Petroleum activities in the sensitive Arctic environment require increased focus on well integrity, since even small leaks can affect production and surrounding ecosystems. It is therefore of the utmost importance that the sealing ability of the annular well cement can be maintained here. This is challenging in normal locations, and difficulties are intensified when moving north. Due to the harsh topside conditions in the Arctic, the operational windows are short -and production will necessarily be turned on/off repeatedly. The temperature of any unheated injected fluid will also be lower here. As a result, Arctic wells will be subjected to strong downhole temperature variations over their life cycles. These cause the volume of well construction materials, like casing steel and annular well cement, to repeatedly expand and contract, which might lead to loss of well integrity through debonding or cracking of the annular cement sheath.In the present paper we describe an experimental laboratory set-up that has been designed for studying the sealing ability of annular cement as a well is exposed to thermal cycling. The samples studied are small-scale well sections including casing, annular cement and rock formation. These are exposed to thermal cycles by using a computer controlled thermal platform, which heats up by means of electrical resistance and cools down through expansion of liquefied nitrogen. It has a temperature span from -50°C to +200°C, and adjustable heating/cooling rates and holding times. During the thermal cycling experiments, any cracking and debonding occurring in the system is continuously monitored in-situ by Acoustic Emission (AE). To demonstrate the functioning of the set-up we present some initial results obtained using ordinary Portland G cement as annular sealant. In this work, the AE events collected during cycling are compared with data from post-experiment computed tomography (CT) scans.The testing methodology presented in this paper is flexible, thus rock type, annular sealant type and casing type can be varied at will. Mud or filter cake effects can also be included. For all samples, the procedure will enable determination of when leakage paths appear (as a function of applied thermal cycles and time), where they appear (in the bulk cement or at its interfaces) and what their sizes, geometries and distributions are. This opens for improved material choices for Arctic well construction, and optimization of operational patterns and remediation strategies for the high north. Most of today's Arctic research and development (R&D) aims to overcome the many topside challenges associated with petroleum operations in the north. These are of obvious importance, comprising extremely cold Arctic temperatures, large temperature variations, harsh weather conditions, drifting ice and ice loads, strong ocean currents, long periods of darkness and remote locations. In fact, the strong focus on topside challenges has led to a down-prioritization of the many subsurface challenges in the Arctic -which stil...