diff --git a/aucma-report/boardAPIComplete.md b/aucma-report/boardAPIComplete.md
index f44168a..d57c026 100644
--- a/aucma-report/boardAPIComplete.md
+++ b/aucma-report/boardAPIComplete.md
@@ -725,7 +725,7 @@ public class Board5ProductionTopVo {
 | orderWorkStatistics | DMS_BILLS_FAULT_INSTANCE, DMS_BILLS_MAINT_INSTANCE, DMS_BILLS_INSPECT_INSTANCE | 工单总数=维修+保养+巡检；已完成=维修完成(BILLS_STATUS='2')+保养完成(MAINT_STATUS='3')+巡检完成(INSPECT_STATUS='3')；停机=维修中(BILLS_STATUS='1') |
 | repairTimeStatistics | DMS_BILLS_FAULT_INSTANCE | 平均响应时间=(REAL_BEGIN_TIME-APPLY_TIME)*24*60分钟；平均维修时间=(REAL_END_TIME-REAL_BEGIN_TIME)*24小时；仅统计BILLS_STATUS='2' |
 | maintStatistics | DMS_BILLS_MAINT_INSTANCE | 按MAINT_STATUS分组统计：1=待保养，2=保养中，3=已完成，4=待验证（无此数据返回0） |
-| inspectStatistics | DMS_BILLS_INSPECT_INSTANCE | 应检设备=COUNT(DISTINCT PLAN_INSPECT_ID)*18；已检设备=INSPECT_STATUS='3'的数量 |
+| inspectStatistics | DMS_BILLS_INSPECT_INSTANCE, DMS_PLAN_INSPECT | 应检设备=关联计划表SUM(DEVICE_AMOUNT)；已检设备=INSPECT_STATUS='3'的数量 |
 | deviceStatusDistribution | BASE_DEVICELEDGER | 按DEVICE_STATUS和DEVICE_TYPE分组统计，WHERE IS_FLAG=1 |
 | faultSourceDistribution | DMS_BILLS_FAULT_INSTANCE | 按FAULT_SOURCE_TYPE分组：1=维护故障，2/3=运行故障，9=自然故障；WHERE IS_FLAG='1' |
 | faultTrendList | DMS_RECORD_ALARM_INFO | 近30天每日告警数量，按ALARM_BEGIN_TIME日期分组 |
@@ -791,15 +791,23 @@ FROM DMS_BILLS_MAINT_INSTANCE
 #### 4.2.4 巡检执行情况统计
 
 ```sql
-SELECT (SELECT COUNT(DISTINCT PLAN_INSPECT_ID) * 18 FROM DMS_BILLS_INSPECT_INSTANCE WHERE IS_FLAG = '1') AS DEVICE_COUNT,
+SELECT (SELECT NVL(SUM(p.DEVICE_AMOUNT), 0)
+        FROM DMS_PLAN_INSPECT p
+        WHERE p.PLAN_INSPECT_ID IN (
+            SELECT DISTINCT PLAN_INSPECT_ID
+            FROM DMS_BILLS_INSPECT_INSTANCE
+            WHERE IS_FLAG = '1'
+        )
+       ) AS DEVICE_COUNT,
        (SELECT COUNT(*) FROM DMS_BILLS_INSPECT_INSTANCE WHERE IS_FLAG = '1' AND INSPECT_STATUS = '3') AS COMPLETE_COUNT
 FROM DUAL
 ```
 
 **业务逻辑说明**：
-- **应检设备数** = 不同巡检计划数 × 18（每计划覆盖18台设备）
+- **应检设备数** = 关联 `DMS_PLAN_INSPECT` 计划表，汇总 `DEVICE_AMOUNT`（每个计划覆盖的设备数量）
 - **已检设备数** = 已完成的巡检工单数（`INSPECT_STATUS = '3'`）
-- **数据来源**：`DMS_BILLS_INSPECT_INSTANCE` 巡检工单表
+- **数据来源**：`DMS_BILLS_INSPECT_INSTANCE` 巡检工单表 + `DMS_PLAN_INSPECT` 巡检计划表
+- **关联原因**：工单表无设备ID字段，需通过计划表获取设备数量信息
 
 #### 4.2.5 设备状态分布
 
@@ -1041,8 +1049,9 @@ export function getProductionTop5() {
    - 巡检工单：`DMS_BILLS_INSPECT_INSTANCE`（WHERE IS_FLAG='1'）
 
 2. **❗重要：巡检应检设备数计算**
-   - 应检设备数 = COUNT(DISTINCT PLAN_INSPECT_ID) × 18
-   - 每个巡检计划覆盖18台设备（固定值）
+   - 应检设备数 = 关联 `DMS_PLAN_INSPECT` 表，汇总 `DEVICE_AMOUNT`
+   - 每个巡检计划的 `DEVICE_AMOUNT` 可能不同，不能硬编码 × 18
+   - 原先的计算逻辑 `COUNT(DISTINCT PLAN_INSPECT_ID) × 18` 导致完成率可能超过 100%
 
 3. **保养状态映射**（DMS_BILLS_MAINT_INSTANCE.MAINT_STATUS）：
    - 1 = 待保养
@@ -1081,3 +1090,4 @@ export function getProductionTop5() {
 | 2026-01-21 | 1.2 | 产量机台TOP5替代备件消耗费用TOP5 | YinQ |
 | 2026-01-22 | 1.3 | 故障数量曲线切换为DMS_RECORD_ALARM_INFO表 | YinQ |
 | 2026-01-26 | 2.0 | 根据Mapper XML更新业务逻辑：工单统计改用工单表汇总、保养/巡检改用工单实例表、更新所有SQL示例和表结构 | Claude |
+| 2026-01-26 | 2.1 | 修复巡检完成率超过100%问题：应检设备数改为关联DMS_PLAN_INSPECT表SUM(DEVICE_AMOUNT)，替代硬编码×18 | Claude |