Effect of rewinding on flux-cored welding wires

The execution of welding and surface coating involving an electric arc is a complex, multidisciplinary process that often requires careful attention and precision in selecting the appropriate parameters, conditions, and process procedures [1‐3]. Gas shielded arc welding is a group of well-established material joining methods that allow for high welding efficiency as well as good quality of welded joints [4, 5]. As a result, these methods are used for welding various material groups in many different industries, including power, offshore, petrochemical, construction, and automotive [6‐8]. Gas metal arc welding (GMAW), flux cored arc welding (FCAW) and metal cored arc welding (MCAW) are methods that use welding wire as an electrode, which is unwound from a spool during the process [9, 10]. Even feeding of the welding wire to the arc is an essential requirement for maintaining a stable welding process [11, 12]. This requirement is important both during traditional production of welded components and during repairs and regeneration carried out using welding techniques, particularly those involving wire arc additive manufacturing (WAAM) [13, 14]. These works can be performed either in large industrial plants with advanced storage systems and extensive machinery, or in smaller enterprises where it is sometimes necessary to adapt the consumable material to existing conditions. One of the actions undertaken by users of welding wires is to rewind the wire themselves. This is most often due to the need to rewind the wire from a large spool onto a smaller one, suitable for many types of welding machines, especially smaller, workshop-sized ones. However, even with a large machine park, there may be a need to rewind the wire when the welding process is carried out in unfavourable environmental conditions or in a small space. In such cases, an appropriate length of wire suitable for working in difficult conditions is unwound, allowing the full spool to be stored in the welding consumables warehouse. Rewinding is sometimes carried out on specialised equipment, on simple stations dedicated to this purpose, and even using a drill drive.

The issue of rewinding welding wires is not mentioned in the manufacturers’ recommendations for use. A thorough literature search also revealed the lack of scientific publications addressing the influence of wire rewinding on the welding process and the quality of welded joints. The few available publications concern only selective aspects of this process. Xu et al. [15] conducted an analysis of a welding wire winding system. Experimental validation of the developed system was carried out using a mechanical prototype designed for automatic spool loading. The proposed method further improved the speed and reliability of the matching process, with the results demonstrating high robustness, accuracy, and efficiency, making it suitable for mass production applications. Moysés et al. [16] used an additional cleaning and lubricating system for GMAW wires to assess whether the introduction of such an additional stage in the wire feeding path would change the process conditions. They found that the system was effective in reducing the feeding force without increasing electrical resistance, but only for wires with better surface condition.

A fundamental requirement of any welding process is that it maintains adequate stability [17, 18]. It depends on many factors, such as the type and composition of the filler material, the welding arc atmosphere, the quality of the base material, and its surface condition [19‐21]. Dynamic changes in welding parameters should be characterized by the smallest possible amplitude, the shortest possible response time — returning to stable values — and the shortest total duration of their occurrence. This makes it possible to avoid welding imperfections such as lack of fusion, porosity, and undercut. With proper process control and monitoring, this can be achieved even in the unfavorable environment of underwater welding [22, 23].

There are many systems for monitoring the welding process. The majority focus on monitoring welding current and arc voltage [24], but stability can also be observed through acoustic emission analysis, arc light emission, by monitoring the frictional resistance of the moving wire, or by using a high-speed camera that records the arc behavior and metal transfer. De Miranda et al. [25] found that using optical sensor can be an efficient and consistent method of setting parameters that provide stable transfer. The issue of accurately assessing process stability has not been fully resolved, and new publications continue to appear that make significant contributions to this topic. In their article, Ferreira et al. [26] presented a new method for assessing stability based on arc reignition voltage peak distribution. On this basis, they introduced a novel stability index, whose effectiveness for short-circuiting GMAW processes they confirmed. Kah et al. [19] based on the measurement of the welding current and arc voltage varying with time, supported by a high-speed camera, assessed the possibility of evaluating stability in GMAW welding with spray and pulse transfer mode.

The specific construction of cored wires, which consist of a metal sheath and a core filled with flux or metal powder, means that their use offers possibilities but also presents challenges and risks [27‐30]. The selection of appropriate welding parameters ensuring stable metal transfer and adequate properties of welded joints is the subject of numerous studies [31‐34]. Cored wires may be more sensitive to any unusual conditions, which manifests itself in their susceptibility to changes in properties caused by storage conditions. In the articles by Wolski et al. [35] and Świerczyńska [36], an increase in the diffusible hydrogen content, a change in stiffness, and a decrease in the plasticity of wires stored in urban conditions were observed. Similarly, in the article by Świerczyńska et al. [37], wires stored in urban and rural conditions were compared, and it was found that wire with a seam significantly increased the diffusible hydrogen content in deposited metal while simultaneously reducing the ductility of the weld metal. It implies that the behavior of cored wires is also highly dependent on their structure. It has been found that the manufacturing process, surface condition, and seam structure have a significant impact on the use of flux-cored wires. For this reason, they may also be more sensitive to the rewinding process.

The aim of the study was to assess the impact of rewinding on the stability of the welding process using different types of welding flux-cored wires with different surface conditions.

The research was divided into three stages:

To ensure that the rewinding process was carried out in a stable, repeatable, and controlled manner, a dedicated test stand was constructed. The basic assumptions were the versatility of the rewinder for different wire diameters, different types of spools, and different materials from which the wires are made. Therefore, it was necessary to use a suitable motor and braking system to control the rewinding process. An additional assumptions were the possible compactness of the test stand and simplicity of construction, which would allow for more intuitive operation and bring the rewinding conditions closer to typical workshop conditions.

The second stage of the research consisted in verifying the possibility of using the rewinder for cored wires. For this purpose, two grades of welding wires for low-carbon steel with a diameter of 1.2 mm were used — a non-copper-coated, seamed wire wound on a basket spool, and a copper-coated, seamless wire wound on a solid plastic spool. For each wire, three layers were rewound onto solid plastic spools. The rotational speed of the rewinding spool was 45.14 rpm. In both cases, 5 coils were manually arranged, after which the motor was switched on. Rewinding took place without leading the wire.

For each spool, three surfacing welds were made using an ABB IRB 2400 robot and a Migatronic BDH 320 welding power source. The following welding parameters were used: wire feed speed 10 m/min, arc voltage 25 V, welding speed 30 cm/min, shielding gas flow rate 12 l/min, shielding gas type M21 (Ar + 18% CO₂), stick-out length 15 mm.

An effective method to indicate the differences between rewound and non-rewound wires is the measurement of welding process stability. These measurements were carried out using the Kemppi DataCatch measurement module, which allows welding stability to be monitored by continuously measuring welding current and arc voltage values. The measurement frequency was 10 Hz. Cyclograms with relationships between welding current and arc voltage were used to present the results – the more concentrated the cyclogram area, the more stable the welding process [38].

An additional test that can be useful in assessing the effectiveness of the rewinding process is the measurement of diffusible hydrogen content in deposited metal. For two chosen wires measurements were made by high-temperature extraction method by using a Bruker G4 Phoenix device equipped with a thermal conductivity detector. Test assemblies of type B from mild steel were utilized for testing, in accordance with ISO 3690. The specimens were weighed with a precision of 0.01 g. For all wires under test, padding welds were made using an automated process, with identical parameters to those used in stability tests. The specimen was held in a copper fixture with copper foil strips, and a weld bead was then deposited. After welding, the specimen was quenched in water with ice, followed by storage in a liquid nitrogen vessel. Hydrogen extraction was performed at 400 °C for 30 min.

The third stage of the research aimed to verify the possibility of using cored wires stored in unfavorable conditions for rewinding and welding. The selected wires were rewound after being stored and then both the wire on the original spool and the rewound wire were subjected to the same analyses. The storage was carried out for one month in the city of Katowice (Poland), which is the center of the Upper Silesian Industrial District. The conditions were unregulated and recorded through temperature and relative humidity measurements by a Termioplus recorders with the SHT15 sensor (temperature recorded with a resolution of 0.1 °C, humidity recorded with a resolution of 0.1% RH). The wires were fully unpackaged and exposed to the ambient air, but protected from direct exposure to precipitation as well as moisture from the ground.

A full-factorial experimental design was employed, involving five independent factors: rewinding (Y/N), storage (Y/N), wire type (seamed/seamless), arc voltage (24/30 V), and wire feed speed (8/12 cm/min). The factors and their levels are summarized in Table 1, while the complete test matrix comprising all 32 experimental runs is presented in Table 2.

The welding was performed on 20 mm thick S355JR steel sheet. The produced surface welds were subjected to visual testing to detect surface imperfections. For the analysis, the following stability indexes were used: average voltage (U_a), average current (I_a), standard deviation of voltage (S_U), standard deviation of current (S_I), voltage root mean square (U_RMS), current root mean square (I_RMS), coefficient of voltage variation (CV_U), coefficient of current variation (CV_I), voltage Carrera coefficient (CC_U), current Carrera coefficient (CC_I). Coefficients of variation are defined as the ratio of the standard deviation to the average value. The Carrera coefficient refers to the number of recorded occurrences in which the measurement value exceeded the fluctuation threshold value. The applied fluctuation threshold for the current Carrera coefficient was 5 A, while for the voltage Carrera coefficient it was 0.5 V. To enhance the clarity of result interpretation, a General Linear Model (GLM) was applied, within which a five-factor analysis of variance (ANOVA) was conducted. As response variables, two coefficients of variation and two Carrera coefficients were considered.

The rewinding of welding wire is often a routine practice in welding operations. It gives the possibility to reduce the consumption of consumables, as well as to adapt it to the available equipment. The rewinding device, built for research purposes, enabled the uniform rewinding of several types of rutile flux-cored welding wires onto spools. The arrangement was even and proceeded without interruption or additional leading, thanks to which the surface of the wire was not deprived of its protective layer or covered with an additional layer of lubricant. After rewinding, the wires were visually inspected, which confirmed that the individual spools did not differ from each other and that the rewinding was uniform.

The first test of rewinding effectiveness involved stability tests of welding parameters carried out on both non-rewound and rewound wires, of seamless and seamed construction. The results showed that rewinding had a significantly greater effect on the wire with a seam. Non-rewound wires showed greater arc voltage stability at a given welding current compared to rewound wires, for which a greater arc voltage spread was observed. This may indicate that the wire rewinding had a negative impact on its quality, e.g. through surface deformation, coating damage, microcracks, which resulted in process instability. Seamless wire also showed a noticeable deterioration in stability after rewinding, but to a lesser extent than the previous wire tested. Samples made with this wire, while still unwound, showed the most concentrated clusters of all the tests. Suban et al. [39] also found in their research that a smaller area of cyclograms indicates a more stable arc. This may indicate a more uniform wire surface, which is maintained even after rewinding. It is also the result of the material’s superior conductive properties, which is a direct stem effect of copper coating on its surface.

An additional quantitative test performed on the seamed wire, which is more susceptible to changes resulting from rewinding, was the measurement of diffusible hydrogen content in deposited metal. It was found that the average hydrogen content before rewinding was within the H10 level, which was also the manufacturer’s specification for this wire. Although after rewinding two of the measurements still showed a value below 10 ml/100 g, the average content was 10.35 ml/100 g, which exceeded the expected level. Cored wires are inherently nonhomogeneous due to their structure, which is a factor that may increase the risk of greater scatter in the obtained results. Both the increased diffusible hydrogen content in non-rewound Sample 1 and the decreased hydrogen content in rewound Sample 1 could have resulted from the initial stage of the welding process, which is often less stable. Additionally, for the non-rewound wire, the first sample was taken from the outermost layer, which, despite being stored under good conditions, could have been slightly more degraded, for example, due to mechanical wire defects formed during packaging and transport. However, for rewound wire, the first sample was made with wire that was originally deeper on the spool compared to subsequent samples. So, for rewound wire, the wire parts with mechanical wire defects were placed closest to the inside of the spool. The increase in hydrogen content caused by wire degradation as a result of wire storage, has been pointed out in earlier publications [35]. Rewinding might affect the properties of the wire surface, e.g., by causing microdamage that promotes chemical reactions, as well as damaging the seam, which may result in its opening and increase the absorption of moisture from the environment by the flux from the wire core, but based on the Mann-Whitney test, the differences between the averages of the two groups (rewound and non-rewound) are not statistically significant at the assumed significance level of 0.05.

A subsequent set of tests was conducted again on seamed and seamless wires, but in order to describe the effect of rewinding more accurately, additional factors were introduced: deliberate changes to the surface condition of the wires, as well as the use of changeable welding parameters. The use of stored and rewound wire causes arc instability with each set of parameters – this can be seen in both the welding current and the arc voltage. This indicates difficulties in carrying out the welding process - fluctuations in current and voltage may signify arc interruptions or the presence of contamination. The most common imperfections found in tested samples were spatters, irregular penetration, and porosity in the weld.

Various stability indexes quantitatively demonstrate a significant influence of rewinding on the welding process with seamed wire (Table 4). When comparing wire with a specific surface condition used for welding with the given parameters, it is clear that after rewinding, the stability is similar or worse. This is particularly evident through the S, CV, and CC indexes. Only the non-stored, non-rewound wire, when welded with the lowest parameters, showed a greater variation of results. This is probably mainly due to insufficient arc power, which is a factor that significantly deteriorates stability. However, after storage, the wire may have had slightly higher resistance, which increased the actual values of welding current and arc voltage. In this case, however, rewinding eliminated this slight difference, and after rewinding, the stored wire showed greater variation in results. Comparison of average and root mean square results provides an additional perspective on the variability of the recorded parameters. If the second of these indexes was significantly higher than the first, it would indicate that the average value was calculated from a data set with a large spread. In the case of seamed wire, there are almost no such differences. Only the rewound and stored wire, at low voltage, caused a slight increase in the root mean square current compared to the average current, but this increase was negligible. This confirms that despite the deterioration in stability, it did not disqualify the welding process.

The storage of wire with an non-copper-coated surface resulted in a slight increase in the average welding current, while for wire with a copper-coated surface it caused a slight decrease in the average welding current (Table 5). In the case of seamless wire, the rewinding process proved to be insignificant for stability index values. Welding with the new wire, regardless of the parameters and rewinding, was characterized by good stability and low variability in both welding current and arc voltage. This wire also slightly expanded the range of recorded parameters when a lower initial voltage was applied, but these changes were minor. After storage, the same trends could be observed – wires behaving proportionally before and after rewinding, as well as a slight deterioration in stability at lower initial voltage. There were no differences between the average and root mean square indexes for both arc voltage and welding current. The factor that most affected the stability of the tested seamless wire was the applied welding parameters.

In addition to the subjective authorial analysis, statistical analysis was used to facilitate the interpretation of the obtained results, with the corresponding p-values presented in Table 6. A significance level of 0.05 was assumed, which means that independent factors for which the p-values are lower than 0.05 are statistically significant. For all coefficients, statistical significance was found for two factors: wire type and arc voltage. This means that they had a direct impact on the stability of the welding process.

For both tested wires, the non-rewound, non-stored wire showed the highest stability. The same wire that had been stored still provided good stability for welding. Its deterioration was caused by rewinding. However, the rewinding effect was different for each wire. In the case of seamed wire, only rewound wire that had previously been affected by storage in unsuitable conditions clearly changed its characteristics. This indicates that the condition of the wire surface is an important factor in its ability to be rewound and then used for welding. For seamless wire, the differences in stability indexes were considerably smaller and less significant compared to seamed wire. The initial welding parameters affect the recorded values of current intensity and arc voltage, but the type of wire used dominates over changes in the initial parameters in terms of process stability. This additionally confirms the influence of the surface condition on the possibility of using rewinding for flux-cored wires. For each of the wires, it was found that the use of higher arc voltage and wire feed speeds, close to the upper limits suggested by the wire manufacturers, significantly increases the chance of achieving a stable welding process after rewinding, even for wires with a partially degraded surface.

The use of flux-cored wires is a direction that promotes increased welding efficiency and can potentially improve the quality of welded joints. However, this can only be achieved if the requirements of a strict technological regime are met, which is reflected, among other things, in proper handling of the consumables. It has been found that inappropriate storage conditions may affect the properties of both the wires [35] and the welds [36, 37]. So far, the literature has described that under urban conditions, features such as the diffusible hydrogen content, which can increase above 15 ml/100 g of deposited metal, wire stiffness, targetability issues, and changes in the electrical resistance of the wires may undergo changes [35]. The marine environment also clearly affects the appearance of the wires, their surface condition, the mechanical properties of the welds, as well as the type of weld fractures obtained [36, 37]. A special case among those studied was the long, two-year storage of wires in shipyard conditions, which showed the dynamics of changes in diffusible hydrogen content and the elongation of padding welds made after different storage periods, taking into account five different popular rutile flux-cored wires [40]. Based on previous articles it can be stated that the performance of flux-cored wires is closely influenced by their structure, with the manufacturing process, surface conditions, and seam structure playing a key role in their effectiveness. Consequently, these wires may be more susceptible to changes due to the rewinding process.

The results presented in this paper expanding the current state of knowledge show how sensitive flux-cored wires are to various factors, and consequently, how this affects the welding process. This broad perspective on the set of parameters and the condition of the wire helps to better understand the challenges faced by the industry transitioning from GMAW to FCAW, as well as to better understand the mechanism behind the disturbances occurring during arc ignition and burning. It should be emphasized that the scope of the study was limited to rutile-type FCAW wires for mild steel. Further research may be conducted in the direction of other wire types (metal-cored, basic) as well as other materials (high-strength steels and high-alloy steels).

The paper presents the effect of rewinding process on non-alloy steel rutile flux-cored welding wires. On the basis of the conducted research, the following conclusions were drawn: