结论：过载保护要在真实的过载场景里验证，不是在实验台上通电
电机过载保护的作用是防止电机在超出设计范围的工况下损坏。但实验室里用可调电源模拟的过载，与轮椅在坡道上堵转时的过载，不是一回事。
真实场景里的过载伴随着机械负载、散热条件变化、电池电压下降，这些因素共同作用。所以验证应当尽量在真实工况下做，或者至少在模拟时把这些因素考虑进去。
典型的过载场景
| 场景 | 特征 | 风险 |
|---|---|---|
| 堵转 | 电机不转但持续通电 | 电流大、散热差，发热最快 |
| 长坡持续爬升 | 大电流持续时间长 | 热量累积 |
| 超载运行 | 载荷超出设计值 | 长期过载加速老化 |
| 机械卡阻 | 传动部件卡住 | 类似堵转 |
| 低速大扭矩 | 转速低导致散热差 | 温升快 |
| 环境温度过高 | 散热条件恶化 | 保护阈值容易触发 |
第一行是最严苛的。 堵转时电机转速为零，风扇或自然对流的散热效果最差，而电流接近最大值，温升速率远高于正常运行。保护如果不及时，几十秒就可能造成损伤。
保护策略的类型
常见的保护策略有几种，各有特点：
电流限制。 检测到电流超过阈值就限制输出。响应快，但单纯的电流限制无法区分短时大电流（正常起步）和持续大电流（堵转）。
电流时间积分。 综合考虑电流大小和持续时间，模拟发热的累积。比单纯电流限制合理，能允许短时大电流而在持续过载时动作。
温度直接测量。 在电机或控制器上装温度传感器，超温时保护。最直接，但传感器位置影响很大——装在外壳上测到的温度滞后于绕组实际温度。
温度模型估算。 根据电流和时间推算绕组温度。不需要额外传感器，但模型的准确性需要验证。
实际产品常采用组合策略，比如电流时间积分加温度测量互为补充。
验证方法
堵转试验。 机械上固定驱动轮使其无法转动，施加驱动指令，记录电流、温度和保护动作的时间。这是最直接的验证。
要注意的是记录点：在电机绕组附近测温度，而不只测外壳。 外壳温度滞后，用它判断保护是否及时会偏乐观。
持续爬坡试验。 在规定坡度上持续行驶，观察温度上升曲线和保护是否动作。这个试验贴近实际工况。
超载试验。 在超出标称载荷的条件下运行，观察保护行为。要注意的是载荷超多少——要有依据，通常参考实际可能出现的超载程度。
高温环境试验。 在环境温度上限条件下重复上述试验，因为散热条件变差，保护动作的时机会改变。
判定要看什么
保护动作了不等于合格。还要看：
动作是否及时。 在电机或控制器受到不可逆损伤之前动作。判断依据是绕组温度是否超过绝缘等级允许的范围。
动作方式是否安全。 保护动作时轮椅的行为——是平缓停止还是突然断电。突然断电在坡道上可能造成溜车（如果制动没同步生效）。
是否有提示。 使用者应当知道发生了什么，而不是以为产品坏了。
恢复方式是否合理。 温度降下来自动恢复，还是需要手动复位。自动恢复要有足够的回差，避免反复触发。
保护动作后的行为要求
这一点容易被忽略。保护动作后，系统应当：
确保轮椅处于安全状态（停止并制动）；给出明确指示；如果是可恢复的保护，说明恢复条件；记录保护事件，便于后续分析。
最后一项对售后很有价值。 如果系统能记录保护动作的次数和时间，售后就能判断使用者是否经常在过载条件下使用，从而给出针对性建议。
与产品规格的关系
过载保护的阈值设定与产品的标称规格相关。如果标称承重较高、标称爬坡角度较大，那么正常使用时的电流就接近保护阈值，保护容易误动作；反之如果阈值定得太高，保护形同虚设。
这个平衡要在选型和标定阶段确定，并用验证数据支撑。常见的问题是标称规格是市场部定的，而电机选型没有相应余量，结果保护在正常使用中频繁动作。
与制动系统的协同
过载保护动作时切断电机输出，此时制动系统的行为决定了轮椅的安全状态。如果制动是断电抱闸型，输出切断的同时制动生效，轮椅停住；如果制动需要主动指令，而控制器已经因保护进入异常状态，可能出现既无动力也无制动的情形。
这个组合在坡道上是危险的。 验证时应当在坡道条件下触发过载保护，确认轮椅不会溜车。
使用者可感知的过载征兆
在保护动作之前，产品通常已经有征兆：速度下降、声音变化、外壳发烫。如果使用者能识别这些征兆并主动减轻负载，就能避免保护频繁动作。
说明书里可以给出这类提示：遇到什么情况说明负载过重，应当怎么处理。 这比单纯依赖自动保护更好，因为保护动作本身对使用体验是负面的。
多电机系统的保护配合
双驱或多驱产品上，每个电机可能有独立的保护，也可能共用一套。需要验证的是配合关系：一侧电机保护动作时，另一侧如何响应。
如果一侧断电而另一侧仍在输出，轮椅会原地打转或偏转，在坡道或狭窄空间里是危险的。正确的设计应当是一侧保护动作时整车进入安全状态，而不是各自为政。
我们的做法
做过载保护验证时，我们会记录温度曲线而不只是保护动作的时刻。曲线能看出温升速率，据此可以判断保护的裕度够不够——刚好在临界点动作和留有余量动作，是两种不同的设计质量。
对于户外或多坡地区使用的产品，建议补做高温环境下的持续爬坡试验，这是实际中最容易触发过载的组合工况。
有需要可以把电机参数和保护策略发过来一起设计验证方案，或者直接联系：132 4819 8029。检测能力见服务介绍，产品分类见电动轮椅检测与代步车检测，标准信息见标准查询。
English version
Conclusion. Motor overload protection exists to prevent damage when the motor operates outside its design envelope. Overload simulated on a bench with an adjustable supply is not the same thing as overload experienced by a wheelchair stalled on a slope. Real overload comes with mechanical load, changed cooling conditions and falling battery voltage acting together, so verification should take place under realistic conditions or at least account for those factors.
Typical overload scenarios. Stall, where the motor is energised but not turning, produces high current with the poorest cooling and the fastest temperature rise. Sustained climbing produces high current for a long period and accumulates heat. Operation above rated load accelerates ageing. Mechanical jamming behaves like stall. Low-speed high-torque operation cools poorly and heats quickly. And high ambient temperature degrades cooling so that protection thresholds are reached sooner.
Stall is the most severe. With the rotor stationary, fan or convective cooling is at its worst while current approaches maximum, so the rate of temperature rise far exceeds normal running. Without prompt protection, damage can occur within tens of seconds.
Protection strategies. Current limiting responds quickly but cannot distinguish a brief high current during normal starting from sustained high current during stall. Current-time integration accounts for both magnitude and duration, approximating heat accumulation, and so permits brief surges while acting on sustained overload. Direct temperature measurement is the most straightforward but depends heavily on sensor placement, since a sensor on the housing lags the actual winding temperature. Thermal model estimation infers winding temperature from current and time without an additional sensor, but the model's accuracy needs verifying. Products commonly combine strategies, such as current-time integration complemented by temperature measurement.
Verification methods. Stall testing, mechanically restraining the drive wheel and applying a drive command while recording current, temperature and the time to protection, is the most direct approach. Measure temperature near the winding rather than only on the housing, since housing temperature lags and judging the timeliness of protection from it is optimistic. Sustained climbing on a specified gradient, observing the temperature curve and whether protection acts, is closest to real conditions. Overload testing above rated load requires a defensible choice of how much overload, normally referenced to what is plausible in service. And repeating these at the upper ambient temperature limit, since degraded cooling changes when protection acts.
What to judge. Protection acting is not by itself a pass. Judge whether it acted in time, before irreversible damage occurred, with winding temperature against the insulation class as the criterion. Judge whether the manner of action is safe: a controlled stop or an abrupt loss of power, since abrupt loss on a slope can allow the chair to run away if braking does not engage simultaneously. Judge whether the user is informed, rather than left to conclude the product has failed. And judge whether recovery is sensible, whether automatic once temperature falls or requiring manual reset, with sufficient hysteresis in the automatic case to prevent repeated triggering.
Behaviour after protection. The system should ensure the chair is safe, stopped and braked; indicate clearly what has happened; state the recovery condition where recovery is possible; and log the event for later analysis. Logging has real service value: a record of how often and when protection has acted shows whether the user routinely operates in overload conditions and supports specific advice.
Relationship to product specifications. Protection thresholds relate to rated specifications. A high rated capacity or steep rated gradient means normal operation runs close to the threshold and nuisance tripping becomes likely; set the threshold too high and protection becomes nominal. This balance belongs to the selection and calibration stage, supported by verification data. A familiar problem is rated specifications set by marketing without corresponding margin in motor selection, so that protection acts repeatedly during normal use.
How we handle it. We record the temperature curve rather than only the moment protection acts, because the curve shows the rate of rise and therefore how much margin the protection has; acting exactly at the limit and acting with margin represent different design quality. For products used outdoors or in hilly areas we suggest supplementary sustained-climb testing at elevated ambient temperature, which is the combination most likely to produce overload in service.
Send us the motor parameters and protection strategy and we will design the verification. Phone or WeChat: +86 132 4819 8029.