Influence of Heat Control on Welding Stresses in Multilayer-component Welds of High-strength Steel S960QL

Today high-strength structural steels (yield strength ≥ 960 MPa) are increasingly applied. Therefore, weldments have to achieve equal strength. Yet, high residual stresses in those welds diminish the components safety. Especially high restraint intensities can lead to crack-critical stresslevels. A special 2-MN-test facility allowed online-measurements of global reaction forces under defined restraint conditions during welding and cooling of multilayer-component MAG-welds. Local residual stresses were measured via X-ray diffraction before and after relief of the restraint. Local and global stresses were highly affected by heat control.


Introduction
In modern steel constructions more often high-strength steels are required to provide higher energy and cost efficiencies. Therefore, in the recent years steel producers developed various highstrength base materials and filler metals. As a result of the increasing material strength a slender design of components and moreover substantial material and weight savings are possible [1]. Whereas the requirements to the mechanical properties of the weld increase, the demands on the component safety are becoming increasingly important. To ensure the metallurgical requirements, a defined working range regarding the heat control must be maintained. Thus, an appropriate heat control provides the required micro structure properties and a sufficient hydrogen effusion as well as the avoidance of any crack initiations. For that purpose a convenient indicator, the cooling time from 800 °C to 500 °C (Δt8/5-cooling time), was established and generally accepted [2]. Various empirical approaches, nomograms e.g. based on the carbon equivalent and weld tests are available to evaluate a proper heat control [3]. High residual stresses are another contributor for crack initiating. Hence, a reduction and forecast of welding stresses should also be considered, especially for high strength steels. For an adequate heat control before, during and after welding, a better understanding about the interaction between heat control, material and restraint of the surrounding structures is necessary. However, most of recent research activities concerned with residual stress characteristics in welds of free shrinking laboratory samples, disregarding supporting effects and heat conduction terms of real component structures [4,5]. Several numerical and experimental studies revealed a high effect of restraint intensity and component dimensions on reaction stresses in welded structures [6][7][8][9]. These works conclude a superposition of local residual stresses and global reaction stresses due to restraint, cp. Fig. 1. Therein, the global restraint strongly affects the amplitude of the weld seams residual stress maxima and results in higher stress levels.  Furthermore, several recent analyses [10][11][12] involved the interaction between heat control and stress build-up in welds under global restraint. A strong influence of preheat and interpass temperature on the resulting reaction forces was found. However, expected welding stresses are usually considered by global safety factors in existing standards and technical guidelines. Continuative studies for an improved knowledge regarding stress forecast and reduction in high strength steel component welds are absent yet. Therefore, this present work analyses the effect of heat control on welding stresses considering realistic restraint conditions by a special test facility.

Experimental
Test Material. Plates of high strength quenched and tempered fine grained structural steel S960QL with a thickness of 20 mm were welded. A similar high strength solid wire according to ISO 16834 was used as filler metal. To provide a crack free root layer an undermatching weld metal was used. The chemical composition and mechanical properties are given in Table 1 and Table 2.  Weld Process and Parameters. All welds were performed with an automated MAG multilayerwelding process. The welding parameters are shown in Table 3. The heat control was varied by means of interpass temperature Ti, cp. Table 4. Regarding production relevant process conditions, the Δt8/5-cooling times were considered while selecting these parameters and determined for each specific weld test. The achieved Δt8/5-cooling times comply the producer specifications, cp. Table 4.

Residual Stresses IX
Weld Tests and Dimensions. A free shrinking test specimen was welded to get a reference for residual stress and weld properties. Fig. 2 shows the specimen dimensions and build-up sequence of the weld tests. An implementation of defined restraint intensity during welding and cooling was accomplished with a special in-house developed 2-MN-test facility. The restraint intensity in weld transverse direction RFy is the components stiffness towards the weld seam related to the seam length. For simple butt joints it can be estimated according to [13]. Realistic restraint conditions of RFy = 3 kN/(mm·mm) in transverse direction were realized. Therefore test specimens were clamped into the 2-MN-test facility, cp. Fig. 3. Run on-/off plates were tack welded to one side of the specimen couple to minimise influences of instable arc conditions during ignition and shut-off.

Measurements.
To monitor preheat and interpass temperatures type-K thermo couples were attached 10 mm adjacent to the weld. Additionally, the interpass temperature was observed with a contact thermocouple on the weld seam. Δt8/5-cooling times were determined by an optical temperature measurement at the weld layer surface using a two colour pyrometer (measuring range: 350 °C to 1300 °C). Using the test facility an online-measurement of the reaction forces due to restraint was achievable. Those forces and temperatures were observed while preheating, welding and subsequent cooling to ambient temperature. In the weld local transverse residual stresses were determined via X-ray diffraction using the sin 2 ψ-method on the top surface after cooling to ambient temperature. These measurements were performed before and after stress relief of the restraint. Fig. 4a shows the reaction force Fy(t) and temperature T(t) for test no. 2, cp. Table 4. At the beginning of the root weld the reaction force is -82 kN, since preheating of the tack welded specimen initiates a compressive force. During welding of the root a force build-up is obvious due to shrinkage of the already solidified inserted weld metal, which generates transversal shrinking forces. While the root weld is cooling down to interpass temperature the reaction force is increasing further with a first maximum of Fy = 121 kN at Ti = 100 °C. A force reduction was detected while welding the next layer, since a local heat input combined with stress relief and partial fusion of the root occurs. Subsequent cooling to Ti leads to a new continuous increasing of the reaction force and the next weld run to a force reduction. This force evolution with an increase of the force amplitude was observed for every weld sequence. The amplitude of the force reduction is the result of welding heat input and rises with each layer. Likewise, the reaction force build-up during cooling to interpass temperature grows with every weld sequence. After the completion of the weld and subsequent cooling to ambient temperature a maximum reaction force of Fy,end = 398 kN is reached. In Fig. 4b  higher interpass temperature due to shorter cooling cycles. A higher heat level thereby remains between each weld run and leads to a lower decrease of the transvers shrinking forces at each pass. At the last weld run the reaction force is half as high compared to weld including lower interpass temperature. After cooling to ambient temperature a remarkable increase of the reaction force for the higher interpass temperature occurs, since the specimen contains a distinct higher heat quantity. The resulting end reaction stress σy,end is about 30 % higher for Ti = 200 °C, cp. Fig. 4b and Fig. 7b. Local Residual Stress. Fig. 5a presents the transversal residual stress distributions across the weld for two different restraint conditions at the centre line (x = 0 mm) of the specimen. Both graphs show characteristic residual stress distributions for steels including phase transformation according to the common concepts [14]. Tensile residual stress maxima occur next to the weld centre line of the last weld run and side peaks at the weld metal next to the fusion line. In each weld run centre stress dips occur due to transformation. The tensile residual stress maximum of 650 MPa (70 % of Rp0.2 of the weld metal) and the overall stress level tends to be higher with global restraint applied. The average transverse residual stresses σ rs y,avg in the weld (y1 = -20 mm to y2 = 20 mm) are about 80 % higher compared to the free shrinking specimen, cp. Eq. 1 and Fig. 5b. In Fig. 6a two transverse residual stress distributions for different interpass temperatures at restrained condition are compared. The tensile stress peaks and levels tend to be higher when a higher interpass temperature is used. One reason is the increased temperature level of the weld while the top layer is welded due to shorter cooling cycles. Therefore, subsequent cooling to ambient temperature increases the inhomogeneous shrinkage restraint in the weld area and involves a higher corresponding tensile stress initiation. Comparing the resulting average transverse residual

478
Residual Stresses IX stress values σ rs y,avg an interpass temperature variation from Ti = 200 °C to 100 °C leads to reduction of about 20 %, cp. Fig. 6b. Another reason for this effect is described by the superposition of global reaction stresses as discussed in the next section. Stress Relief. Fig. 7a presents a comparison of transverse residual stress distributions before and after the stress relief of the restraint of test no. 2. The stress relief leads to an almost parallel shift of the transverse residual stress distribution. The estimated difference between the average transverse stress before and after stress relief revealed to be nearly the same as the determined end reaction stress with ca. 115 MPa for Ti = 200 °C, cp. Fig. 7b.

Summary
In the present study, the influence of heat control on welding stresses in high-strength steel component was analysed. Plates of 20 mm thickness were multilayer-MAG-welded under free shrinkage and defined restraint conditions. The interpass temperature was varied between 100 °C and 200 °C. Online-measurements allowed an observation of the occurring reaction forces while welding and cooling. The local residual stress determination was performed using X-ray diffraction at weld test specimen. From this work, the following conclusions can be drawn:  The heat control significantly affects the reaction stress. Reaction force amplitudes increase with each run during the welding process and are obviously lower with higher interpass temperatures.  However, low interpass temperatures lead to reduced reaction stress after subsequent cooling in accordance to recent studies [10][11][12].  In this work a first quantification of this effect was accomplished for high strength steel components of S960QL.  Based on extensive recent studies concerned with the influence of heat control on reaction stresses under restraint conditions [11,12] in this work local welding stresses were involved.

Advanced Materials Research Vol. 996
 With a new approach of average transverse residual stresses influences of heat control and restraint condition on these stresses were analysed. Therefore, a comparison of total welding stresses of specimens welded under restraint and free shrinkage was achieved.  The superposition of global and local welding stresses leads to higher residual stress levels under restraint conditions. The increase of average transverse residual stresses by approximately 100 MPa is almost equal to the resulting reaction stress.  Moreover, an adapted heat control revealed to be a sufficient way to lower average tensile residual stresses by more than 20 %.  The local residual stress determination after stress relief revealed an almost parallel shift in the range of the measured reaction stress.