Reinforced concrete (RC) structures are deteriorated with time when they are subjected to the aggressive environment. In order to maintain the safety and serviceability, most structures need the appropriate maintenance program to be applied during their service life. In 2002, Japan allocated approximately 13.5 trillion yen, which is 21.5% of the total construction budget, to the maintenance projects of the existent infrastructure. The ratio of maintenance budget to the overall construction budget is expected to continuously increase in the future because of increasing number of aging structures. In the near future, infrastructure maintenance will become the major task which has to be significantly concerned

Corrosion of reinforcing steel due to salt attack is one of major mechanisms deteriorated RC structure around the world. Due to expansion of corrosion product of reinforcing steel, pressure is generated inside the concrete and covering concrete is under tensile stress. Corrosion crack, spalling, or failure of structure is the results of steel corrosion. In order to ensure safety and serviceability of deteriorated structure, maintenance planning program, which is related to inspect current structural condition, predict future structural condition, and decide necessary action to be performed, has to be conducted.

Currently, there are many standard specifications of maintenance program for RC structure proposed by various organizations as well as various researchers. Due to variations of structural properties as a reason of workmanship, material properties, or environmental conditions are normally observed in the reality. Therefore, most of current maintenance program uses deterministic model to predict future condition of structure and is normally specified safety factor to cover the variations of structural properties and ensure the safety of structure. This leads to over-design and high cost of structure than the actual requirement. Some researchers also proposed maintenance program with deterioration prediction model stochastically based on Markov process. The Markov process assumes that progress of deterioration only depends on the current structural condition and neglects the improvement actions taken in the past. As well as, database of deterioration rate of similar structure and environmental conditions are required as an input in the prediction. So its application is limited to only the similar group and environment of structures. Therefore, in this study, a new probabilistic based deterioration prediction model is proposed. Benefits of both deterministic model of deterioration prediction and stochastic model of variation of structural properties in reality are combined in the method proposed in this study.

The main objective of this study is to propose a maintenance planning program based on optimization of life cycle cost that considers actual uncertainties of structural properties, and environmental conditions; considers deterioration mechanisms of structure both before and after being repaired. Variation of deterioration degree and probability of failure can be calculated. As a result, expected repairing cost and failure cost can be estimated. Finally, life cycle cost is determined and maintenance planning is decided based on the intervention that shows the minimum life cycle cost. However, scope of this study is still limited to RC structure deteriorated by salt attack as it is main mechanism rapidly deteriorated RC structure.

Firstly, deterioration prediction model for chloride induced steel corrosion of existing RC structure and repaired RC structure are proposed. Various effects such as crack width, macrocell corrosion, and performance of repairing system are considered to predict the future structural condition along the service life. Chloride diffusion along the crack width, macrocell corrosion due to non uniform chloride distribution, and corrosion crack width propagation are considered in the deterioration prediction model. Surface coating, patching repair, cathodic protection, and its combination are considered as options to repair deteriorated RC structures. For surface coating, durability against chloride penetration and cracking of surface coating are discussed. For patching repair, durability against chloride penetration as well as macrocell corrosion due to different in chloride concentration and material properties are considered. For cathodic protection, service life of cathodic protection is considered and corrosion can be neglected during the service life of cathodic protection.

Inspection program is recommended to determine actual variation of structural performance and environmental conditions. Parameters to be inspected are recommended based on result of sensitivity analysis. The minimum number of sample to be inspected are recommended based on required level of confidence, acceptable level of error, and ratio of target portion to the whole structures. Goodness of fit test is used to determine the most suitable probability density function in order to define the inspection result and use in the calculation.

Monte Carlo simulation is selected to solve the reliability problem. Probabilities of failure of structure due to chloride induced corrosion are determined. Limit states that are considered include corrosion initiation, corrosion crack width, and concrete spalling. Inspection results of actual structure are considered to be utilized as an actual variation of structural properties in the determination of probability of failure. Number of evaluation times of Monte Carlo simulation is tested in order to produce reliable and reproducible result. Probability of failure and variation of deterioration degree can be determined annually.

Life cycle cost is used to decide the most suitable maintenance planning. In this study, repairing cost and failure cost are considered as life cycle cost. Repairing is conducted when corrosion crack width reach the limited value. Repairing cost is composed of fixed repairing cost, variable repairing cost, and annual repairing cost. Fixed repairing cost is assumed to be fixed at every time of repairing. Variable repairing cost is assumed to vary with the predicted variation of deterioration degree. Annual repairing cost is also considered in case of cathodic protection that is required the maintenance of the system and electric power. History price of repairing cost is used in the calculation. Failure cost is composed of user cost and cost of death. User cost is calculated from time value and user loss time that relating to variation of deterioration degree. Cost of death is included economic loss of death of a person. Total repairing cost and failure cost are determined throughout the service life of structure with also considering the effect of discount rate. Maintenance methods, schedule of maintenance, cost, and discount rate are considered to affect the result of maintenance planning. The schedule of repairing, method of repairing, expected repairing cost can be obtained from the result of this study. Both of maintenance planning and budget allocation can be achieved by using the method proposed in this study.

Finally, examples of application of method proposed in this study to decide maintenance program for the actual RC structures are given. Actual inspection program is conducted and obtained results are used in the prediction. The lowest life cycle cost of scheduled and method of repairing can be decided for each case studies. Different of life cycle cost between different repairing method, and different durability of structure are shown in the result.

However, result of optimization of maintenance program largely depends on repairing and failure cost. There is also a difficulty to obtain an update and reliable cost data. So it is strongly recommended that database about repairing and failure cost relating to deteriorated structure should be focused. In order to determine actual variation of structural properties, the suitable inspection methods have to be considered. However, there is still difficulty to inspect resistance of concrete against penetration of aggressive ion. Develop of inspection method to be reliable inspect this property with affordable resources should be considered. Corrosion of reinforcing steel is actually time-dependent mechanisms. Due to formation of rust, changing in temperature or resistance of concrete, corrosion current can be affected. In the future, these factors should be considered in the prediction model. Moreover, there are many repairing methods newly developed repeatedly. Long term durability of those repairing methods should be carefully confirmed as many repairing methods found to accelerate deterioration of structure after being repaired. Spalling is one of failure that can cause severe damage to users and its prediction model for spalling of covering concrete has not been well developed yet. Deterministic model to predict spalling should be developed in the future. Effects of macrocell corrosion on corrosion crack width are significant. However, in this study only the different in material properties and chloride concentration are considered. There are other factors such as distance between cathodic and anodic area, time dependent properties of macrocell, etc. should also be considered in the future. Finally, not only salt attack that significantly deteriorates RC structure, but also other mechanisms such as carbonation, freezing and thawing, or loading should also be considered. Maintenance planning program for those deterioration mechanisms should also be proposed. In the final, integration of all mechanisms into one maintenance planning program is the final intention.

高度経済成長期に急速に整備された社会資本ストックは,2001年度現在,約405兆円であり,これまでの維持管理および更新費用を実績として計算した場合,2005~2025までの20年間に必要とされる費用は,約135兆円にのぼると試算されているこのようなことからも明らかなように,膨大な社会資本ストックを抱える我が国においては,効率的に構造物の維持管理を行うことが必要不可欠である。このような現状に対して,社会資本ストックを効果的・効率的に運用することを目的として,アセットマネジメント技術の社会資本ストックへの応用に関する検討が実施され始めてきている。ここでは,アセットマネジメントの確立に向けた第1段階として,ライフサイクルコスト(以下,LCC)の最小化を目的とする維持管理の最適化を定義している。現在では,LCCを考慮した維持管理対策に関する研究も,各研究機関において活発に検討が行われているが,LCC算定の基本情報となる構造物の劣化予測に関しては,未だに多くの問題を抱えている。実際の構造物の劣化は,様々な不確定な要因によって支配されるため,その劣化の進行は同一構造物内においても一様でないが,これらの不確定な要因を考慮せずに確定論的に予測しているのが一般的である。このため,幸いなことに我が国では未だ橋梁の崩落事故は発生していないが,最近のアメリカや中国のような事例のように,突如,落橋する危険性を推測することができない。また,最近では,構造物の維持管理の重要性が認識されてきているため,定期的な点検を各管理機関が作成しているマニュアルに沿って実施されてきている。しかし,構造物の劣化は,構造物の品質および環境条件によって支配されるものであり,構造物毎の特異な劣化状況を把握するためには,画一的なマニュアルに沿っても適切な点検結果を得られない場合がある。このため,点検結果がその後の劣化予測をするためには不十分である場合や,逆に必要以上の点検を実施している場合もあるなどの問題がある。当然のことながら点検は費用を要するものであり,我が国全体の社会資本ストック量を考慮すれば,強弱のある点検を実施することが極めて重要であるといえる。本論文は,これらの問題を解決すべく,様々な不確定な要因を考慮した確率論的手法に基づき維持管理計画の策定方法を提案するものである。

第1章は序論であり,本研究の背景と目的および論文の構成を取りまとめている。

第2章は,既往の研究を取りまとめたものであり,特に,最近の各国で提案されている維持管理計画策定方法に関する様々な方法論に関し,特徴を対比しながら取りまとめている。維持管理計画策定のために必要となる,劣化予測モデル,費用分析手法,信頼性解析等に関して詳細に調査している。

第3章は,劣化予測の基本情報となる点検計画の策定方法に関して,塩害劣化を受けるコンクリート構造物の場合に関して説明を行っている。特に,感度解析の結果を活用して,劣化予測に多大な影響を及ぼす点検結果の抽出,信頼性の高い点検結果を得るために必要なデータ数の決定方法を提案している。

第4章は,本研究で用いる劣化予測モデルの詳細を説明している。従来,予測手法が殆ど無かった,ひび割れの影響,マクロセル腐食,補修の効果を予測する手法を提案している。

第5章は,不確定性を考慮した劣化予測手法の提案を行っている。点検結果から得られた実際の結果を活用し,モンテカルロシミュレーションを用いて,構造物の損失確率を計算する手法を提案している。提案した手法を用いて,様々な補修方法が損失確立に及ぼす影響の評価を行っている。

第6章は,構造物のライフサイクルコスト,主に補修費用と損失費用の計算方法の提案を行っている。提案した手法を用いて,補修方法,補修回数,補修費用,割引率,維持管理費用の制約条件などの各条件が,計算結果に及ぼす影響を検討している。これにより,ライフサイクルコストを最小とする維持管理計画が策定可能となる。

第7章は,提案した手法のケーススタディーを行っており,タイ国にあるRC構造物を調査し,その結果に基づいて,将来の維持管理計画の提案を行っている。

第8章は,本研究で得られた成果および今後の課題を取りまとめている。

以上,本研究は,実際の構造物を効果的・効率的に維持管理するために必要となる,点検計画の策定,劣化予測,ライフサイクルコストの算出方法,を提案し,構造物の管理者が維持管理計画を策定する際の意思決定に資する情報を提供することを可能にしたものであり,有用性に富む独創的な研究成果と評価できる。よって本論文は博士(工学)の学位請求論文として合格と認められる。