The growing demands for gas storage capacity encouraged the Polish Gas and Oil Company to build underground gas stores in salt deposits located in the north-eastern part of Poland, in the area bordering on the Gulf of Gdańsk (South Baltic Sea). The construction site is located about 4 km away from the Baltic Sea coast. The drilling of boreholes and diluting of salt rock was proposed as a method of creating the chambers. The aim of the paper is to analyze the mixing processes due to a brine discharge into the marine environment in order to confirm the efficiency of the constructed installation. Assessment of brine mixing can be carried out using different approaches: laboratory and/or field experiments, analytical methods, modelling. The laboratory approach, the most common one (e.g. Abraham 1967; James et al. 1983; Roberts & Toms 1987, 1988; Roberts et al 1997; Cipollina et al. 2005; Shao & Law 2010; Bashitialshaaer et al. 2012), focuses on the analysis of a single jet shape as a function of selected parameters, such as initial velocity, angle of incidence, a density difference between effluent and ambient water. The majority of laboratory experiments was carried out for stagnant water (e.g. Zhang & Baddour 1998; Shao & Law 2010; Papakonstantis et al. 2011), however, a few experiments for non-stagnant conditions in the ambient environment were also described (e.g. Gungor & Roberts 2009; Roberts et al. 2011; Lai & Lee 2014). Field experiments are quite rare due to difficulties associated with their execution. In marine conditions, simultaneous registrations of all parameters of interest usually require a support of a vessel and divers. It should be emphasized that data measured in-situ always represent a unique, non-repeatable, situations. Despite all difficulties and limitations, field measurements are an invaluable source of information on all natural processes. Field measurements focus mainly on mixing processes in the far field of the installation (e.g. Marti et al. 2011).
Considerable interest in the mixing processes is related to the brine discharged from desalination plants. Brine produced in the desalination process is usually characterized by the salt concentration below 100 kg m-3, whereas the present study deals with the discharged brine characterized by the concentration of salt reaching 250 kg m-3. Due to environmental reasons, the discharge located in Puck Bay, the area protected by the Natura 2000 program, is limited to 300 m3 h-1 to meet the restrictions imposed by local administration to reduce the salinity excess to 0.5 PSU in the near-field of the discharge installation.
The Gulf of Gdańsk is situated in the south-eastern part of the Baltic Sea (Fig. 1a) in the area limited by the imaginary line between the Cape of Rozewie and the Cape of Taran (Fig. 1b). Puck Bay is a shallow, western sub-region of the Gulf of Gdańsk separated from the open sea by the Hel Peninsula. In the middle of the Hel Peninsula, there is a shallow sandbank (called Rybitwia Mielizna) stretching from the Rewa Cape to Kuźnica, which divides the Bay of Puck into two parts characterized by different circulation patterns: the eastern part, called the Outer Puck Bay (av. depth ~20.5 m), and the western part, called the Puck Lagoon (av. depth ~3 m). Currents in this non-tidal region are generated mainly by wind and the accompanying water level variations in time and space. In addition, the circulation in the Puck Lagoon is influenced by water exchange with the eastern part of the bay. The water circulation patterns were described by Nowacki (1993) on the basis of occasional field observations of currents. It is quite well documented that water circulation patterns in the Outer Puck Bay depend on the wind direction. The direction of surface currents usually follows the wind direction, although the opposite current direction can be observed in the south-eastern part of the Outer Puck Bay. The direction of bottom currents frequently follows the surface currents. However, the available field measurements of the currents are insufficient to present the flow patterns that are generated by spatially and temporarily varying wind conditions in the region. Wind statistics for coastal stations, Gdynia and Hel, show the prevalence of W-SW wind (i.e. Gdynia: SW – 18.7%, 4.2 ms1, W – 24.0%, 4.3 m s-1; Hel: SW – 15.6%, 5.1 m s-1, W – 22.4%, 5.5 m s-1; Kwiecień 1990). The maximum measured current in Puck Bay in the vicinity of Rybitwia Mielizna, 0.37 m s-1 at the surface and 0.34 m s-1 at the bottom, was associated with the northern wind (Nowacki 1993).
Water salinity in the Puck Bay is determined by interactions between marine and fresh waters. In the Outer Puck Bay, salinity is closely related to the inflow of more saline water from the Gdańsk Basin through surface and bottom layers. In addition, the surface layer is modified by fresh water from the Vistula River discharging into the Gulf of Gdańsk about 40 km away from the center of the brine discharge installation. In the Puck Lagoon, salinity depends mainly on the intensity of water exchange with the Outer Puck Bay. The influence of small rivers is limited to their outlets. Statistics indicate (Fig. 1c) that the mean monthly salinity in the Outer Puck Bay is significantly lower compared to the Puck Lagoon during periods of increased discharge from the Vistula River, i.e. early spring, because of snow melting. Long-term statistics (1951-1990) for the Vistula River show a mean annual discharge equal to 1080 m3 s-1, and the highest annual discharge equal to 3840 m3 s-1 (Fal et al. 1997).
Water density in the Puck Bay changes seasonally, like water temperature, with spatial modifications due to weather conditions (i.e. air pressure, wind) and bottom configuration in the analyzed region. Due to intensive mixing, stratification cannot develop in the Puck Lagoon and in the shallow parts of the Outer Puck Bay. In the eastern, deeper, part of the Outer Puck Bay, the density stratification can develop occasionally. In this context, the density stratification in the area of discharge was not taken into account.
The analysis of brine mixing in Puck Bay was carried for the constructed diffusing system located 2300 meters from the shore designed by the EMPORIUM design office. The discharge installation consists of 16 diffuser’s blocks (Fig. 2), each with 3 nozzles distributed every 120o of the circumference. The installation was constructed to discharge brine with the density (ρe ) of 1160 kg m-3 (salinity 215 PSU), salt concentration of 250 kg m-3 at a rate of 300m3 h-1 . The installation covers a rectangular area of 180 × 180 m.
The in-situ measurements were carried out from October 2010 to December 2012 to monitor mixing processes during the early functioning of the brine discharge installation. The monitoring program included three types of in-situ measurements: (1) continuous measurements of salinity, temperature, and water currents in two locations carried out by the use of autonomous Aanderaa Instruments rotor type current meters (A and B, Fig. 3); (2) CTD measurements taken by the InterOcean probe (conductivity, temperature, depth) in the near-field of the installation; (3) CTD measurements in the vicinity of a single block. Information on the quality and quantity of brine was made available by the owner of the installation, while wind conditions come from meteorological station in Gdynia (2010, 2011) or measurements carried out by the owner of the installation using an instrument located at a fixed platform in the center of the brine discharge installation (location S, Fig. 2a).
Continuous measurements of environmental conditions in the ambient environment, including salinity, temperature, and water currents 1 meter above the bottom, were carried out in two locations, A and B (Fig. 3), situated on the eastern and western sides of the installation. Because of the technology applied for the construction of the underground stores, the amount, the concentration of salt and the density of brine increased gradually until the summer of 2012, when the maximum values were reached. During the first months, brine was discharged only for a few hours a day. The operating time was gradually extended until, starting from the summer of 2011, the discharge was carried out continuously, with short breaks for the maintenance of the installation. At the same time, the concentration of salt and the density of the effluent were gradually increasing, reaching the target density of 1160 kg m-3 and the concentration of salt of 250 kg m-3 in the summer of 2012. Three periods, 13 October – 26 November 2010; 12 July –26 August 2011; and 22 May – 9 August 2012, represent three stages of brine discharge (early, intermediate, target), characterized by different salt concentration in brine and density. Sample results for those measurements are presented in Fig. 4. All the results represent periods of relatively weak wind conditions, when wind velocity was generally lower than 6 m s-1, and wind direction was varied. Water currents in the region are generated by wind. In such conditions, the measured flow velocity hardly ever exceeded 0.1 m s-1 (Fig. 4). It is characteristic that the predominant flow directions were to the north and to the south, which resulted from the local bathymetry and configuration of the coastline. It should be mentioned, however, that the measurements of currents conducted in the 1990s in similar locations, but under more severe wind conditions, showed velocities of up to 0.35 m s-1 (Jasińska 1996).
In the autumn of 2010 (the early stage), when brine was discharged for only a few hours a day, salt concentration in brine was relatively low (Fig. 4). At that time, the values of salinity in locations A and B were very similar, differing at most by 0.1 PSU. In the summer of 2011, the amount of brine increased to 250 m3 h-1, and the concentration of salt to 225 kg m-3. During periods of weak dynamics, salinity differences between the two gauges did not exceed 0.25 PSU. In the summer of 2012, the discharged brine reached the maximum permissible parameters (total discharge – 300 m3 h-1, concentration of salt – 250 kg m-3). Under such conditions, the maximum salinity differences between the two gauges occasionally reached 0.35 PSU.
The
Parameters of brine discharged during selected in-situ measurements
Date | Q (m3h–1) | ρ (kg m–3) | d (m) | wind direction (deg) | wind velocity (m s–1) | ||||
---|---|---|---|---|---|---|---|---|---|
arm 1 | arm 2 | arm 3 | arm 4 | total | |||||
27.10.2010 | 0.0 | 0.0 | 76.8 | 57.8 | 135.3 | 1055 | 0.008 | 225 | 2 measured in Gdynia station |
26.07.2011 | 75.5 | 55.9 | 0.0 | 61.5 | 192.9 | 1135 | 0.009 | 315 | 4 measured in Gdynia station |
24.01.2012 | 74.9 | 74.5 | 74.7 | 75.8 | 299.9 | 1164 | 0.009 | 315 | 2 measured in Gdynia station |
1.08.2012 | 74.0 | 67.0 | 78.0 | 76.0 | 295.0 | 1160 | 0.009 | 120 | 3.5 measured at a center of installation (location S) |
The collected data make it possible to analyze the salinity distribution in selected verticals (Fig. 5), and the spatial distributions in selected layers (Fig. 6) or cross-sections (Fig. 7). It is important to note differences in salinity values, especially in the upper part of verticals, between the results from selected dates. Those values represent natural, or almost natural, salinity values in that region. The observed salinity variations in time are associated with the inflows of water from the Gulf of Gdańsk to Puck Bay.
The most pronounced influence of brine discharge on the salinity distribution can be expected in the lower part of verticals located at the center and in the closest vicinity of the installation (locations S, 9, 10, 17, 18 – Fig. 3). In the presented cases, a full range of vertical profiles can be seen (Fig. 5), from a well-mixed one (27.10.2010) to a well-pronounced halocline (24.01.2011). In calm weather, the highest salinity value can be expected in location S; such a situation was observed on 24 January 2012. On 26 July 2011 and 1 August 2012, the highest values of salinity in the lower part of verticals were observed in some locations in the close vicinity of the installation. This salinity pattern may have been caused by water currents and a non-uniform distribution of discharge among arms.
On the selected dates, salinity distributions in the bottom layer (Fig. 6) and in two perpendicular cross-sections (Fig. 7) present four different situations. In the first case – on 27 October 2010 – the amount and the concentration of salt in brine were low, leading to small variation in salinity values at the bottom (7.14-7.24 PSU, Fig. 6a). However, owing to water currents, a “plume” of higher salinity was shifted in the north-western direction. The asymmetry of water salinity is quite well seen in vertical cross-sections (Fig. 7a). Similar distributions on 26 July 2011 (Figs 6b, 7b) present a rather unusual situation when salinity at the center of the installation was lower than in the surrounding area. In that case, the salinity differences at the bottom ranged from 6.88 to 7.28 PSU, while the minimum value in the water column did not exceed 6.7 PSU.
On 24 January 2012, a “plume” of increased salinity at the bottom (7.35-7.90 PSU) was located quite symmetrically in relation to the center of the installation (Fig. 6c). This observation is confirmed by salinity distributions in two perpendicular cross-sections (Fig. 7c). On 1 August 2012, a clear, elongated “plume” of increased salinity at the bottom (7.50-7.95 PSU, Fig. 6d) moved northward from the center of the installation. It is very characteristic that the salinity “plume” is limited to 1-1.5 meters in thickness and a few hundred meters in the horizontal plane (Fig. 7d).
The above examples make it sufficiently clear that measurements in a single vertical at the center of the installation do not provide adequate information on brine spreading in the surrounding environment. A full analysis must be based on measurements conducted in a number of verticals. The above results show the effect of water currents on the shape of the high-salinity “plume”, and demonstrate the close relationship between brine parameters and excess salinity patterns. The detailed analysis of salinity changes in Puck Bay in the period of October 2010 – December 2012 has shown (Robakiewicz 2014) that salinity excess hardly ever reached a value of 0.5 PSU.
Numerous laboratory experiments (e.g. Abraham 1967; James et al. 1983; Roberts & Toms 1987, 1988; Roberts et al. 1997; Cipollina, 2005; Shao & Law 2010; Bashitialshaaer et al. 2012) and some theoretical investigations (e.g. Oliver et al. 2013b) have been carried out to describe characteristic dimensions of a single jet (Fig. 8) and to assess the dilution of discharged brine in the ambient environment. The aim of the
In most cases, laboratory experiments were carried out for stagnant water (e.g. Roberts & Toms 1987, 1988; Roberts et al. 1997, Bashitialshaaer et al. 2012), whereas calm weather lasting more than a few hours is very unusual for the Puck Bay region. During the
The detailed CTD measurements, covering 71 locations in an area of 17 × 17 meters (Fig. 11) were carried out in the vicinity of the D1-1 block equipped with nozzles of 0.009 m in diameter. The position of each CTD vertical was registered by a DGPS gauge. On 26 August 2011, salt concentration in brine 237 kg m-3 was discharged through the D1 arm at a rate of 61.1-61.4 m3 h-1, which means that the initial velocity at the nozzle was 22.20-22.45 m s-1.
The spatial distribution of the maximum salinity values in verticals (Fig. 11a) shows three directions of salinity increase tracing the jets ejected from the D1-1 head. The spatial distribution of salinity in the bottom layer (Fig. 11b), obtained with the same data, clearly shows the area of increased salinity where a single jet reaches the bottom. The two above-mentioned distributions were used to select verticals for the presentation of the A-A vertical cross-section through a single jet.
The salinity distribution in Fig. 12 represents the A-A cross-section along the estimated axis of a single jet starting about 2 meters from the diffuser’s block; measurements closer to the block do not coincide with the jet axis (Fig. 11). The trajectory of a single jet determined by measurements can be compared with characteristic dimensions predicted on the basis of laboratory experiments and derived in a pre-investment study using the JET3D model (Robakiewicz & Robakiewicz 2008). It should be mentioned that the discharge conditions assumed in that study were quite different from those encountered at the time of in-situ measurements. The major difference (Table 2) can be seen in the exit flow velocity, which was smaller during the in-situ measurements as a result of an increased nozzle diameter by 0.001m. In addition, a gentle water movement was registered under natural conditions.
Comparison of brine discharge conditions assumed in the pre-investment study and actual brine discharge conditions.
Parameter | Assumed | Actual |
---|---|---|
Nozzle diameter d(m) | 0.008 | 0.009 |
Exit flow velocity U0(ms-1) | 30 | 22.2÷22.45 |
Density of effluent ρe(kgm-3) | 1160 | 1160 |
Density of ambient water ρa(kgm-3) | 1005 | 1005 |
Densimetric Froude number Fr(-) | 273 | 190-192 |
Angle of incidence(deg) | 45 | 45 |
Velocity in reservoir v(ms-1) | 0 | 0.02÷0.06 |
The measured jet of brine was characterized by Fr ~190, whereas most of the previous laboratory investigations were performed for jets characterized by Fr<100. Recent laboratory experiments of Bashitialshaaer et al. (2012), conducted under a wide range of conditions (Fr=21÷567), have shown substantial differences between characteristic dimensions of a jet in the cases of Fr<100 and Fr>100.
Formulas derived from laboratory experiments for Fr<100 (Cipollina, 2005; Bashitialshaaer et al. 2012) and from theoretical investigations (Oliver et al. 2013) were used to estimate the characteristic dimensions of a single jet. It appears that the results obtained by those formulas slightly underestimate the horizontal distances to the jet centerline maximum (Xy) and to the edge point (Xe), as well as the vertical distance to the trajectory centerline maximum (Y) (Fig. 13) in comparison with the
One of the main purposes of the discharge installation was to improve the mixing of brine with the ambient environment. To assess the effectiveness of mixing processes, the dilution coefficient is commonly used. It can be estimated as:
where
On the basis of measurements taken in Puck Bay, the dilution coefficient was estimated at
Values of the return/impact point dilution coefficient
Dilution coefficient | ||
---|---|---|
Laboratory experiments | ||
Nemlioglu & Roberts(2006) | Not started,boundary interaction | 323 |
Shao & Law(2010) | 0.05-0.47 | 239 |
Papakonstantis et al.(2011) | 0.37-1.39 | 295±26.6 |
Lai & Lee(2012) | 0.24-0.92 | 203 |
Analytical solution | ||
Kikkert et al.(2007) | No boundary | 182 |
Predictive models | ||
VisJet (Lai 2010) | No boundary | 148 |
CorJet(Jirka 2008) | No boundary | 123 |
JET 3D -condition as assumed in the pre-investment study |
1.37 | 360 |
Measurements | 1.75 | 458 |
Shao and Law (2010) noted the potential influence of the bottom boundary on the dilution coefficient. They found that the minimum source height, necessary to avoid the influence of boundary interaction at the return point, should satisfy the relation Hs/(Frd) > 0.74 (
The
Measurements of a single jet behavior in the marine environment are quite unique. So far the jet configuration has been investigated by analytical approaches and laboratory experiments. Most of them dealt with jets characterized by low values of the densimetric Froude number (Fr<100) discharged into stagnant water. Measurements taken in 2011 in Puck Bay represent a situation when water velocity was 0.02-0.06 m s-1 and the jet was characterized by the densimetric Froude number Fr=190. The results of field measurements were compared with estimates done using the formulas derived from laboratory experiments and analytical investigations in terms of characteristic dimensions of a single jet and the dilution coefficient. It was found that the estimated characteristic dimensions of a single jet discharged into stagnant water were underestimated compared to the those observed during the field experiment. The estimates of the dilution coefficient done using the formulas derived from laboratory experiments, predictive models and analytical solutions for stagnant water conditions vary in a wide range of values (123÷360). These differences can be attributed to the influence of the bottom boundary represented by the