It　is　difficult　to　detect　and　evaluate　the　structural　damage　in　a　shield　tunnel　during　operation　because　many　traditional　techniques　based　on　the　observation　of　vibrations　are　limited　in　daily　monitoring　in　tunnels.　Thus,　the　curvature　radius　of　a　static　longitudinal　settlement　curve　is　used　to　identify　the　residual　health　and　safety　of　an　in-service　shield　tunnel.　However,　there　are　still　two　problems.　The　curvature　radius　is　suitable　for　a　qualitative　judgment　rather　than　a　quantitative　evaluation　for　longitudinal　damage　detection.　Moreover,　the　curvature　radius,　which　is　calculated　from　the　measured　settlements　of　three　neighboring　points,　gives　an　average　damage　degree　in　a　wide　scope　only　and　is　difficult　to　use　to　identify　the　damage’s　precise　location.　By　means　of　the　analysis　of　three　kinds　of　longitudinal　failure　modes　in　a　shield　tunnel,　this　paper　proposes:　(1)　a　damage　detection　method　based　on　the　monitored　increment　of　the　neutral　axis　depth;　and　(2)　an　index　to　evaluate　longitudinal　damage.　The　index　is　composed　of　the　residual　ratios　of　the　equivalent　flexural　stiffness　(HFM1)　and　the　equivalent　shear　stiffness　(HFM3).　The　neutral　axis　position　and　the　proposed　damage　index　can　be　determined　using　long-gauge　Fiber　Bragg　Grating　sensors.　Results　from　numerical　simulations　show　that　the　deviation　between　the　HFM1　and　the　true　value　residual　ratio　of　the　equivalent　flexural　stiffness　is　no　more　than　1.7%.　The　HFM3　is　equal　to　its　true　value　in　the　entire　damage　process.　A　loading　experiment　for　a　scaled-down　model　of　a　shield　tunnel　using　long-gauge　Fiber　Bragg　Grating　sensors　indicated　that　the　errors　in　the　HFM1　were　no　more　than　5.0%　in　the　case　of　early　damage　development　(HFM1　≥　0.5).　The　maximum　error　did　not　exceed　9.0%　even　under　severe　damage　conditions　in　the　model.　Meanwhile,　the　HFM3　also　coincided　with　its　true　value　in　the　entire　testing　process.　©　2019　by　the　authors.　Licensee　MDPI,　Basel,　Switzerland.