Particular diligence in the preparation of documentation and conducting a tender procedure for construction works must be made by public contracting entities, who are subject to additional restrictions. In Poland, the largest public contracting entities are the sectoral ones. These are the entities (defined in Art. 3 of the Public Procurement Law Act) operating in the sectors of water management, energy, transport and postal services. The paper presents the analysis of tender offers for the execution of construction works or design and execution of construction works related to the reconstruction, construction and revitalization of railway lines, announced after the amendment to the Public Procurement Law on June 22, 2016. The considered examples are selected tender procedures covering the scope of construction and assembly works on railway lines throughout the country. The authors paid particular attention to the tender criteria applied and the requirements for the potential contractors for works in the field of railway infrastructure.

Every decision or action taken as part of a construction project involves risk. Unforeseen branch works that may occur during the construction investment are the so-called additional work. They cause risk, both for the contractor and the investor. Skilful management of this risk may lead to minimizing the change in the investment duration or minimizing the change in the cost of the contractual amount. The work proposes a method of analysing the risk of industrial works that may occur during additional works in railway construction investments. A constructed Bayesian network based on the risk component of industrial works was used for the analysis. Bayesian networks are listed as one of the 31 techniques suggested for risk analysis in accordance with the ISO 31010 standard, which enables the correct analysis of the examined problem with satisfactory accuracy. During the construction of the network, historical data was obtained from completed and settled railway infrastructure construction projects, and 125 unique records corresponding to additional works were identified. The created Bayesian network combines technological aspects resulting from the specificity of the implementation of branch works in railway construction projects with a practical assessment of their risk. The proposed network model allows for risk analysis by defining various event scenarios, and has high application capacity resulting from the ease of applying its results in practice in the implementation of railway investments.