1 Introduction

Pumped-storage hydropower plants are becoming increasingly significant for a reliable supply of industrial countries with electrical energy, especially since the so-called “energy transition” in Germany, currently enabling an economic storage of surplus electrical energy. Welded penstocks are key components of such plants at which the welding technology, however, plays a crucial role.

Figure 1 shows the historical development of materials and joining technologies which have already been successfully applied in penstocks for more than 100 years. Table 1 reviews the main properties of the used materials numbered in Fig. 1. For the last 50 years, quenched and tempered fine-grained steel plates are applied, whereas, thermomechanically rolled steel plates has become increasingly used over the past 25 years as well. However, the steel grade of type S690Q has currently become the preferred option. Detailed reports of the developments and experiences made with these materials and their welds can be found in [1–6].

Fig. 1
figure 1

Historical development of materials and joining technologies for pressure shafts in Central Europe [1]. The materials designations are further given in Table 1

Full size image
Table 1 Properties of steel grades for penstocks in pumped storage hydropower plants [1] (* current standard: DIN EN 10025–3)
Full size table

2 Initiation

During a subsequently performed material testing on samples, which have been removed from welds of a completed penstock of a pumped-storage power plant just before commissioning, an unexpected embrittlement of the weld metal has been discovered.

The penstock with a diameter of 5500 mm has been manufactured from multiple pipe segments of rolled steel plates of grade S690Q, each 3 m in width with wall thicknesses between 30 and 47 mm, whereas, every single segment has been produced by two longitudinal weld seams. Two to three pipe segments have been connected with circumferential weld seams to specific units, which then have been exposed to a post weld heat treatment (PWHT) in an electrically heated furnace prior to installation into an inclined shaft. The welding has been performed by means of the submerged arc welding (SAW) procedure by using seamless copper coated basic, matching flux cored filler materials of type AWS A5.23: F11A4-EC-F5.

After on site welding, the installed units within the inclined shaft have been connected with circumferential weld seams by means of the metal active gas (MAG) process towards lugs from the inner side of the pipes. Then, the pipes have been encased in concrete without PWHT.

The observed poor toughness of the already mounted and PWHT-treated welds questioned a safe service behavior of the entire plant. For this reason, a solution has been sought in order to develop applicable procedures for recovering the toughness of the brittle weld deposit in the inclined shaft. Since the longitudinal and circumferential welds have only been accessible from the inside, a repair through weld removal and repeated welding was not an option because a possible break-through of the steel wall might have caused water intake which then might have prevented quality re-welding.

For time and cost reasons, a full replacement of the existing inclined shaft or even a design of a new tunnel was no option. Apart from that, a fabrication of a sleeving new pipe (pipe in pipe) has been considered to be unfeasible as well. This variant was not taken into consideration due to time and cost reasons, but also due to hydraulical reasons with regard to the already installed turbine and pumping systems.

3 Causes of embrittlement

First, the root causes of the occurred embrittlement had to be addressed before searching for possible solutions to this problem. The analysis of the manufacturing process showed that acceptable toughness values have been reported during the respective welding procedure qualification (WPS) tests. However, it was very conspicuous that slow cooling rates of 15 °C/h have been applied during the described PWHT procedure (peak temperature 570 °C and 2 h holding time) in contrast to the specified 50 °C/h. Subsequent laboratory experiments on representative test welds (Fig. 2a) confirmed that slow heating and cooling rates have a strong influence on the weld materials toughness. This observed behavior of the test weld material is illustrated in Fig. 2b. It can be observed, that with specified cooling, much lower transition temperature is achieved.

Fig. 2
figure 2

Influence of annealing treatment. a Stress relief annealing curves with heating and cooling rates of 50 °C/h (blue) and 15 °C/h (red). b Results of Charpy-V impact tests at different removal temperatures during slow cooling of 15 °C/h (solid symbols) in comparison with 50 °C/h (open symbol)

Full size image

The investigations additionally confirmed that only the cooling rate had a major impact on the materials’ tendency to embrittlement. The heating and the holding time at peak temperature, on the other hand, did not play a significant role. This result suggested that the respective effect was temper embrittlement. This phenomenon has already been known to occur on heavy forgings and is described in detail in literature [7, 8]. Furthermore, various investigations have already been performed for the weld material as well [9]. Stein [10], for instance, published an extensive review of the temper embrittlement susceptibility of low-alloyed steels and presumed that temper embrittlement is primarily induced by the segregation of alloying and impurity elements such as Sb, Sn, As, or P. According to the author, reducing the amount of the latter would be the most beneficial, since the atomic concentration of P is much higher than for the other elements. However, the detailed mechanism is not clear up to now, and is subject of further investigations. Currently, P segregation along grain boundaries seems to be a significant driving force for the observed embrittlement [11].

Fracture surface analyses on Charpy impact specimens, which have been removed from the original welds of the penstock, clearly indicated the observed temper embrittlement effect. Examinations performed by scanning electron microscopy (SEM) revealed high proportions of intergranular fracture, Fig. 3a. This appearance is typical of temper embrittled steels. After the cause of the occurred embrittlement of the weld material has been identified, the aim was to provide potential solutions to this problem, taking into account the aspects mentioned above.

Fig. 3
figure 3

Fracture surfaces obtained by scanning electron microscopy on Charpy-V impact specimens. a Intergranular fracture of as-received sample (Charpy-V toughness of 8J at −20 °C). b Ductile fracture after a short heat treatment at 660 °C for 60 s (Charpy-V toughness at −20 °C: 140J)

Full size image

4 Reversible temper embrittlement (RTE)

The effect of temper embrittlement has been well known for decades in the literature and that it can be reversed by quickly heating the samples to a peak temperature of around 600 °C followed by a subsequent fast cooling (reversible temper embrittlement, RTE) [10, 12]. Based on this knowledge, samples of the brittle weld materials have been removed from the penstock and have been exposed to a short-time heat treatment in order to investigate its influence on the materials toughness behavior. The results of these experiments as a function of peak temperature and holding time are presented in Fig. 4. The positive impact on the recovery of the Charpy-V toughness therefore deduced the expected RTE effect [1]. Figure 3b shows the impact of the RTE-treatment on the fracture surface appearance of the Charpy-V specimens revealing clear ductile fracture proportions.

Fig. 4
figure 4

Increase of Charpy-V impact toughness at −20 °C after short RTE-heat treatment [1]

Full size image

Systematic investigations using the Gleeble simulator have been performed to quantify appropriate parameters for the practical application of the RTE-treatment in the existing inclined shaft. The respective results are illustrated in Fig. 5. It can thus be derived, that a short-time heat treatment between 550 and 750 °C of the investigated brittle weld metal causes the toughness to become recovered as requested.

Fig. 5
figure 5

Results of Charpy-V impact tests on specimens which have been heated to the respective peak temperatures, held for 15 s and subsequently cooled to room temperature with a t8/5-time of 200 s (Gleeble tests) by free cooling

Full size image

5 Practical application of the RTE-treatment

After verifying the basic effects of the RTE-treatment, continuing research primarily concentrated on selecting adequate heat treating parameters to be practically applied in the inclined shaft on-site. Different approaches including gas heating, inductive heating as well as laser heating have been intensively investigated. However, the well-proven electrical resistance mat heating process has been found to be the most appropriate method for this purpose, Fig. 6.

Fig. 6
figure 6

Mat heating preliminary test setup (left) on a 40 mm thick embedded plate, (right) during heat treatment both without insulation

Full size image

Extensive qualification activities have been performed by specifying a minimum required temperature within the entire volume of the weld deposit during the heat treatment while maintaining a maximum temperature to prevent overheating and its negative consequences on the materials properties. A quality assurance program for the RTE-treatment has been developed based on standard quality assurance measures in welding. After qualification of the process, an industrial application of the selected procedure has been successfully started to recover the toughness of the brittle welds in the existing inclined shaft. Approximate heating and cooling rates are between 135 and 500 °C/h under different conditions applied.

6 Conclusions

6.1 For welding-related processing

If high-strength weld materials are used, it is of crucial importance to appropriately qualify the heat treating parameters in the WPS such as heating rate, peak temperature, holding time, and, above all, the cooling rate when applying PWHT. Additionally, potential changes in the materials toughness must be verified on fully representative test coupons which have to be attached to the component to be heat treated on-site. This is the only way to ensure that any unexpected and undetected embrittlement may occur in the component, potentially leading to serious problems during service.

6.2 For manufacturers of filler materials

It is strongly recommended to examine the weld materials embrittlement sensitivity during PWHT by careful investigations. Moreover, these results shall be published in data sheets and guidelines in order to ensure a safe use of the products.

7 Outlook

  • The aspects described in 6.1 and 6.2 should be considered within the regulation and standards.

  • Welding residual stresses in tough materials are reduced once plastic deformation takes place. Therefore, it should be discussed whether a PWHT on high-strength steels and weld metals with their currently achievable toughness levels is necessary as prescribed in the regulations so far.

  • Performing on-site PWHT on large components, such as on manifolds and inclined shafts, is in many cases very complex and difficult to execute. This raises the question whether higher stresses due to temperature differences unavoidably remain in the component rather than relieving the stresses by means of PWHT.