HELLO CHINA

INDUSTRY CLUSTER · Automotive manufacturing

Wuhan Hannan intelligent connected NEV components

Hannan component work becomes difficult when fuel cells, battery packs, body parts and controllers meet in one vehicle and every test still has to point to an exact hardware and software revision.

Formation and regional foundations

Vehicle stations define interfaces and cycle time, power systems carry safety limits, sensing and control hardware preserve revisions, and tooling holds dimensional consistency. Bench, environmental and in-vehicle retesting connect those responsibilities.

Fuel-cell systems, traction batteries and body parts may pass their own stations yet expose new conflicts in cooling, communications, dimensions and control timing when fitted to one vehicle. Replaying a vehicle fault shows how Hannan connects bench traces, tooling lots and software calibration to one vehicle identity.

Core geography: A Hannan manufacturing network linking new-energy power systems, automotive electronics, body and chassis parts, and vehicle validation in Wuhan Economic and Technological Development Zone

Six revision interfaces on one validation vehicle

Which record returns a vehicle fault to its component

Vehicle, hardware and calibration identities span all six stations; a changed identity redraws usable tests from the point of change.

  • Vehicle interface freeze sheet — Envelope, voltage class, cooling interfaces and communication protocols are signed separately on the freeze sheet, so every later change retains an owner and an effective vehicle number.
  • Fuel-cell system bench trace — The fuel-cell bench places stack current, compressor draw, hydrogen pressure and coolant temperature difference on one second-by-second trace, preventing net power from becoming an isolated number.
  • Traction-battery lot history — Incoming cell lots follow module positions into the pack history and align with sampling harnesses, thermal sensors and the BMS revision, allowing an exception to return to one physical location.
  • First and last body parts — First and last parts retain tool number, material heat, press period and gauge result, using within-shift drift to distinguish a component change from an assembly change.
  • Controller and sensor timestamps — Sensor time, controller receipt time and actuator response time are replayed together, separating dropped frames, bus congestion and control-logic delay.
  • Combined-environment test unit — An assembly with fixed hardware and software revisions enters vibration, damp heat, salt spray and voltage disturbance, retaining test order and every interruption with the specimen.

01 · Freeze the vehicle interface before the component

Envelope, voltage, cooling and communications are four different drawings.

The vehicle programme fixes package, power, cooling and communications before prototype design. The package drawing, harness table, cooling circuit and control protocol must cite one baseline. If the prototype receives a temporary connector or bracket change, that revision joins the same file set. Each component report then names its baseline so engineers can decide whether older data still applies to the current vehicle. The baseline also names each interface owner for direct clarification.

Moving a coolant port by a few millimetres changes hose stress, flow, and service access together, while adding one bus signal may require fresh controller calibration. Such vehicle-interface changes can resemble component faults in test results, and mixed timing or revisions send rework to the wrong object. Every interface change creates a new baseline, and old tests remain usable only after the unchanged boundaries are identified.

Give every interface change a new identity and state which bench and road tests must be repeated. The review signs mechanical, electrical, thermal and communication boundaries separately and prepares an impact list for every change. That list names the bench, environmental or road items to repeat rather than merely announcing a revision. Unaffected evidence remains valid, while affected results receive a clear expiry date. Every added test cites its impact-list identity and approval chain.

  • Interface freeze defines the vehicle numbers to which a result applies; it does not prohibit engineering change.

02 · Balance four flows on the fuel-cell bench

Hydrogen, air, coolant and current jointly determine power.

Stack output belongs on the same timeline as air compression, hydrogen delivery, cooling and controller demand. The system trace retains stack voltage and current together with hydrogen pressure, air flow, coolant inlet and outlet temperatures, and auxiliary demand. During a cold start or rapid power ramp, the order of these signals says more about air, hydrogen or cooling limits than any isolated peak. Auxiliary demand has its own column rather than disappearing under stack output.

Net power alone hides cold start, transients and auxiliary consumption in system efficiency. Two duties with the same net power may differ: one sustains output through high compressor demand, while another reaches a cooling limit. Vehicle-side output alone cannot separate them. If an auxiliary saturates first during a transient, the later power reduction should not be attributed immediately to stack material. The order of power limiting separates auxiliary and stack responsibility.

Replay the vehicle demand curve on the bench and record the direct trigger for each power limit. The bench uses vehicle power demand as its input and replays launch, acceleration, grade and idle recovery. Around every power limit, it locks several seconds of complete raw signals. After changing a hydrogen valve, compressor control or cooling parameter, the same demand trace is used so a different duty cannot masquerade as a repair. Data near a trigger remains unsmoothed so a transient peak survives export.

  • Peak stack output gains vehicle meaning only after air-path and cooling demand are deducted.

03 · Battery-pack safety begins with one cell identity

Thermal, electrical, structural and traceability conditions must all hold.

Cells, modules and packs retain a continuous relation among material lot, assembly position and test result. Cell incoming lot, module position and pack serial number form a continuous chain, with busbars, temperature sensors and insulators tied to their stations. Capacity, resistance, leak and insulation results are not stored only as averages; they remain locatable at pack edges, centre positions or near a coolant inlet. Any pack temperature spread can then return to cell position and assembly condition.

Average capacity cannot reveal local temperature, joint resistance or post-impact insulation risk. A local rise in joint resistance may first appear as a temperature spread, and a small post-impact displacement may affect insulation in only one region. Pack-average capacity is insensitive to both. If teardown loses the original position, an abnormal cell cannot be matched to welding, adhesive or cooling-plate assembly. Without position evidence, teardown leaves material speculation rather than a station.

Sample different pack positions and return teardown findings to the responsible cell and operation. The sampling matrix covers different module positions and keeps the same points after thermal cycling, vibration or impact. Before teardown, photographs and coordinates are frozen; cells, joints and operation numbers are then written back one by one. Corrective work targets the confirmed lot and station and is verified on a pack not used for teardown. The verification pack retains cell layout for point-by-point thermal comparison.

  • A thermal event cannot be located if cell lot, module position or BMS revision is missing.

04 · How one tool holds vehicle body gaps

Dimensional stability comes from tooling, material and press cadence.

First, middle and final parts from one forming lot are compared on a common datum. First, middle and final stamped or moulded parts are measured on one gauge datum, while appearance defects are recorded separately for visible and mounting surfaces. Coil or resin lot, tool cavity, equipment temperature and production time stay with the reading so dimensional drift returns to a particular manufacturing condition. Appearance checks use fixed light and angle so crews do not shift defect criteria.

Springback, tool heating and clamping shifts can move assembly gaps away from the first piece. A passing first piece does not establish lot stability. Tool heating changes clearances, springback shifts with material, and worn clamping faces can move holes gradually. A flushness problem on the vehicle may combine several parts, so selecting one worst item does not describe the complete tolerance stack. Tolerance direction and combination matter; one worst part may not create the worst fit.

Merge gauge data, tool maintenance and vehicle-fit feedback by lot and trace drift to a station. Gauge readings are plotted in production order and aligned with tool maintenance, material changes and vehicle-fit feedback. When one-way drift appears, inspection intervals narrow before cavity, temperature or material is isolated. After adjustment, a new first piece starts the series; a repaired final part never overwrites the original trend. The trend also shows tool cavity, revealing drift confined to one cavity.

  • A body-gap trend needs both first-to-last parts and assembly points; one conforming part cannot explain vehicle deviation.

05 · Sensors and controllers need one clock

Latency and dropped frames alter intelligent-function behaviour.

Sensor raw values, controller inputs and actuator commands require a comparable time base. Camera, radar or other sensor timestamps must be convertible to the controller receive time, algorithm output and actuator command clock. Bus load, firmware revision and dropped-frame count join the event package so offline replay does not confuse latency with a perception error. Each event package retains the clock-calibration method for one shared zero.

A module can pass self-test yet misbehave under bus congestion or synchronisation error. A module self-test rarely reveals timing faults caused by bus congestion, clock drift or several sensors competing for resources. A vehicle action delayed by tens of milliseconds may arise from message arrival or a calibration threshold, and those causes require different corrective ownership. Timing evidence separates algorithm decisions from message transport.

Retain raw data for offline replay of critical scenes and compare action timing after calibration updates. Critical scenes retain raw data without resampling and reconstruct input, decision and actuation on one clock. After calibration or firmware revision, the same event package is replayed frame by frame to compare triggers, dropped frames and actuation delay. Road retesting begins only after the timing difference is resolved. Road retesting keeps the same clock scheme across both stages.

  • A millisecond-scale clock error can reverse the apparent causal order of a driver-assistance event.

06 · Environmental durability is more than one temperature

Vibration, damp heat, salt and voltage variation interact.

Components may fail under combined temperature, humidity, vibration and supply variation despite passing each test alone. Combined environmental testing sequences temperature cycling, damp heat, vibration and supply variation around the component’s vehicle location instead of adding four separate certificates. The same function, connection resistance, sealing and appearance checks are made before and after exposure, and any intermediate failure is tied to its stage. The profile states exposure order because wet-then-vibration differs from the reverse.

Excess acceleration creates unrealistic damage, while weak exposure misses cumulative road fatigue. An excessive combined profile may create resonance or condensation absent from service, while a weak profile misses accumulated fatigue; duration alone does not prove acceleration. Mounting zones face different gradients, salt and vibration directions, so one profile cannot cover every vehicle component without qualification. An acceleration factor holds only when failure morphology remains the same.

Define the combined profile from vehicle location and measure the same functions and connections before and after. The team first measures the in-vehicle envelope, then selects acceleration that preserves the main failure mechanisms. Functional checkpoints separate each exposure stage and direct deterioration back to connections, sealing or material. A revised item repeats the sequence and is compared with a road sample to confirm the accelerated test did not change failure type. A road control calibrates reality and keeps laboratory conditions tied to service.

  • Exposure order changes damage accumulation, so individual passes cannot simply be added together.

07 · Vehicle validation takes bench assumptions onto the road

Road, driver and traffic make loads irregular.

Road testing marks grades, congestion, acceleration, charging or refuelling events and aligns them to component signals. The road-event package places grade, speed, traffic state, driver demand and charging or refuelling points on the same timeline as component signals. Vehicle load, tyre condition and ambient temperature are also retained because they change power demand, vibration and heat rejection. These environmental measures separate vehicle change from weather between road runs.

A road exception that cannot return to the bench cannot be classified as random disturbance or stable fault. An isolated alert that cannot repeat on the same road section or bench duty is hard to classify as random communication, driver action or a stable hardware fault. If a replacement part is followed by a different route, the absence of another alert does not prove the new part corrected the original problem. No alert on a different route records an observation but does not confirm repair.

Engineers extract the shortest repeatable segment from a long road test, fix vehicle load, driver action, and software revision, and translate the same demand into a bench duty. The bench runs before and after replacement and then returns to the original road, with weather and traffic checked when reproduction rate changes. If the bench repeats reliably while the road does not, route conditions take priority over another blind part replacement.

  • A road replay tests the bench hypothesis only when it returns with the same vehicle and calibration identity.

08 · Lot release signs the changes too

Release must state how this lot differs from the previous one.

The production-release package lists part lot, hardware revision, software calibration, inspection result, approved deviations, and every change from the prior lot. Material substitution, process setting, sub-supplier, and firmware adjustment cannot remain only with the supplier because each may alter completed vehicle validation. A temporary deviation also states applicable quantity, expiry, and closure owner so an emergency disposition does not become permanent practice.

A supplier change that misses vehicle impact review removes the starting point for later faults. Two lots may both show pass while using different inspection methods or an unevaluated material change, leaving no starting point for a later fault. Repeating every test for every change is also inefficient. The useful boundary follows affected interfaces and failure mechanisms. Impact assessment concentrates validation on interfaces that alter failure risk.

Maintain a list of material, process, software and sub-supplier changes requiring revalidation, with approver and date. Component, controls, test and vehicle roles classify each change as still valid, requiring added verification, or invalidating prior evidence. Release signatures carry date and owner. If production later shows an exception, investigation starts from the corresponding change and affected test instead of reopening the whole supply chain. The change history gives later investigations a precise starting point.

  • Production release should state every difference from the prior lot and assign an expiry point to each deviation.

Sources and reading notes

Vehicle sales do not substitute for component validation, and a passing prototype is not production release. Hardware, software and vehicle revisions are followed separately and interpreted only when their timestamps align.

  • Wuhan Hannan intelligent connected NEV components: National recognition year is 2024.
  • Wuhan Hannan intelligent connected NEV components: List validity is 3 years.
  • SMEs reported in the cluster: more than 280.
  • Component fields in the public demand list: 5 fields.

Related companies

These companies have a locally relevant industry role supported by public information. Order is not a ranking, inclusion does not imply authorization, partnership or endorsement, and Hello China does not participate in or guarantee transactions.

A company may request a free correction or removal through the formal channel if its record is inaccurate or it does not wish to be shown. Do not send business licences, authorization letters or other sensitive files through insecure channels.

Sources

Use the original publishers to check dates, scope and definitions; public material is not a transaction guarantee.