Evaluation of Surface Texture of Hydrogen Pressure Regulators

At the present time, the primary application of hydrogen fuel systems is in the automotive industry, where hydrogen fuel cell electric vehicles (FCEVs) are capable of achieving zero exhaust emissions and zero pollution [5].

The FCV is regarded as a significant zero-emission alternative, presenting a range and refueling time comparable to that of internal combustion engine vehicles. A multitude of model-based studies have been conducted with the objective of enhancing fuel cell system and management in automotive applications. One of the principal components of the fuel system is the pressure regulator, which has a considerable influence on the overall performance of the automotive powertrain (Figure 1) [6].

1.png

Figure 1. Pressure regulator as a part of the hydrogen supply system in the LDV automotive fuel cell system [7]

The efficacy of the system is contingent upon the functionality of the pressure-regulating components, which are responsible for ensuring the optimal delivery of hydrogen at a specified pressure [6].

A hydrogen FCEV is capable of achieving effectively zero exhaust emissions and zero pollution. It is not only an optimal substitute for conventional FCVs, but also the principal method of utilizing hydrogen energy in an efficient manner [8]. The primary method of providing mileage comparable to that of conventional internal combustion engine vehicles is through the utilization of high-pressure hydrogen storage. This necessitates that the hydrogen within the gas cylinder be maintained at a pressure of 30-35 MPa [9]. The typical operational pressure of a proton exchange membrane hydrogen fuel cell is 0.16 MPa. Accordingly, a pressure reduction system has been devised for the FCEV. In consideration of the thermodynamic properties of hydrogen, a two-stage pressure reduction method is currently employed for the FCEV. The two-stage process entails a reduction in pressure from 35 MPa to 5 MPa in the initial stage and from 5 MPa to 0.16 MPa in the subsequent stage [5].

In a hydrogen power system, an increase in pressure can result in enhanced corrosion of the cathode, which may ultimately culminate in a short circuit within the cell and reduced efficiency or inoperability. Accordingly, the pressure selected for supplying the fuel cell with hydrogen has a direct correlation with the cell's lifespan, fuel consumption, and the commercial viability of the system [10].

The considerable difference between inlet and reduced pressure levels gives rise to the specific operation of the regulators, which in turn necessitates particular requirements for the geometry and surface roughness of the components, as well as the utilization of materials that are conducive to the sealing of the hydrogen molecule.

The continuous increase in demand for improved reliability and mechanical parts, coupled with a reduction in frictional losses and, in particular, an increase in power density, is resulting in a notable rise in the tribological load on contact surfaces. The characteristics of the contact between the surfaces in question are becoming increasingly significant. Two of the most crucial surface properties are surface roughness and topography (Figure 5). For the optimal design of contact surfaces, it is of paramount importance to comprehend the impact of surface roughness parameters on friction, wear and tear, particularly in the context of hydrogen systems [16].

The experimental setup for measuring the roughness and topography of a Memtex RH242 hydrogen pressure regulator is presented in Figure 5. The measurement was performed with a Taylor Hobson Form Talysurf i-Series. The graphical results of the roughness parameter measurement (Graphic 1) and the profile parameters (Graphic 2) are presented, with additional lines located at 11.5 mm and 13.5 mm showing an area with a profile close to linear, which reflects the piston working area. The graphical results thus presented justify the need for increased control of surface texture parameters, at the expense of less tight control of surface shape deviation.

In order to ascertain the primary surface characteristics of the critical elements of the Memtex RH242 hydrogen pressure regulators [11], the surface texture and topography parameters have been measured using the Taylor Hobson Form Talysurf i-Series [17]. Based on the preliminary observations, the surface roughness parameters have been selected for control, and the data has been statistically processed and analyzed.

A number of different surface roughness parameters may be utilized, depending on the specific circumstances. These characteristics are associated with the geometrical properties of the workpiece. These parameters are defined and can be found in several standards, including ISO 4287:1997 [18], ISO 1302:2003 [19], and ISO 21920:2021 [20]. The selection of an appropriate measurement parameter is crucial for the accurate determination of roughness. The surface parameters can be calculated from the filtered roughness profile. The most commonly utilized surface parameters are outlined below [21].

5.png

Figure 5. Experimental installation and measurement results of roughness and topography for Memtex RH242 hydrogen pressure regulator using Taylor Hobson Form Talysurf i-series [22]

The arithmetic mean of the absolute ordinate values, Z(x), within the sampling length is defined as Ra (Eq. (4) [18]). It is important to note that this roughness parameter of a surface can vary considerably without affecting the performance of the surface. Therefore, it is common practice to specify a tolerance band or a maximum Ra value that is acceptable on the drawing [23]:

$R a=\frac{1}{l} \int_0^l|Z(x)| d x$           (4)

The Ra value for a given sample length represents the average roughness, and thus the effect of an atypical peak or valley will have only a minor influence on the value. It is recommended that the Ra value be estimated for several successive sampling lengths and the resulting values averaged. This will ensure that the Ra value is representative of the surface being inspected. It is essential that the measurements be made perpendicular to the surface. The Ra value does not provide information about the shape of irregularities on the surface. Similar Ra values can be obtained for surfaces with markedly different profiles. It is useful to indicate the machining process used to produce the surface [23].

As the requirements for surface texture have evolved, so have the standards. The most significant change in ISO 21920 [20] concerns the manner in which the calculation is conducted, it is no longer performed repeatedly and subsequently averaged. The revised calculation method entails a single value of Ra (or Rq). However, this does not apply to Rp, Rv and Rz, which continue to be averaged to minimize the impact of outliers [24].

Rq (Eq. (5) [18]) is the root mean square value of the ordinate values Z(x) within the sampling length and expressed mathematically [23]:

$R q=\sqrt{\frac{1}{l} \int_0^l Z^2(x) d x}$          (5)

When compared to the arithmetic average, the root mean square parameter has the effect of giving extra weight to the numerically higher values of surface height [23].

When roughness parameters are determined instrumentally Rq has the advantage that phase effects from electrical filters can be neglected. The Ra parameter using the arithmetic average is affected by phase effects that cannot be ignored. Ra has almost superseded Rq on machining specifications. However, Rq still has value in optical applications where it is more directly related to the optical quality of a surface [23].

Rz, the maximum height of the profile is defined in accordance with the sampling length. This parameter is frequently used to ascertain whether the profile exhibits protruding peaks that may potentially impact the static or sliding contact function [18, 25].

The different ISO standards provide a variety of definitions and methodologies for calculating the Ra parameter.

Rt (Eq. (6)), (total height of the profile) is parameter that highly susceptible to the influence of elevated peaks or deep scratches. The Rt value is defined as the vertical distance between the highest peak (Rp) and the lowest valley (Rv) along the specified assessment length of the profile [18, 26].

$R t=\max R p_i+\max R v_i$          (6)

Rp (Eq. (7)), maximum profile peak height: height of the highest peak from the mean line, defined on the sampling length [18, 25].

Use a natural cubic spline to interpolate through the discrete data values. For each sample length i=1 to CN. Determine portions of the profile above the mean line, these are the profile peaks. For each profile peak j=1 to m, determine the supremum height Z_pj:

$R p=\frac{1}{C N} \sum_{i=1}^{c N} R p_i$          (7)

where,

$R p_i=\max _{1 \leq j \leq m} Z p_i$          (8)

In many tribological applications, Rp and Ry is an important parameter because damage may be done to the interface by the few extreme deviations from the mean line [27]. Highest peak and the lowest valley (Rv) may affect smooth translation or the isolation of hydrogen from the environment.

The technological process for manufacturing of hydrogen pressure regulators could be evaluated by statistical analysis of the qualitative parameter roughness. Statistical analysis of stability, accuracy, and process tuning could be used for evaluation of the technological process [28].

As the requirements for surface texture have evolved, so have the standards. The most significant change in ISO 21920 [20] concerns the manner in which the calculation is conducted, it is no longer performed repeatedly and subsequently averaged. The revised calculation method entails a single value of Ra (or Rq). However, this does not apply to Rp, Rv and Rz, which continue to be averaged to minimize the impact of outliers [24].

In order to analyze the different surface roughness parameters, the results obtained by measuring the internal cylindrical surface of the first stage of the hydrogen pressure regulators Memtex RH242 [11] with the Roughness Tester INSIZE ISR-C002 [29] will be utilized. Anodized Aluminium T6511, with additional polishing of the surfaces, was used for the fabrication of the individual housing parts.

As illustrated in Figure 9, the graph presents a measurement obtained with the Insize ISR-C002 Roughness Tester [29] and displays the result for Ra on a cylindrical surface of a Memtex RH242 hydrogen pressure regulator. This result is presented graphically using Insize DataView software. Following the final finishing process, which involved polishing and Anodising, the surface macro geometry was characterised by the absence of sharp peaks on the roughness profile. This characteristic is of significant importance, as it ensures minimal friction between the sealing elements while providing effective hydrogen isolation from the atmosphere. The tabular representation of the findings is presented in Table 1.

9.png

Figure 9. Roughness curve displayed on Insize DataView for surface roughness tester V1.4. software

Standard measuring equipment and established techniques and methods are utilized for the analysis of surface roughness. An adapted methodology for surface roughness measurement was applied, based on an analysis of the design and technical specifics of hydrogen pressure regulators. The availability of statistical processing of the results allows for the correction to tolerances allowed, and enables rapid evaluation of process changes manifested in increased parameter variance. In contrast to popular ways of evaluating roughness measurement results, the adapted analysis methodology yields dynamically changing tolerance limits to signal potential problems.

The measurement is performed in four directions: +X, -X, +Y and –Y (Figure 7). Each measurement is repeated five times (Figure 10). In order to estimate the repeatability for each of the surface roughness parameters in the different directions, the sample mean (Eq. (9)) and the standard deviation (Eq. (10)) are calculated [30, 31].

$\overline{R(k)}=\frac{1}{m} \sum_{j=1}^m R(k)_j$          (9)

where, k – parameter determines the type of surface roughness calculated (Ra, Rz, Rq, Rt, Rp), j – consecutive number of attempts (j=1÷5), m – total number of attempts (m=5).

$s_{R(k)}^2=\frac{\sum_{i=1}^m\left(R(k)_i-\overline{R(k)}\right)^2}{m-1}$           (10)

The calculation of the sample mean (Eq. (9)) for each roughness parameter $\overline{R(k)}$ in each of the four roughness measurement directions (see Figure 7) enables the evaluation of the process capability, particularly in the context of post-surface finish, which is of significance for hydrogen embrittlement, corrosion resistance and internal/external leakage protection. The standard deviation s_R(k) (Eq. (10)) helps to compare the scatter of results in different measurement directions and provides a rapid estimate of random errors.

It is considered that an average value of the different roughness parameters in each direction can be used as a further estimate of the circularity deviation in the working section.

Table 1. Roughness table with measurement results of different surface roughness parameters: Ra, Rz, Rq, Rt and Rp

10.png

Figure 10. A scatter plot of the roughness Ra measurement in various cross-sections, with the data being derived from five separate measurements

As demonstrated in Figure 10, which presents the graphical outcomes for the surface roughness parameter Ra, it is evident that there is an absence of visible scatterings in the various measurement directions, with the exception of random deviations that are attributable to incorrect basing or subjective error.

A statistical evaluation of the mean values of surface roughness parameters across various cross-sections is conducted. This involves the calculation of the mean value, R(k)_av, for all cross-sections (see Eq. (11)), and the standard deviation, s_R(k)av, for all cross-sections (see Eq. (12)).

$R(k)_{a v}=\frac{R(k)_{+X}+R(k)_{-X}+R(k)_{+Y}+R(k)_{-Y}}{4}$            (11)

$s_{R(k) a v}=\frac{s_{R(k)+X}+s_{R(k)-X}+s_{R(k)+Y}+s_{R(k)-Y}}{4}$            (12)

In accordance with the final estimate, the maximum allowable parameters R(k)_lim are defined on the basis of statistical analysis:

$R(k)_{l i m}=R(k)_{a v}+c . s_{R(k) a v}$            (13)

where, c is the security index, which is a function of the available database and varies depending on the process and may be the manufacturer's know-how.

For the purpose of the study, we choose c=1.

According to the established methodology for the control of surface texture parameters, the measurement results obtained are compared with the tolerance limits:

$\overline{R(k)} \leq R(k)_{t o l}$            (14)

In an adapted control methodology for hydrogen pressure regulators, we compare the mean parameter values with dynamically varying statistically calculated limits:

$\overline{R(k)} \leq R(k)_{l i m}$            (15)

In order to evaluate the roughness parameters according to the adapted method, Eq. (15) is utilised, thereby yielding dynamically changing parameters. The results presented in Table 2 illustrate the application of statistical analysis. In the yellow cells of the table in Table 2, values of Rz and Rq are observed in the +Y and -Y directions, indicating a possible problem with shape deviation or insufficient surface stability. According to the established control methods, the parameters Rz and Rq correspond to the tolerance indicated by Equation 14. The values of the parameters Rz and Rq marked in yellow indicate elevated levels of roughness in the specified area, which has the potential to result in H₂ leaks due to material corrosion or accelerated wear of the sealing element. It is noteworthy that the elevated values are confined to two out of four measurement directions, a factor that has the potential to induce enhanced friction in these specific areas, consequently resulting in conversion characteristic unevenness.

Table 2. Roughness table with statistical parameters for surface roughness

The applied aspect of the adapted surface roughness control methodology can be considered as an analytical tool to evaluate the performance and technical characteristics of the regulators as a function of the roughness parameters.

As the main characteristic directly influencing the performance of pressure regulators directly affecting the hydrogen fuel systems, the conversion characteristic of the input pressure Pin to the output pressure Pout is considered. This conversion characteristic is strongly influenced by the smooth and uniform movement of the piston, and this movement depends on the friction between the contact surfaces. A direct functional relationship can therefore be established between the characteristics of the surface texture and the output pressure of the regulator. Ensuring a constant regulator output pressure to the hydrogen fuel cell is the basis for its efficiency and reliable, trouble-free operation.

Another important performance characteristic that can be ensured by the developed surface roughness control methodology is related to the reliable and safe operation of hydrogen pressure regulators. Surface roughness has a direct influence on sealing element wear, which can lead to internal/external leakage. Similarly, increased roughness can lead to internal/external leakage due to the small size of the hydrogen (H₂) molecule.

In batch production, the proposed control methodology may prove cumbersome and slow. The analysis of the specific requirements and design solutions of different manufacturers, combined with their know-how, can greatly help in the adaptation and implementation of the methodology. The statistical methods applied can help to facilitate the control of the geometric dimensions and the availability of a database can facilitate the measurements and reduce the number of measurements from five repetitions in four sections to the measurement of five hydrogen pressure regulators again in four sections. This possible adaptation of the developed methodology will reduce the impact of each individual regulator on the overall evaluation through the dynamic parameter R(k)_lim, but will help to better represent the overall process and indicate changes in it in a timely manner. Another parameter that allows the possibility of enhancing and reducing the influence of process changes is c (safety index), and it largely reflects the accumulated know-how from the application of the adapted methodology to the control of surface roughness by statistical methods.

This research is supported by the Bulgarian Ministry of Education and Science under the second stage of the National Program "Young Scientists and Postdoctoral Students – 2".