Statistical analysis of past kicks and blowouts occurred in a Middle Eastern oilfield

In 2017, blowouts and then explosions occurred in a Middle Eastern oilfield. A root cause of the incident is lack of study and investigation of the past kicks and blowouts data. Therefore, as a pioneer work in the region, data gathering and analysis of past kicks and blowouts were made in the studied oilfield to learn lessons and find gaps. Out of the 149 drilled wells, a total of 117 kicks and three (3) blowouts occurred. In this work, a list of drilling parameters to be considered in data gathering and analysis were suggested as a guideline for future works elsewhere. The statistical analysis not only showed all the three exploration wells kicked which is not a surprise, but it also showed that 39 out of 146 (26.71%) also experienced kicks during reservoir drilling. The large number of kicks in development wells proved that possibility of kick occurrence in development wells is not low. In exploration wells, the predominant kick causes were gas-cut mud and insufficient mud weight which indicates the necessity of using pressure while drilling in addition to drilling rate control systems in exploration wells. However, in development wells, lost circulation was the predominant kick cause indicating the necessity of using low-weight drilling fluids and managed pressure drilling systems. The direct role of human error exists at least in 60% of kicks occurred in this field, which shows the great importance of improved drilling personnel training. Although only 3.45% kicks in development wells occurred due to improper hole fill-up during tripping, this cause should not only be deemed trivial, but it should also be taken seriously as being the cause of the blowout. The 2.56% possibility of kick conversion to blowouts and 67% risk of blowout conversion to explosion emphasize the necessity of maintaining primary well control and using efficient and early kick detection systems. Bullhead was the more commonly used method than standard well control methods; as this kill method may not always be safe, its application should be revised.


Introduction
Kick flows are considered serious drilling problems signaling loss of well integrity. If kicks are converted to blowouts, the consequences would be extremely uncontrolled in terms of fatalities, injuries, detrimental environmental impacts, loss of reputation and obviously cost of equipment lost and the required well control cost. Many blowouts have occurred during the drilling history. Etkin et al. (2017) listed 20 largest historical blowouts and considered the 1979 Ixtoc I and the 2020 Macondo as two example catastrophic ones. Grace et al. (2000) discussed a gas blowout in Taiwan; US-CSB (2019) discussed a blowout in Pryor Trust well 1H-9 in state of Oklahoma due to not filling in the hole during tripping in a horizontal well with oil-based mud ;St John (2016) discussed and compared the Macondo blowout (in Gulf of Mexico, GOM) and the Bardolino blowout (in North Sea); Khalifeh and Saasen (2020) briefly described a 2016's blowout in North Sea which was controlled before leading to explosion; Ashena et al. (2011) andNabaei et al. (2011) discussed two blowouts in onshore Iranian wells; Ashena et al. (2021a) discussed a blowout occurred in a development well of a Middle Eastern field in 2017, which was similar in several terms to that of Pryor Trust well. The loss of well control happened following sudden release of gas during tripping the drill string to the surface. Tripping began to change the bit at a depth of 2610 m (8563.4 ft) when drilling the reservoir formation. Due to the blowout and subsequent explosions, the drilling rig was burned, and two rig crew members lost their lives. Some well capping methods were simultaneously applied to extinguish the blaze and finally control the well. The financial implication of the blowout included loss of the super heavy rig of nearly $110 million and blowout control expenses.
Analysis of past kicks is crucial for several reasons. First of all, via statistical analysis of past kicks data, improvement of the drilling learning curve can be made to prevent future well control incidents (Editions Technip 1981). Second, through analysis of the kick data and artificial intelligence methods, early kick detection systems can be developed. In addition, kick data contributes to eliciting formation fluid properties and reservoir pressure study. It is highly important that the drilling crew report the kicks details and data as correctly and exactly as possible. However, in reality, such a proper recording is not always the case unfortunately in some field practices. Based on a statistical 10-year (1979-1988) analysis of kicks and blowouts in conventional wells of the USA, a large percentage of blowouts still occur in development wells. In addition, the same reference indicated that 51% of kicks and blowouts in development wells occur during tripping, which indicated the vivid role of human error role in the Reason's Swiss Cheese Model (Reason 1997(Reason , 2000, the frequency of blowouts in an exploration well was found to be 2.85 times that of a development well (Wylie and Visram 1990). In unconventional high-pressure deep-water conditions, this frequency proportion is 5.14 (Etkin et al. 2017). Several researchers emphasized the importance of development of early kick detection systems (Fraser et al. 2014;Jacobs 2015;Shaker and Reynolds 2020;Habib et al. 2020). Yin et al. (2019) used past kicks data in order to obtain a kick risk forecasting model. Dobson (2009) and Bybee (2010) analyzed the kicks and blowouts occurred in the UK continental shelf and linked them mostly to geological conditions at the wellsite, cementing issues and then to human error as a minor causal factor. Shi et al. (2019) adopted the random forest and support vector machine algorithms to detect a kick during drilling in real-time. Alouhali et al. (2018) trained and evaluated five models-decision tree, K-nearest neighbor (KNN), sequential minimal optimization (SMO) algorithm, artificial neural network (ANN), and Bayesian network-to detect kicks based on actual kick cases. Some researchers (Miska et al. 1996;Baldino et al. 2019a, b) used kick data to elicit formation fluid properties including pressure.
One of the root causes of the blowout occurred in 2017 (Ashena et al. 2021a) was lack of past kicks and blowouts statistical analysis. Therefore, in this work, statistical data analysis of the previous kicks and blowouts occurred in all drilling wells of the oilfield is conducted. In this analysis, the number and percentages of kicks due to several causes and the well control or kill methods are investigated. The occurred kicks are investigated for exploration and development wells separately. The role of human error in kick causes is quantitatively found. The oilfield has a very highly pressured gas cap which makes drilling in gas-cap really challenging. The reservoirs in this oilfield are naturally fractured carbonate formations. Following several decades of production, pressures of the reservoirs, particularly the shallower ones have depleted, which has put development drilling at high risk of lost circulation. Finally, some lessons are elicited out from the data for future drilling operations T, which can be extended to other fields particularly with naturally fractured formations.

Past kicks and blowouts data
Using previous drilling history records of each well, past kicks and blowouts which occurred in naturally fractured reservoirs of the investigated oilfield are gathered in Table 5 (in Appendix). The authors suggest that the following details and parameters be considered in kick data gathering for statistical analysis: • Well type: It is important to indicate the type of drilling wells: exploration (abbreviated as "Exp"), delineation ("Del") or development (Dev). Based on Norwegian Petroleum Directorate, an appraisal or delineation well is an exploration well drilled to define the extent of the reserves. Following the first three exploration wells drilled in this field, two delineation wells were drilled in the banks of the reservoir to determine or locate the boundaries of the reservoir. • Hole size: The hole size where kicks occurred was indicated consisting of 8 1/2, 8 3/8, 6 1/8, 6, 5 7/8, and 4 1/8-in.
The depth where kicks occurred is another important parameter which was gathered for each case. • Reservoir formation: There are two separate oil reservoir formations in this field. All the wells drilled into and penetrated only the overlying reservoir formation (abbreviated as A), except for the exploration well which drilled deeper and also penetrated into reservoir formation B. • Operation type: It is also crucial to record the type of drilling operation or phase which was ongoing at the time of kick occurrence. This could be either during drilling, tripping, BOP testing, etc. • Kick cause: The cause of the kick or the reason why the kick occurred is another important parameter which was found and collected for each kick. • Pit gain: It is important to record the kick influx volume immediately following well shut-in. Unfortunately, except for some exceptions, the pit gains were not recorded in daily drilling reports. • Well control method: There are several available well kill methods. However, it is important to select a suitable kill method for each specified kick. • Whether blowout or explosion occurred: Conversion of kick to blowout or later explosion is an important indicator of severity of a well control situation which was found by the authors.

Number of kicks and blowouts
Based on Table 5, it was found that 117 kicks occurred in the 149 wells drilled in the oilfield. Out of 117 kicks, 30 kicks occurred in three exploration wells and 87 kicks in the development wells (Table 1 and Fig. 1). Three kicks converted to blowouts, one of the blowouts occurred in an exploration well and two of them were in development wells. It is noted that another blowout occurred in a workover well; however, as workover wells are not discussed in this work, that blowout is not discussed here.
Following statistical analysis of the kick data in Table 1, the following points were reached:  Fig. 2). • In all the wells, 2.01% of them experienced blowouts and 2.56% of the kicks converted to blowouts; in exploration wells, 33% of them kicked and 3.3% of kicks converted to blowouts. In development wells, 1.36% of wells kicked and 2.3% of kicks converted to blowouts (see Fig. 3). • Out of three blowouts occurred in this field, one of them was in an exploration well which was successfully controlled, whereas the other two blowouts which were in development converted to explosions. Therefore, consid-ering all wells, 2/3 (66.67%) of the blowouts converted to explosions (see Fig. 3).

Well operations
Using the gathered data, it is possible to investigate during which specific drilling operations kicks occurred. The operations are as follows: • Drilling/cutting rocks • Tripping (either POOH or run in hole/RIH), • Condition trip or wash and ream, • Displacing mud or circulating mud,  Table 2 shows the numbers and percentages of kicks occurred during different well operations in exploration wells (also shown in Fig. 4) and in development wells (also shown in Fig. 5). According to Table 2 and Fig. 6, in all wells, most (46.15%) kicks occurred during "drilling" or cutting the rock. In exploration wells, following the "drilling" during which 63.33% of the kicks occurred, "Condition Trip or Wash and Ream" and "RIH" are the next operations when most kicks occurred with, respectively, 16.67% and 13.33% records. However, in development wells, following "drilling" (40.23%), "POOH" is the next operation when most kicks occurred (14.94%).  Table 3 lists the breakdown of kicks based on their causes. Figures 7 and 8, respectively, show percentages of kicks due to different causes, respectively, in exploration and development wells. In exploration wells, gas-cut mud due to gas-cap drilling (63.33%) and insufficient mud weight (30%) are the two most predominant causes (see Fig. 7). However, in development wells, lost circulation (37.93%), insufficient mud weight (21.84%), and cement slurry loss (17.24%) are the three predominant causes of kicks (see Fig. 8). Figure 9 shows percentages of kicks due to different causes in all wells. In all the wells, lost circulation (28.2%) and insufficient mud weight (23.9%) are the two predominant kick causes. The role of human error in the kick causes is quantitatively found. 1 3 Kick causes are briefly explained as follows: Insufficient mud weight When a permeable zone is drilled using a lower than enough mud weight which exerts less pressure than the formation pore pressure, an underbalance condition occurs resulting in kick occurrence. This kick cause is one of the predominant kick causes encountered in the world (Adams 1980), including in the studied field with 2.6% of all the kicks occurring due to this kick cause. In exploration wells, the percentages are greater than exploration wells with, respectively, 30% and 21.84%. It is noted that in development wells, insufficient mud weight is either caused by the drilling planning or the operational crew, and it is thus considered associated with human error.
Improper hole fill-up during pulling out of the hole Not filling the hole properly/adequately during tripping or due to mud loss is probably an important reason for kicks occurring in development wells. This kick cause is usually attributed to human error or neglect of the drilling crew. In exploration wells, no kicks occurred due to this cause; however, in development wells, 3.45% kicks occurred due to this cause. Overall, 2.6% of all the kicks occurred due to that.

Complete loss and improper hole fill-up
During drilling or tripping in naturally fractured formations, complete loss or lost circulation may occur. Although the crew are expected to proceed to filling-in the well with mud immediately, it may take a while to proceed, and due to possible underbalance, a kick influx may already enter the wellbore. In exploration wells, no kicks occurred due to this cause; however, in development wells, a large percentage of kicks (37.93%) occurred due to that. Overall, 28.2% of all kicks occurred due to this cause. This kick cause is associated with human error.

Gas-cut mud
During drilling of gas-bearing zones particularly gas-cap, the gas born on the cuttings is mixed with the mud and causes lightening of the mud. Sometimes, it is observed as "Gas Bubble" as an indication at the surface. Depending on the mud weight, it may cause underbalance and then a kick occurrence. This phenomenon usually occurs during drilling. However, if good mud circulation is not carried out particularly for heavy weight muds prior to pulling the drill string out of the hole, it may be observed after a while due to gas expansion during pulling out of the hole, or later during running the next drill string into the hole or during BOP testing. In drilling the first exploration well in the investigated oilfield, this was observed.
Another occasion that gas-cut mud may be observed is while drill-out of cement plugs. In this case, gas is accumulated below the cement plug (level-off or balance) which will be mixed with the mud to cause gas-cut mud or lightening of the mud upon drill-out. Therefore, in such cases, it is recommended to drill-out the cement plug with a high mud weight to counteract possible lightening effect of the gas. In exploration wells, most kicks (63.33%) occurred due to this cause; however, in development wells, only 5.75% kicks occurred due to that. Overall, 20.5% of all the kicks occurred due to this cause. In development wells, this kick cause is associated with human error.

Surge and inducing fractures
Due to too quickly running in the drill string into the hole, the formation may become fractured or further fractured (in naturally fractured reservoir formations), which would cause mud loss. When mud loss becomes excessive, it may not be possible to control it quickly, and the problem is replaced by another problem (i.e., kick). In exploration wells and development wells, similar percentages of kicks occurred due to this cause with, respectively, 3.33% and 1.15%. Overall, 1.7% of all the kicks occurred due to this cause. This kick cause is associated with human error.

Swabbing
Due to too quickly pulling the drill string or other assemblies out of the hole, a pressure drop or an underbalance condition may be resulted particularly in heavy and viscous muds, causing a kick influx entering the wellbore.
Following kick detection, if pressure rating of the wellhead and the downhole tubulars allow, bullheading the influx back to the formation may be proceeded. Otherwise, it is recommended to strip-in the well (with the annular preventer being closed), then apply the Driller's method by which the kick influx is circulated out of the hole (in the first circulation). In the meantime, mud weight is increased to proceed to killing the well in the second circulation. In exploration wells, there were no kicks due to this cause; however, in development wells, 5.75% of the kicks occurred due to that. Overall, 4.3% of all the kicks occurred due to this cause. This kick cause is associated with human error.

Cement loss
Cement slurry loss may occur during primary cementing or during remedial cementing or cement plugging. In either case, due to the resultant underbalance condition, a kick influx would enter the wellbore. In case of partial mud loss or sometimes severe mud loss in the reservoir formation, a balance cement plug may be spot in the loss zone. In case of complete loss, a level-off cement plug is set to cure the loss (Ashena et al. 2021c). In exploration wells, no kicks occurred to this cause; however, in development wells, 17.24% of kicks occurred due to that. Overall, 12.8% of all the kicks occurred due to this cause. This kick cause can be associated with human error, but in this work, its association has been ignored.

Insufficient cement slurry weight
The other kick cause is cementing with insufficient or lower than enough cement slurry weight to balance formation pressure. This would result in an underbalance condition and consequently a kick influx. This can be either due to issues in planning or due to a pump truck issue. No kicks with such a cause occurred in this oilfield.

U-tubing during cementing liners
At the end of cementing liners, some cement slurry may rise up the liner lap. Therefore, after escaping the cement slurry (raising the drill pipes out of the slurry), due to u-tubing, the slurry would u-tube from the annulus into the drill pipes and show itself as a flow. Although this flow is not a kick flow, the annulus should be essentially filled with mud to prevent any possible underbalance. If the annulus is not filled properly, there is a high risk of kick occurrence. No kicks with such a cause occurred in this oilfield.

Pumping salt water or pipe-lax to release stuck pipes
If the drill string is stuck, an option is to pump salt water or pipe lax to release the drill string. Pipe-lax is mixed with diesel and is then pumped to spot on the stuck area. Due to its density, an underbalance is created in the wellbore, and a kick flow is intentionally allowed to occur with the purpose of releasing the drill string or elimination of the stuck. Next, the kick influx is bullheaded back to the formation. Similarly, in case of using salt water, if the density of salt water is lower than the mud weight, kick flow may occur. In exploration wells, only 3.33% of kicks occurred due to this cause; however, in development wells, a slightly larger percentage of kicks (6.9%) occurred due to that. Overall, 6% of all the kicks occurred due to this cause.
As a summary of this section, in exploration wells, the predominant kick causes were gas-cut mud and insufficient mud weight indicating the necessity of using pore pressure, while drilling systems and drilling rate control in exploration wells. However, in development wells, lost circulation was the predominant kick cause indicating the necessity of using low-weight drilling fluids and underbalanced or managed pressure drilling systems in such depleted reservoirs. Human error, either in planning or operation phase, is directly involved in kick causes consisting of improper hole fill up during lost circulation or POOH, surge and swab, and insufficient mud weight (for development wells). Therefore, as shown in Table 3, direct role of human error exists in 3.33% of kicks in exploration wells and 69% of development wells, or overall, 60% of all kicks occurred in this field, which shows the great importance of improved drilling personnel training in the future.  Well control methods Table 4 shows the well control methods used to control kicks based on their specified kick causes. Figure 10 generally presents the breakdown of kicks based on different well control methods used to put an end to all the kicks occurred in the studied oilfield. Based on the figure, most kicks (39.2%) were controlled by the bullhead method. The next more commonly used well control methods were wait and weight (15.4%) and Driller's method (13.7%). The well control methods are briefly explained as follows: Circulation to balance mud weight In case of gas-cut mud during drilling gas-bearing formation, a simple method of controlling the situation is by circulating the mud with the purpose of balancing the mud weight. Therefore, by one or more circulations bottoms-up, it is expected that the entrained gas (which has entered the mud) be displaced out of the hole. This method is abbreviated as "Circ to Balance." Almost one fifth of all the kicks (20.5%) were controlled by this control method.

Increasing mud weight
If the kick influx is minor in volume, it may not be necessary to shut-in the well in case of drilling in non-reservoir formations. However, most drilling companies mandate well shut-in in reservoir drilling due to possible oil and gas flow hazards of exposure to the reservoir. However, a drilling company may not mandate well-shut-in as is case in this field; instead, the crew will proceed to increasing the mud weight gradually (in stepwise manner) while keeping mud circulation. As shown in Fig. 10, 6% of all the kicks were controlled by this control method.

Bullhead
If due to complete loss and not filling in the annulus, kick occurs, a quick well control method is by bullheading the kick influx back into the formation. As high pressure is applied from the surface, this method can be used provided that the BOP, wellhead and tubulars pressure rating allow that and if the formation rock is rather dense and does not allow penetration of the influx back. Advantages of this method are its quick effect, controlling the well without necessity of handling the hydrocarbons at the surface, and its applicability in case of being off-bottom. The operator and the drilling companies involved in this field consider bullhead as the first well control method to be applied. Most kicks (39.3%) were controlled by this method.
If bullhead is not possible, e.g., due to limited pressure rating of BOP, etc., it is essential to strip-in the hole (with the annular preventer closed) and then apply the driller's method; displace the gas out in the first circulation, and then in the second circulation, increase mud weight to kill the well.

Driller's method
In some cases, pressure rating of tubulars may not allow bullheading. Therefore, if the drill string is on-bottom, the driller's method may be used. This method is also called the "Two-Circulation" method by some authors such as Well Control Methods Adams (2006). 13.7% of all the kicks were controlled by this method.

Wait and weight method
In the reservoir, it is strongly recommended by many drilling companies to shut-in the well even for small pit gains (say 2-3 bbl). For considerable pit gains (> 10 bbl), it is standard by almost all drilling companies to shut-in the well first and then either practice the driller's method or wait and weight method. Wait and Weight method is preferred over the driller's in some cases including when the pressure rating of the BOP and wellhead is not high and requires quick action. This method is also called "One-Circulation" method by some authors such as Adams (2006). 15.4% of all the kicks in the studied field were controlled by this method.

Volumetric method
The volumetric method (also called "bleed-off" or "lubricate and bleed") is considered an unconventional well control method. It is usually applied when the drill string is far off-bottom or out of the hole, and when bullheading, the influx back to the formation is not possible. However, this well control method has some calculation complexities. The involved drilling companies in this field used this well control method just in special cases and preferred it than stripping-in the drill string back to the bottom and then using Driller's or Wait and Weight method. Only 5.1% of all the kicks were controlled by this method.

Relief well(s)
In case the kick is not regained by using conventional and unconventional well control methods, it is most likely converted to a blowout. In case of a blowout, the loss of well control is usually regained by drilling relief wells as a tertiary well control method. This technique is sometimes called the "bottom-kill" method. Only one kick (0.9% of all kicks), which was transformed into the blowout occurred in 2017, was controlled by this method.

Summary and conclusions
Following the blowout that occurred in 2017 in this Middle East field, lack of investigation of the past kicks and blowouts data was recognized as one of the root causes of the incident. Therefore, in this work, following scrutiny into drilling records, 117 kicks and three (3) blowouts were found to occur during drilling the reservoir formations of total 149 wells in the studied oilfield. The following conclusions were drawn from this statistical investigation: 1. The proper drilling parameters to be considered in data gathering and analysis were mentioned as a guideline for future works elsewhere. 2. The number of 39 wells out of total 146 development wells (26.71%) kicked during reservoir drilling, most of which occurred during drilling and tripping. Contrary to expectations, more blowouts occurred in development than in exploration wells. This emphasized the high possibility of kicks and blowouts occurrence in development wells. Kick possibility during tripping in development wells should not be ignored, and careful monitoring of the crew is vital to prevent catastrophes like the blowout in 2017. 3. In exploration wells, the predominant kick causes were gas-cut mud and insufficient mud weight, which indicates the necessity of using pressure while drilling systems and drilling rate control in exploration wells. However, in development wells, lost circulation was the predominant kick cause in the depleted naturally fractured reservoir, which indicates the necessity of using low-weight drilling fluids and underbalanced or managed pressure drilling systems in similar reservoirs. 4. The direct role of human error exists at least in 60% of kicks in this field, which shows the great importance of improved drilling personnel training in the future. 5. The 2.56% possibility of kick conversion to blowouts and 67% risk of blowout conversion to explosion in all wells shed another light to the necessity of maintaining primary well control and using efficient and early kick detection systems to prevent kicks. 6. Bullhead was the more commonly used method (~ 40%) than standard well control methods "Driller's and Wait and Weight" (~ 30%). This represents that well kill policy is traditionally set not to let the kick influx reach the surface; however, this kill method may not always be safe and should be replaced with standard ones. 7. As a gap in recorded kick data, pit gains were not mostly recorded in the drilling and well history records. As this parameter is of great importance for improved statistical analysis on severity of well control events, it is advisable to be strictly recorded in the future.