical state can be influenced by external stimuli such as noise, temperature, or vibration, the psychological state is affected by the previous user experience and expectations [29, 30, 31]. Therefore, to be able to evaluate both, physical and psychological impact on the users’ comfort, objective and subjective methods should be considered. The present research is developed following the De Looze model. Therefore, the interaction between the product, the user, and the environment is considered and evaluated.

3. Ride comfort evaluation

Ride comfort, also called long-term comfort, of rail vehicles is affected by multiple parameters such as vibration, noise, temperature, smell, and visual stimuli. Derived from wheel-track interaction and rail motion, vibration is the critical factor affecting users’ comfort and health, limiting their performance. Therefore, evaluating the vibration transmission in a rail environment is fundamental to quantify passengers’ comfort levels and assess the harmful consequences of vibration [31, 32, 33, 34, 35].

Three main standards are fully dedicated and employed to evaluating passengers’ comfort based on vibration analysis, ISO 2631, EN 12299, and Sperling ride index. The human body has its natural vibration mode, which affects the human vibration feeling. When this mode matches an externally induced vibration, resonance may occur, which, if absorbed, can lead to the physical stress of tissues and organs [10, 33, 35, 36, 37, 38, 39]. Furthermore, depending on the human tissue’s characteristics, vibration with similar intensities but different spectral content will induce different dynamic responses in the human body. Thus, acceleration needs to be weighted based on human body sensitivity to obtain an index that can reflect the vibration feeling. Although different methodologies exist, the three methods mentioned above share the application of frequency weighting curves. Those capable of producing the highest effect are ranked with the highest weight, and other frequencies are attenuated based on their relative importance [10, 40]. All methods state acceleration measurements in three directions: vertical, fore-and-aft, and lateral.

It should be highlighted that, besides sharing the same assumptions and goals, the methods use different calculation techniques to evaluate comfort. Therefore, one method cannot be transformed into another just by analyzing its results. Instead, a complete analysis and correlation amongst indexes need to be performed [10]. ISO 2631, EN 12299, and Sperling ride index methods and their vibration evaluation techniques will be given. The present work followed the ISO 2631 standard approach. Thus, it is used as a reference methodology to identify rail track infrastructure sections and train maintenance needs will be demonstrated.

3.1 ISO 2631 comfort evaluation

ISO 2631 standard quantifies WBV regarding comfort, human health, and motion sickness. Comfort and health evaluations are related in many ways; thus, frequencies between 0.5 and 80 Hz are defined as the most relevant ones since, at this range, vibration affects the body as a whole, which can lead to discomfort and fatigue.

Acceleration measurements should occur on the vibration transmission interfaces: floor, seat surface, and seatback. Then, the root-mean-square (rms) acceleration is calculated for each axis and the corresponding weighting curve is applied [36, 37, 38, 40]. The weighting process is calculated according to Eq. (1):

$$a_{\mathrm{w}}={\left[\sum {\left(W_i{a}_i\right)}^2\right]}^{\raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$2$}\right.}$$E1

where W_i represents the weighting frequencies and a_i the rms accelerations. Weighting curves application depends on the measurement location and purpose. The total vibration (a_v) is achieved following Eq. (2):

$$a_{\mathrm{v}}={\left(k_x^2{a}_{\mathrm{w}x}^2+k_y^2{a}_{\mathrm{w}y}^2+k_z^2{a}_{\mathrm{w}z}^2\right)}^{\raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$2$}\right.}$$E2

where aw are the rms accelerations for each axis, and k represents the multiplying factor dependent on the measuring position, presented in Table 1.

	X-axis	Y-axis	Z-axis
Floor	Wk and kx = 0.25	Wk and ky = 0.25	Wk and kz = 0.40
Seat surface	Wd and kx = 1.0	Wd and ky = 1.0	Wk and kz = 1.0
Seatback	Wc and kx = 0.80	Wd and ky = 0.50	Wd and kz = 0.40

Table 1.

Frequency weighting curves and multiplying factors defined by ISO 2631 for comfort analysis of a seated passenger.

Finally, based on $a_{\mathrm{v}}$, the discomfort is evaluated by a defined scale, Table 2, where accelerations higher than 0.315 m/s² are ranked as uncomfortable.

$a_{\mathrm{v}}\left(\mathrm{m}/{\mathrm{s}}^2\right)$	Ride comfort
≤0.315	Not uncomfortable
0.5–0.63	Little uncomfortable
0.63–0.8	Little uncomfortable to fairly uncomfortable
0.8–1.0	Fairly uncomfortable to uncomfortable
1.0–1.25	Uncomfortable
1.25–1.6	Uncomfortable to very uncomfortable
1.6–2.0	Very uncomfortable
2.0–2.5	Very uncomfortable to extremely uncomfortable
≥2.5	Extremely uncomfortable

Table 2.

ISO 2631 comfort evaluation scale.

3.2 EN 12299—mean comfort method

The EN 12299 standard is a statistical method based on the rms method. The mean comfort is divided into two methods, the standard method and the complete method. The standard method only considers the floor vibration, whilst the complete method uses floor and seat locations. Thus, the standard method is a simplification of the complete method.

The two method variations quantify the passenger mean comfort during a continuous 5 minutes run. Therefore, the measurement duration shall be a multiple of five, and a minimum of four zones traveled at constant speed must be accomplished to apply the method [11, 41, 42].

In opposition to the ISO 2631 method, the frequencies are initially weighted, and then the rms acceleration over 5 seconds is calculated for each axis. Finally, the 95th and 50th percentiles are determined for periods of 5 minutes, and the mean comfort index is obtained.

The mean comfort ($N_{\mathrm{MV}})$ is calculated following Eq. (3):

$$N_{\mathrm{MV}}=6\sqrt{{\left(a_{x\mathrm{P}95}^{\mathrm{w}}\right)}^2+{\left(a_{y\mathrm{P}95}^{\mathrm{w}}\right)}^2+{\left(a_{z\mathrm{P}95}^{\mathrm{w}}\right)}^2}$$E3

where, $a_{\mathrm{P}95}^{\mathrm{w}}$ represents the 95th percentile of the weighted accelerations in the three directions, x, y, and z. The evaluation of $N_{\mathrm{MV}}$ is defined based on a scale, Table 3. The scale considers values between 1 and 5, where a ride comfort index under 1 is considered a “very comfortable ride”, and above 5 is considered a “very uncomfortable ride” [42].

NMV	Ride comfort
≤1	Very comfortable
1–2	Comfortable
2–4	Medium
4–5	Uncomfortable
≥5	Very uncomfortable

Table 3.

EN 12299 evaluation scale.

This method presents some significant limitations; the use of the 95th percentile leads to data exclusion and the lack of possibility to correspond the track irregularity’s location with the $N_{\mathrm{MV}}$ values (the highest $N_{\mathrm{MV}}$ values can occur during different 5 seconds time intervals). Moreover, measurements must occur at a constant speed for 5 continuous minutes, which is difficult to achieve during passenger service [42].

3.3 Sperling’s method

The special characteristic of Sperling’s method is that the ride comfort index $\left(W_z\right)$ is evaluated individually for vertical and lateral directions. The calculation goes following Eq. (4).

$$W_{\mathit{Zi}}={\left[{\int }_{0.5}^{30}{G}_i\left(f\right)B_i^2\left(f\right)\mathrm{d}f\right]}^{\raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$6.67$}\right.}$$E4

Where $G_i$ corresponds to the double-side square acceleration [(cm/s²)²] and $B_i$ represents the frequency weighting curve. As with the previous methods, the WBV level is evaluated based on a scale, Table 4. The passengers will not feel discomfort for values under 3 and will feel extreme discomfort for results above 3.5 [10].

$W_z$	Ride comfort
1	Just noticeable
2	Clearly noticeable
2.5	More pronounced but not unpleasant.
3	Strong, irregular but still tolerable.
3.25	Very irregular
3.5	Extremely irregular, unpleasant, annoying; prolonged exposure intolerable.
4	Extremely unpleasant; prolonged exposure harmful.

Table 4.

Sperling’s method ride comfort evaluation scale.

This method is mainly applied to evaluate the vibration level of the vehicle rather than the users. Therefore, Sperling’s method is specially used when comparing two or more train comfort rides. This method’s major limitation is that vibration influences in different frequency bands and directions relating to sitting comfort are ignored [39].

4. Ride comfort index and rail and train maintenance identification

Besides affecting passengers’ comfort, vibration is also associated with safety. High safety levels, guaranteed by maintenance, are crucial to secure the reliability and longevity of rails and trains, furthermore nothing drives passengers away more than safety failures. Maintenance is performed by corrective and preventive interventions. The reactive action after failure recognition characterizes the corrective approach [13, 14, 15, 16]. In opposition, preventive interventions are defined as proactive measures to prevent and minimize failures at reduced costs.

Condition-based maintenance (CBM) is the traditional applied preventive track monitoring system. This is executed based on a defined schedule, allowing action to be taken when there is failure evidence. Inspection vehicles, such as EM 120 are usually applied to detect track failures [43, 44, 45]. These are expensive vehicles, and their passage introduces traffic disruptions, affecting the regular service operation [16, 44].

Railway track abnormalities influence rail vibration. Higher acceleration peaks are noticed on a defective rail compared to a healthy one. Thus, besides affecting safety, track abnormalities lead to increased passenger discomfort [15]. Due to the connection between vibration, track infrastructure abnormalities, and ride comfort, a new CMB methodology capable of identifying track abnormalities and rail vehicles maintenance needs was proposed. Based on the limitations of EN 12299 and Sperling’s method, the ISO 2631 standard was defined as a reference methodology concerning ride comfort analysis. The main goal was to overcome the limitations of the current CBM methods, providing a complementary low-cost solution without disrupting the railway service. Thus, it was hypothesized that if multiple trains with different suspension systems present floor discomfort at the same location, then the track infrastructure requires maintenance. Moreover, it was also considered that the train vehicle needs maintenance if a specific train presents low ride quality at the seat surface.

4.1 Experimental procedure

Different train series have different suspension mechanisms. Those are the critical mechanical structure affecting ride comfort. The suspension mechanism’s function is to attenuate the vibration resulting from the train motion and wheel-track interaction. The primary suspension contains wheel-track interaction vibrations, whereas the secondary suspension suppresses the vibration transmission from the bogie to the carbody. This way, vibration is attenuated from the bogie to the seat, so if passengers feel discomfort, a track abnormality is identified [46, 47, 48].

Based on the mentioned assumption, it was assumed that the track infrastructure needs maintenance if trains with different suspension mechanisms reported floor discomfort at the same geographic location. Pendolino, Intercity and Urban trains vibrations were monitored for 18 journeys whilst running a passenger service. For the long-distance trains, nine measurements were taken for the Pendolino and six for the Intercity train. Measurements occurred at different places inside the train, namely at the lead, middle, and end cars. Regarding Urban rail, three records happened at the first seat in the motion direction. The experiments run on the Northern Line, downward direction, between Porto–Campanhã and Aveiro stations.

Following ISO 2631 recommendations, three-axial accelerometers were placed on the floor and seat surface and aided by a GPS, the train’s geographic location was obtained. This way, track infrastructure maintenance sections were identified by matching the floor discomfort with the train geographic location. Vibration measurements were performed by three-axial seat pad accelerometers (PCE-VDL-24I ±16 g) at a sample rate of 200 Hz, preventing aliasing and respecting Nyquist’s theorem. Moreover, in addition to following ISO 2631 recommendations, the accelerometers also allow data recording on a μSD card [49].

The geographic location was obtained using a RedBoard Qwiic [50] connected with a GPS Logger Shield [51]. The equipment was programmed to retrieve the location, train velocity, and record the data on the μSD card at 1 Hz. Vibration and geographic measurements occurred synchronously.

4.2 Track maintenance needs identification

For this purpose, an algorithm was developed using MATLAB 2020a (from MATHWORKS) [52]. Following ISO 2631 approach, the analysis starts by calculating the instantaneous floor discomfort, corresponding to the total acceleration calculation at each second. Discomfort levels were divided into two categories: equal or under 0.315 m/s², rated as “Not uncomfortable”, and above the same threshold, ranked as “Uncomfortable”. Once identified, the “Uncomfortable” locations were mapped.

The GPS device was installed in the train driver’s cabin to get the best GPS signal. Therefore, a maximum distance of 158.9 m (Pendolino length) between acceleration and GPS measurements was observed for measurements taken at the end of the train. Moreover, considering the worst-case scenario where the Pendolino travels at 220 km/h, around 61.1 m are achieved in 1 second. Therefore, a maximum 220 m offset was applied to obtain the matching segment.

Considering all mentioned parameters, the matching locations and maintenance needs segments are those with less than 220 m distance between discomfort places of three different trains, Figure 1.

Figure 1.

Illustrates the track maintenance segments identified by the MATLAB algorithm.

According to Figure 1, three maintenance segments were identified. The start and end geographic locations and the respective associated kilometer are presented in Table 5 for each zone.

Segment number	Start coordinates		End coordinates		Line kilometer
	Latitude	Longitude	Latitude	Longitude
1	40.9771	−8.6372	40.9747	−8.6366	315
2	40.9005	−8.6221	40.8941	−8.6211	306
3	40.8817	−8.6191	40.8797	−8.6188	304

Table 5.

Maintenance segments geographic locations and line kilometer.

The listed locations match those defined by the Portuguese train company (IP) by passing the EM 120 inspection vehicle as requiring maintenance [53]. This way, the methodology reliability is proven, and ride quality as maintenance needs identification may be applied, resulting in a low-cost and non-disruptive CBM method.

4.3 Train maintenance needs identification

Following ISO 2631 standard methodology, comfort levels were calculated for all journeys regarding the floor and seat surface locations. It was hypothesized that if most trains present “Not Uncomfortable” journeys, those ranked with discomfort levels for floor or seat surface needed maintenance. The previous acceleration records calculated the total acceleration for each trip; results for the Pendolino are shown in Table 6. The train identification number was replaced to keep the train’s ID anonymous.

Train ID	Measurement place	Measurement location	$a_{\mathrm{v}}\left(\mathrm{m}/{\mathrm{s}}^2\right)$	Ride comfort
A	Beginning	Floor	0.06	Not uncomfortable
		Seat surface	0.27	Not uncomfortable
B	Beginning	Floor	0.06	Not uncomfortable
		Seat surface	0.24	Not uncomfortable
B	Beginning	Floor	0.06	Not uncomfortable
		Seat surface	0.22	Not uncomfortable
C	Middle	Floor	0.05	Not uncomfortable
		Seat surface	0.23	Not uncomfortable
D	Middle	Floor	0.05	Not uncomfortable
		Seat surface	0.24	Not uncomfortable
A	Middle	Floor	0.05	Not uncomfortable
		Seat surface	0.22	Not uncomfortable
C	End	Floor	0.05	Not uncomfortable
		Seat surface	0.24	Not uncomfortable
C	End	Floor	0.06	Not uncomfortable
		Seat surface	0.27	Not uncomfortable
E	End	Floor	0.06	Not uncomfortable
		Seat surface	0.26	Not uncomfortable

Table 6.

Pendolino ride comfort evaluation.

Pendolino results presented low acceleration levels associated with the “Not Uncomfortable” ranking. The same tendency was not observed for the InterCity trains. Comfort levels, see Table 7, noticed high discomfort levels for the seat surface location. This way, three trains were identified as needing maintenance.

Train ID	Measurement place	Measurement location	$a_{\mathrm{v}}\left(\mathrm{m}/{\mathrm{s}}^2\right)$	Ride comfort
A	Beginning	Floor	0.06	Not uncomfortable
		Seat surface	0.28	Not uncomfortable
B	Beginning	Floor	0.06	Not uncomfortable
		Seat surface	0.23	Not uncomfortable
C	Middle	Floor	0.08	Not uncomfortable
		Seat surface	0.27	Not uncomfortable
D	Middle	Floor	0.06	Not uncomfortable
		Seat surface	0.38	Little uncomfortable
B	End	Floor	0.06	Not uncomfortable
		Seat surface	0.36	Little uncomfortable
E	End	Floor	0.06	Not uncomfortable
		Seat surface	0.39	Little uncomfortable

Table 7.

Ride comfort evaluation for Intercity trains.

In opposition to the Pendolino train, operated as a single unit, the Intercity train is composed of one locomotive hauling five coaches. Although five hauled coaches perform under the same journey conditions, each car’s comfort does not depend on the other vehicles’ comfort. Trains D, B and E, presented “Little Uncomfortable” comfort levels; thus, it can be concluded that these trains needed maintenance.

Regarding the Urban train, as on Pendolino, all journeys were ranked as “Not Uncomfortable” independently of the measurement location. Table 8, shows Urban train results.

Train ID	Measurement place	Measurement location	$a_{\mathrm{v}}\left(\mathrm{m}/{\mathrm{s}}^2\right)$	Ride comfort
A	Beginning	Floor	0.06	Not uncomfortable
		Seat surface	0.20	Not uncomfortable
B	Beginning	Floor	0.07	Not uncomfortable
		Seat surface	0.26	Not uncomfortable
C	Beginning	Floor	0.06	Not uncomfortable
		Seat surface	0.22	Not uncomfortable

Table 8.

Urban trains ride comfort evaluation.

Train seats have the potential to modify the vibration transmission from the floor to the user. Vibration tends to be amplified by the seat. Thus, ride comfort needs to be evaluated on both floor and seat surface to draw conclusions. Based on this assumption, ride quality assessment was conducted for both locations on multiple Pendolino, Intercity and Urban trains.

Pendolino started its operation in 1999 and was renovated in 2017. This renovation changed the seat design but kept the seat structure and carbody. The Intercity service was introduced earlier than the Pendolino in 1980 and renovated in 2002. In the same year, the 3400 series Urban train was introduced. Thus, besides having different characteristics, it was expected that the Intercity and Urban trains presented similar ride comfort values due to the coincident renovation and service introduction years [54, 55].

However, whilst Pendolino and Urban journeys were ranked as “Not Uncomfortable”, the same tendency was not observed on the Intercity journeys. Three of those vehicles reveal high discomfort levels. Thus, once the journeys were run on the same track, it is possible to conclude that Intercity coaches present maintenance needs.

5. Transmissibility

Passengers spend most of their time seated; thus, vibration is transmitted to the user due to the contact with the seat and floor. This way, the seat is essential to reduce vibration transmission and increase passengers’ comfort. The seat and human body’s dynamic responses affect seat vibration transmission as they constitute a coupled dynamic system [10, 11, 33, 35, 37, 56]. Therefore, seat dynamics are quantified regarding transmissibility, which verifies seat efficiency in vibration discomfort and is an indicator of ride comfort. The seat transmissibility presents the ratio of vibration at the user-seat interface and the floor, according to Eq. (5):

$$H\left(f\right)=\frac{G_{\mathrm{o}}\left(f\right)}{G_{\mathrm{i}}\left(f\right)}$$E5

where H(f) represents the transmissibility, G_o(f) is the output acceleration at the seat-user interface, and G_i(f) represents the input acceleration at the floor [57].

Transmissibility differs in direction (vertical, fore-and-aft, and lateral) and location (seat surface and seatback). Laboratory experiments demonstrated a vertical transmissibility peak between 4 and 6 Hz when sitting upright with backrest support [57, 58, 59, 60]. However, these studies did not implement transmissibility tests on train seats. Instead, they considered a rigid seat frame with distinct characteristics of the standard train seat, such as frame dimensions and support points. Moreover, in opposition to the natural rail environment, the experimented seats are individual instead of double, and the foam is placed on top of the surface without a restricting cover [57, 58, 59, 60]. These are essential parameters because seat transmissibility is affected by the physical properties of the foam, such as thickness, density, or yield strength [59]. Doubling the foam thickness roughly halves its stiffness; thus, transmissibility and discomfort also increase. Patelli and Griffin [57] conducted a study where foam thickness’s effect on transmissibility was observed. When the thickness increased from 40 to 80 mm, the transmissibility reduced from 4 to 7 Hz to 3–5 Hz, respectively, with maximum frequencies of 3–4 Hz for the 40 mm foam and 2–3 Hz for the 80 mm foam. In this study, the individuals did not have contact with the backrest [58]. Zhang et al. [61] also reported a similar tendency and observed higher transmissibility when the seat pan foam was increased from 60, 80, and 100 mm.

The stomach has a resonance frequency between 2 and 20 Hz; thus, passengers may feel sick if the seat transmits vibration within that range.

Ribeiro [62] performed transmissibility tests on Alfa Pendular trains in 2012 before the train renovation using the same method as the one used in this study. Transmissibility peaks around 4.3 Hz were found. At that time, seats were covered by tissue without seams instead of the leather covers with seams now used [54]. Moreover, Ribeiro numerically identified the rigid body frequencies of the carbody and bogie. The former reported frequencies under 1.42 Hz, whilst frequencies between 4 and 12 Hz characterize the latter [62].

To properly quantify the dynamic performance of the Pendolino seats, transmissibility tests were conducted on its two-seat classes, namely, comfort and touristic. Figure 2 illustrates both seat types. The main difference regarding those seats is their dimensions, especially the seat surface thickness. The comfort class seat has a thickness of 190 mm, whereas the touristic class has a thickness of 130 mm. Moreover, the 2017 renovation introduced significant changes to seat covers and foams. All foams were replaced with new ones, and a leather cover with seams was introduced [54].

Figure 2.

Pendolino seats: (a) comfort seat, (b) touristic seat.

A set of dedicated experimental tests were accomplished. According to Table 9, four volunteers (two males and two females), aged between 9 and 39, weighing 33–115 kg and 1.33–1.87 m in height, participated in the study. The subjects sat in a normal posture, placed both hands on their thighs, and made complete contact with the seatback.

Subject	Gender	Age (years)	Weight (kg)	Stature (m)
M1	Female	9	33	1.33
M2	Female	25	58	1.70
M3	Male	26	80	1.87
M4	Male	39	115	1.85

Table 9.

Subjects’ characteristics.

Two seats, one of each class, were instrumented. Seats had the same exact location in the vehicle, particularly near the bogie. Acceleration at the seat surface was measured using a three-axial seat pad accelerometer (PCB 356B41). Uniaxial accelerometers (PCB 393A03) were placed on the floor (1 unit) and the metallic support frame (4 units) for measuring the acceleration at the seat frame. Accelerometers were positioned as shown in Figure 3.

Figure 3.

Accelerometer positioning: (a) seat frame and floor accelerometers, (b) seat surface pad.

The data acquisition system was composed of a NI cDAQ-9172 with NI 9234 IEPE modules connected to a PC to acquire and record data measurements. The vibration was induced by a group of people randomly walking and jumping nearby the seat. A time series of 3 minutes with a sampling frequency of 2048 Hz, posteriorly decimated at 100 Hz, was saved. Data processing was performed using MATLAB scripts previously validated.

5.1 Transmissibility curves

Transmissibility peaks were identified for all individuals within the two seats. Figure 4 shows the transmissibility curves for the comfort and touristic seats for individuals M1 to M4. Three transmissibility peaks were identified, T1 (yellow zone), T2 (gray area), and T3 (green zone).

Figure 4.

Transmissibility curves: (a) comfort seat, (b) touristic seat.

The comfort seat shows T1 transmissibility equal to 0.2 Hz, the T2 transmissibility presents two main peaks, at 1.4 Hz and 2.5 Hz resulting from the same type of movement but with different amplitudes. T3 appears at frequencies equal to 4.9 Hz for M2 and M3, 4.7 Hz for M4, and 4.1 Hz for M1.

The touristic seat presented a peak of 0.2 Hz concerning T1. On this seat, T2 presented two prominent peaks, a common peak of 1.8 Hz for all individuals and a second peak between 2.3 and 3.3 Hz, depending on the subject. The T3 transmissibility area starts at 4.3 Hz for M3, and increases for M1 (4.5 Hz), M2 (4.7 Hz), and M4 (4.9 Hz).

Comparing comfort and touristic transmissibility results, T1 characterizes both seat types at lower frequencies with values of 0.2 Hz. Increased frequencies define the T2 area. The frequency ranges for the comfort and touristic seats are observed to be between 1.4–2.5 Hz and 1.8–3.3 Hz, respectively. Concerning the last transmissibility zone, T3, this presents frequencies between 4.1 and 4.9 Hz for the comfort seat and a range of 4.3–4.9 Hz for the touristic one.

The frequencies reported in this experiment are lower than those reported in the study performed by Ribeiro (4.3 Hz). Introducing new foams and covers with limited seams significantly modifies vibration transmission. These factors restrict the foam cellular morphology movements, and the vibration absorption varies compared to the free foam of the pre-renovation seats. Moreover, the new foams may present different mechanical properties from the previous ones, affecting their vibration-transmission capability. Transmissibility values reported by Ribeiro [62] are similar to those of T3.

The frequency values presented in this study are also lower than those presented by several authors [57, 58, 59, 60], who found vertical transmissibility around 4.3 Hz, similar to Ribeiro’s findings. However, these tests were conducted in a laboratory environment instead of a real rail environment. Moreover, a single seat was considered, and the seat frame was replaced by a simplified rigid metallic structure (approximately 1000 kg) with free foam on top of the surface without covering or other types of movement restrictions.

Matching the present results with the carbody and bogie modal identification conducted by Ribeiro, T1 zone can be associated with foam movements induced by carbody vibrations. T2 and T3 present frequencies within the bogie rigid body movements. Therefore, these transmissibility zones can be associated with this range, which may lead to seat structural movements such as rigid body, torsion, or bending [62].

In agreement with the research of Patelli and Griffin [57] and Zhang et al. [61], transmissibility peaks decreased for higher foam thicknesses; that is, T1 and T2 transmissibility frequencies were higher for the touristic seat than those for the comfort seat. Moreover, the Pendolino seat foam thickness is higher than those of the previous studies; therefore, following the reported trend of those experiments, lower transmissibility frequencies were expected in the present results.

6. Interface pressure

The overall seat discomfort is influenced by both dynamic and static seat characteristics. These two factors are firmly connected, the importance of one depends on the influence of the other. In low or absent vibration, discomfort evaluation depends mostly on the static seat characteristics. On the other hand, dynamic features dominated discomfort evaluation when vibration magnitude increases. Therefore, it is essential to consider both static and dynamic seat characteristics when evaluating passengers’ comfort. Interface pressure is commonly used as a static comfort evaluation technique [63, 64].

When a passenger sits, the pressure at the interface between the seat and the user’s buttocks varies over the seat surface area. This pressure variation defines the “average pressure”, representing the mean pressure induced on the seat surface. On the other hand, some pressure is concentrated around the ischial tuberosities leading to pressure peaks [64, 65]. A maximum 32 mmHg threshold should not be exceeded. This value corresponds to the capillary pressure value, and above that the pressure is considered harmful as it can obstruct the capillaries, restricting blood circulation and, consequently, result in a deprivation of oxygen to the tissues, causing discomfort [63, 66, 67, 68]. The ideal foam equally distributes pressure on the seat surface.

After evaluating the dynamic seat characteristics of the Pendolino train, the static conditions were assessed by measuring its interface pressure. Experiments run with the same conditions (seats and subjects) as the transmissibility tests. The interface pressure was recorded using a CONFORMat sensor (from TekScan) for 10 minutes for each individual, and maximum pressure was recorded. Figure 5 illustrates the pressure distribution for Pendolino seats regarding M2 subjects. Similar tendencies were observed for all individuals.

Figure 5.

Pressure distribution: (a) comfort seat, b) touristic seat.

Maximum pressure of 32 and 37 mmHg were found for the comfort and touristic seats, respectively. Those pressures correspond to the discomfort threshold; thus, passengers may feel discomfort. Apart from the ischial tuberosities, the pressure was equally distributed on the seat surface.

7. Conclusions

Nowadays, railways are one of the most used mass transportation systems worldwide. Compared to air travel or car, their high transportation capacity and low environmental impact are significant advantages. Higher passenger numbers will be achieved by increasing those vehicles’ attractiveness, which depends mainly on comfortable journeys. Passengers define a comfortable journey based on comfort, safety, and user conditions. Seat comfort is characterized by the interaction between seat dynamic and static characteristics. The former is associated with seat behavior under vibration conditions, whilst the latter concerns comfort evaluation in the absence of vibration. Therefore, both evaluation types should be performed to evaluate users’ comfort properly.

The rail-wheel interaction associated with motion leads to a complex vibration environment. Besides affecting comfort, vibration is strongly connected to safety. Based on that assumption, a new CBM methodology capable of identifying train and rail maintenance needs was proposed. Following ISO 2631 approach, ride comfort was calculated for 18 journeys conducted on Pendolino, Intercity and Urban trains connecting Porto–Aveiro stations. Results matched those of IP obtained by the EM 120 inspection vehicle. The low-cost system was then validated, and its high precision was proved.

Improving the seating capability to absorb vibration is crucial to increasing passenger comfort. The extent to which vibration is amplified or attenuated depends primarily on the dynamic properties of the seat. Transmissibility quantifies these properties. Therefore, transmissibility tests were performed on comfort and tourist seats on the Pendolino train. The prominent transmissibility peaks were lower than those obtained for the simplified seat version in the laboratory experiments.

After classifying seat dynamic characteristics, the seat static parameters were evaluated by pressure tests. Maximum pressures were obtained at the discomfort threshold near the ischial tuberosities. Apart from this location, the pressure was equally distributed on the seat surface.

To the author’s knowledge, this is a pioneering study considering a complete passenger comfort evaluation. Moreover, new low-cost and non-disruptive maintenance identification methodologies were proposed based on the passengers’ comfort evaluation. The results of this research can be applied to improve users’ comfort levels and increase the attractiveness of the rail mode of transportation.

Acknowledgments

The first author expresses gratitude to the Fundação para a Ciência e Tecnologia (FCT) for a PhD scholarship under the project iRail (PD/BD/143161/2019). The authors would like to acknowledge the support of the projects FCT LAETA–UIDB/50022/2020, UIDP/50022/2020, and UIDB/04077/2020.

Conflict of interest

The authors declare no conflict of interest.