This paper was adecfed bf pmwnatbn by an SPE Program CommilbeMowing rewiw of intormafii contained In an abafracf 8ubmHfed by the author(s). Contenm of the paper, as presented, have ml taen reviewed by lfw SOcbly of Pelrobum Enginearn and are subjecf to oarreclbn by the author(s). The nwwbd, nY pmsemted, do6a not necewrily Ntbcf any podtii of the SocbIY of Pefrofeum Engineers, W officam, w mentws. POP6M wewnfed at SPE meetirw am 8@cf to publfcatbn revisw by Edbbl Cammlffwa of the society d~~-. Pwmiwbnmqb~mm*M&~wW~tis.lll_tiw Mb~.~ti_*Mtin~wkõ f where and by whom the M b presented. Write Librarian, SPE, P.O. Sox 8S38SS, Rbfwdson, TX 75wHSSE, U.S.A., Tefex 16324S SPEUT.

This paper was selected for presentation by an SPE Program Committee following review of information contained in an abstract submitted by the authorp). Contents of the paper, as presented, have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author@). The material, as presented, does not necessarily reflect anv Dosition of the Societv of Petroleum Enaineers. its officers. or members. P a w s oresented at SPE meetinas are subiect to oublication review bv Editorial Committees of thiSocietv of b9troleum Eng~neers. 6ermission to copy is restricted to en abstract of not more'than'300 words. Illustrations may not be copied. i h e abstract should contain conspicuous acknow~ed~me~ of where and by whom the paper is presented. Write Librarian, SPE, P.O. Box 833836, Richardson, TX 750883836 U S A . Telex, 730989 SPEDAL. AbstractGas kick experiments in oil-and water-based mud have been studied in a full scale inclined research well. One of the main objectives from these experiments has been the study of gas rise velocities. In order to analyse the gas velocities, the parameters varying during a real drilling operation have been systemized and gathered in a data base. The work represents a significant extension to existing correlations for gaspiquid flows in a full scale inclined annulus with real drilling muds.In high concentration gas kicks the gas rises faster than in low and medium concentration kicks, this is obsemed for both oil and water based mud. The rise velocity correlations obtained from these experiments is not significantly dependent on gas void fraction, mud density, inclination, mud rheology, and surface tension. The results are quite different from other previously reported correlations, and this has a major implication on kick simulations.

This paper was selected for presentation by an SPE Program Committee following review of information contained in an abstract submitted by the authorp). Contents of the paper, as presented, have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author@). The material, as presented, does not necessarily reflect anv Dosition of the Societv of Petroleum Enaineers. its officers. or members. P a w s oresented at SPE meetinas are subiect to oublication review bv Editorial Committees of thiSocietv of b9troleum Eng~neers. 6ermission to copy is restricted to en abstract of not more'than'300 words. Illustrations may not be copied. i h e abstract should contain conspicuous acknow~ed~me~ of where and by whom the paper is presented. Write Librarian, SPE, P.O. Box 833836, Richardson, TX 750883836 U S A . Telex, 730989 SPEDAL. AbstractGas kick experiments in oil-and water-based mud have been studied in a full scale inclined research well. One of the main objectives from these experiments has been the study of gas rise velocities. In order to analyse the gas velocities, the parameters varying during a real drilling operation have been systemized and gathered in a data base. The work represents a significant extension to existing correlations for gaspiquid flows in a full scale inclined annulus with real drilling muds.In high concentration gas kicks the gas rises faster than in low and medium concentration kicks, this is obsemed for both oil and water based mud. The rise velocity correlations obtained from these experiments is not significantly dependent on gas void fraction, mud density, inclination, mud rheology, and surface tension. The results are quite different from other previously reported correlations, and this has a major implication on kick simulations.

This paper was adecfed bf pmwnatbn by an SPE Program CommilbeMowing rewiw of intormafii contained In an abafracf 8ubmHfed by the author(s). Contenm of the paper, as presented, have ml taen reviewed by lfw SOcbly of Pelrobum Enginearn and are subjecf to oarreclbn by the author(s). The nwwbd, nY pmsemted, do6a not necewrily Ntbcf any podtii of the SocbIY of Pefrofeum Engineers, W officam, w mentws. POP6M wewnfed at SPE meetirw am 8@cf to publfcatbn revisw by Edbbl Cammlffwa of the society d~~-. Pwmiwbnmqb~mm*M&~wW~tis.lll_tiw Mb~.~ti_*Mtin~wkõ f where and by whom the M b presented. Write Librarian, SPE, P.O. Sox 8S38SS, Rbfwdson, TX 75wHSSE, U.S.A., Tefex 16324S SPEUT.

SPE Members *Now with Petec A/S **Now with Baker Hughes INTEQ Abstract A number of gas kick experiments have been performed in a 2000 m long, 60 inclined research well. During the experiments a several of parameters were varied, such as mud density, mud type, gas concentration, and control rate of the kick. The well was heavily instrumented during the experiments, both downhole and surface. The gas distribution at start of the kick has been computed, and compared to the gas distribution out of the well which was calculated based on, measurements of the mixture density through the choke line, choke pressure and pit gain. This was done for the majority of the 24 kicks which were performed. The effect of mud density and solubility on the return distribution has been analysed. Effects of operational procedures during the kicks have been analysed as well (circulation rates and duration of close-in). The work shows that the return gas rate and gas distribution depends on the following variable; the initial gas distribution, mud type (OBM or WBM), and operational history of the kick. Optimal operational control procedures which will create the most favourable return gas distribution are recommended. Introduction Gas kicks occur seldom during drilling operations. However, for exploration drilling the kick frequency is higher than the average, and for some special wells (like high pressure, high temperature wells in the Central Graben area of the North Sea) the frequency of kicks is extremely high (approx. 2 per well). The main reason for (his is the small difference between the pore pressure and the fracture pressure, in combination with the extreme conditions which makes the control of the well pressure more difficult than in "normal" wells. Kick incidents are generally poorly documented, and the data usually consist of pit volume, shut-in pressures and choke and pump pressure. The initial influx distribution is not logged, only the volume (pit gain). The return gas distribution is not logged either. So data on how the gas distribution change from kick start to end, and how operational factors influence this have been non-existent. In the DEA-E-9 project, 24 gas kick experiments were performed, and data from a number of surface and downhole sensors were logged [1]. The data from the experiments have been extensively analysed [2, 3]. This paper summarize the analysis of the gas distribution for the experiments. GAS KICK EXPERIMENTS In the gas kick experiments performed in the DEA-E-9 project, a number of data was logged. The gas was injected through a coiled tubing which was run into the drill string down to the bit. An overview of the kick experiments is given in Appendix B. The experiments were performed in a 2020 m long research well with a maximum inclination of 63. Nitrogen and Argon gas was injected, and parameters such as mud type, mud density, gas concentration, mud flow rate, gas injection depth etc. was varied. Details of the experiments and the analysis of the data have been documented. GAS INFLOW RATE Knowledge of the gas flowrate into the well is essential to perform a detailed analysis of the kicks. The complete equation for the gas flow into the well is: (1) P. 481^