Knowledge Base · Vehicle Sensor Networks

Ship Sensor Networks

What a merchant or ferry vessel measures, over which buses, at what rates, how it behaves across a voyage, what goes wrong, and how the Sense simulator should generate it.

Prepared 12 Oct 2026 · Feeds: Sense simulator data generator

1 Overview and operating context

A ship is a long-lived, mostly autonomous sensor network. Navigation sensors feed the bridge over NMEA 0183 or IEC 61162-450 Ethernet, engines and machinery report to an alarm and monitoring system (AMS) over Modbus, and smaller vessels and ferries use NMEA 2000 on CAN. A voyage data recorder (VDR) records bridge data continuously. Shore connectivity is satellite based, so the data path is bandwidth limited and intermittent.

Vessel classes in scope for the simulator are container ships (including reefer plugs), tankers (cargo tank, inert gas and gas detection signals), bulkers and ferries. Typical service speeds: container 14-22 kn, tanker 11-15 kn, bulker 10-14 kn, ferry 18-30 kn.

Most numeric values below are plausible industry defaults taken from the research notes, not specification quotes. Each table carries a basis line: "sourced" means the research notes tie it to a cited web source, "estimate" means domain-knowledge typical value to be treated as a tunable default.

Vessel identifiers: IMO number is 7 digits, MMSI is 9 digits with a MID prefix. A noon report (position, distance, fuel, remaining on board, weather) is sent daily at 12:00 local.

2 Sensor inventory

2.1 Navigation (NMEA 0183, NMEA 2000, AIS)

Navigation signals. Basis: PGN names sourced (Garmin list); ranges and rates estimate.
Signal	Unit	Typical range	Sample / report rate	Source bus (0183 sentence / N2K PGN)
Position lat/lon	deg (1e-7)	-90..90, -180..180; noise 1-3 m CEP; HDOP 0.7-1.5 with DGPS	1 Hz	GGA, GLL, RMC / 129025, 129029
COG / SOG	deg true; kn	0-359.9; 0-30 kn cargo, 0-45 kn fast ferry	1 Hz	VTG, RMC / 129026
Heading	deg	0-359.9; gyro drift under 0.1 deg/h; yaw oscillation +-0.5-2 deg in seaway	10 Hz gyro, 1 Hz log	HDT, HDG, THS / 127250
Rate of turn	deg/min	-720..720; typically +-30; hard turn 20-40 deg/min for a large ship	10 Hz	ROT / 127251
Water speed	kn	0-30; differs from SOG by current 0-3 kn	1 Hz	VHW / 128259
Depth	m	0.5-200+; blank in deep water; dropouts 1-3% in aeration	1 Hz	DBT, DPT, DBK / 128267
Wind	kn or m/s; deg	0-60 kn; 0-359 deg; true wind computed from apparent plus SOG/heading	1 Hz	MWV, MWD / 130306
Rudder angle	deg	-35..+35; autopilot trims +-1-3 deg	5-10 Hz	RSA / 127245
UTC time	UTC	-	1 Hz	ZDA / 126992
GGA quality fields	mixed	fix quality 0/1/2 (DGPS)/4 (RTK); 8-14 satellites	1 Hz	GGA

AIS Class A dynamic report intervals. Basis: sourced (ITU-R M.1371 as summarised by Raymarine).
Ship state	Nominal interval
At anchor or moored, up to 3 kn	3 min
At anchor or moored, over 3 kn	10 s
0-14 kn, steady course	10 s
0-14 kn, changing course	3 1/3 s
14-23 kn, steady	6 s
14-23 kn, changing	2 s
Over 23 kn	2 s
Static and voyage data (message 5)	6 min or on change
Class B SOTDMA	3 min (0-2 kn), 30 s (2-14), 15 s (14-23), 5 s (over 23)
Class B CSTDMA	3 min (2 kn or less), 30 s (over 2 kn)
Aid to navigation (message 21)	3 min

AIS field encodings (estimate): lat/lon 1/10000 min; SOG 0.1 kn (1023 = not available); COG 0.1 deg (3600 not available); heading 0-359 (511 not available); ROT -128..127 (-128 not available); navigation status 0 underway on engine, 1 at anchor, 5 moored, 8 sailing, 15 undefined; timestamp second 0-59 (60 not available). Message 5 carries name, IMO, callsign, type, dimensions, draught (0.1 m), ETA and a 20 character destination. Messages 1/2/3 are Class A position, 18/19 Class B, 24 Class B static, 27 long range (3 min via satellite). Terrestrial reception loses about 2-10% of sentences.

2.2 Engine and machinery

Main engine and generator signals. Basis: estimate (domain knowledge); AMS scans at 1 s and stores at 10-60 s.
Signal	Unit	2-stroke main engine, cruise	4-stroke genset / ferry	Sample / report rate	Source bus
Engine speed	rpm	70-105	500-1000 (gensets fixed 720/750/900/1000)	10 Hz N2K; AMS 1 s	N2K 127488, Modbus, J1939
Load	% MCR	60-85 service; 10-40 slow steaming	40-80	1 s scan, 10-60 s store	Modbus / AMS
Exhaust gas T per cylinder	degC	340-440	350-520	1 s scan	Modbus / AMS
Turbo inlet / outlet exhaust T	degC	400-500 / 230-300	500-600 / 350-450	1 s scan	Modbus / AMS
Scavenge air pressure	bar	1.0-3.0 (0.3-0.5 at 25% load)	2.0-4.5 boost	1 s scan	Modbus / AMS
Scavenge air T	degC	35-50	40-55	1 s scan	Modbus / AMS
Turbocharger speed	rpm	8000-20000 large; 20-50 krpm small	25-60 krpm	1 s scan	Modbus / AMS
Jacket cooling water out	degC	80-90 (in-out delta 6-12 K)	80-92	1 Hz N2K 127489	N2K, Modbus
Jacket water pressure	bar	2-4	2-4	1 s scan	Modbus / AMS
Lube oil pressure	bar	3.5-5.0	4.0-6.0 (genset 3.5-5)	1 Hz N2K (kPa)	N2K 127489, Modbus
Lube oil T in	degC	40-50	60-80	1 s scan	Modbus / AMS
Main bearing T	degC	55-75	70-95	1 s scan	Modbus / AMS
Fuel oil inlet T (HFO)	degC	120-140 (12-18 cSt)	110-130	1 s scan	Modbus / AMS
Fuel pressure	bar	7-10	6-10	1 s scan	Modbus / AMS
Fuel rate / SFOC	kg/h; g/kWh	about 1500-3000 kg/h at 70% for a 6S60; SFOC 165-185	SFOC 190-220	1 Hz N2K (L/h)	N2K 127489, Modbus
Daily fuel consumption	t/day	VLCC 45-70 at 14 kn; 8000 TEU 80-200 at 18-22 kn; slow steaming 100-120	ferry 30-100	derived daily	AMS / noon report
Shaft power / torque	kW / kNm	10-40 MW large ship	-	1 s scan	Modbus / AMS
Starting air / control air	bar	25-30 / 6-8	25-30 / 6-8	1 s scan	Modbus / AMS
Engine vibration	mm/s RMS	3-8	4-9	1 s scan	Modbus / AMS
Generator V / Hz / kW	V, Hz, kW	440/450 V, 60/50 Hz, 400-3000 kW	-	1 s scan	Modbus; J1939 on medium-speed gensets
Generator winding T	degC	70-110	-	1 s scan	Modbus / AMS
Stern tube bearing T	degC	40-65	-	1 s scan	Modbus / AMS
Crankcase oil mist	mg/l	about 0	about 0	1 s scan	Modbus / AMS
Battery (N2K)	V, A, K	24 V nominal, 22-29 V	-	1 s	N2K 127508

2.3 Tanks, bilge and ballast

Tank, bilge, ballast and cargo signals. Basis: estimate; N2K fluid level PGN 127505 name sourced.
Signal	Unit	Typical range	Sample / report rate	Source bus
Fuel tanks (HFO, MGO)	m3, %, cm	0-100%; bunker 2000-8000 m3 on a big ship	10-60 s	N2K 127505 / Modbus
Ballast tanks	m3, %	20-40 tanks; ballast volume 20-60% of DWT	10 s	Modbus / AMS
Ballast pump flow	m3/h	1000-3000	10 s	Modbus
Draft fwd / aft / mid	m	6-16 (container 6-14, VLCC 10-22); trim -2..+3 m	1 min	Modbus / AMS
Bilge level	mm	0 normal; alarm about 150-300; high-high 500; auto-start cycles 1-10/day	1 s	Modbus / AMS
Oily water separator	ppm	0-15	10 s	Modbus / AMS
Cargo tank (tanker) level, T, P	m, degC, kPa	pressure 0-20 kPa (inert gas 5-15); cargo T 35-60; IG O2 below 5%	10 s	Modbus / AMS
Gas detection	% LEL	0	1 s	Modbus / AMS
LNG boil-off (dual fuel)	%/day	0.08-0.15	-	Modbus / AMS
Stability GM	m	0.5-5 (container 1.0-3.5); IMO intact minimum 0.15	event	loading computer

2.4 Hull motion and weather

Hull motion and environment. Basis: estimate; N2K attitude and environment PGN numbers from the research notes.
Signal	Unit	Calm / typical	Heavy weather	Sample / report rate	Source bus
Roll	deg	+-1-3; natural period 12-25 s container, 8-14 s other	+-10-30	10 Hz	N2K 127257 / 0183 XDR
Pitch	deg	+-0.5-1.5; period 6-10 s	+-3-8	10 Hz	N2K 127257
Heave	m	+-0.2-1	+-3-6	IMU 25-100 Hz, publish 1-10 Hz	IMU
Bridge acceleration	g	0.02-0.05	0.15-0.3	IMU 25-100 Hz, publish 1-10 Hz	IMU
Hull stress	microstrain	+-100-300 (+-20-60 MPa)	300-1000	10-40 Hz; statistics every 20-30 min	Hull monitoring system
Slamming / whipping	events/h	under 1	5-30	event	Hull monitoring system
Significant wave height Hs	m	0.5-2.5	4-12	-	Wave radar
Wave period Tp	s	5-9	10-16	-	Wave radar
True wind	kn	5-20	35-60	1 Hz	N2K 130306
Air pressure	hPa	940-1050 range	drop over 3 hPa in 3 h signals storm	1-2 s	N2K 130310 / 130311
Sea water T; current	degC; kn	0-32; 0-3	-	1-2 s	N2K 130310

2.5 Alarm system and bridge items

Alarm, ECDIS and bridge data. Basis: estimate.
Signal	Unit	Typical range	Sample / report rate	Source bus
Cross-track distance (XTD)	m	limit 0.05-0.5 nm	1 Hz	ECDIS
Safety contour	m	10-30	1 Hz	ECDIS
CPA / TCPA limits	nm / min	CPA 0.5-2 nm; TCPA 12-20 min	1 Hz	ECDIS / radar
Radar tracked targets	count	40-100 at 24-48 rpm	per scan	Radar
BNWAS dormant period	min	3-12, then visual 15 s, audible 15 s, remote	event	Bridge
AMS alarm and event stream	class, tag, state	see section 6	event	Modbus / AMS
Fire detection, watertight doors	binary	-	event	AMS / VDR

2.6 Reefer containers

Reefer container monitoring. Basis: estimate.
Signal	Unit	Typical range	Sample / report rate	Source bus
Set point	degC	chilled +2..+13; frozen -18..-25; ultra-low -60	on change	Reefer controller
Supply air T	degC	set point +-0.5	5-15 min logger	Reefer controller
Return air T	degC	supply + 1..5	5-15 min logger	Reefer controller
Ambient T	degC	-10..+45	5-15 min	Reefer controller
Humidity	%	60-95	5-15 min	Reefer controller
O2 / CO2 (controlled atmosphere)	%	O2 2-5; CO2 2-8	5-15 min	Reefer controller
Compressor current	A	3-18 (3-phase 380/440 V)	5-15 min	Reefer controller
Power draw	kW	3-5 running; 6-9 defrost	5-15 min	Reefer controller
Defrost cycles	per day	2-4 for frozen, 20-40 min each; supply air spikes +5-15 K	event	Reefer controller
Mode	enum	ON, DEFROST, PULSATION, STANDBY	event	Reefer controller

Reefer data reaches the ship through a gateway (remote container monitoring over PLC or powerline), with ship gateway reports from 1 min to 60 min, and 1 hour to 15 min reporting for remote monitoring. A ship may have 500-2000 reefer plugs (estimate).

3 Shipboard networks and protocols

Network layers. Basis: mix of sourced (bridge notes) and estimate.
Layer	Technology	Rate / cadence	Notes
Nav sensors to bridge	NMEA 0183 (IEC 61162-1), RS-422	4800 baud (38400 for AIS and high speed); about 480 B/s at 4800; 1 Hz typical, 10 Hz heading and ROT	ASCII, XOR checksum, sentence at most 82 chars
Nav, integrated bridge	IEC 61162-450 Ethernet	UDP multicast, same sentences	Newbuild integrated bridges
Auxiliary and small vessel	NMEA 2000 (IEC 61162-3)	CAN 250 kbit/s, 29-bit ID, 8 B frames	Fast packet carries up to 223 B over 32 frames; one missing frame discards the whole PGN
Engine and AMS	Modbus RTU / TCP, OPC UA gateway	poll 0.5-5 s per register block; AMS scan 1 s; trend log 10-60 s	RTU 9600-115200 baud; 3-10% retries on noisy power
Engine controller	CAN J1939 on medium-speed gensets	10-1000 ms per PGN	Sentinel values 0xFF/0xFFFF mean not available
VDR	IEC 61996	continuous; 12 h (VDR) or 48 h (S-VDR)	Radar, ECDIS, audio, nav data, alarms, hull openings, doors
Shore link	VSAT / FBB / Iridium / LEO	see below	Store and forward when down

NMEA 2000 PGN rates. Basis: estimate (bridge notes); PGN names sourced.
PGN	Name	Rate	Framing
129025	Position, rapid update	10 Hz	single frame
129026	COG and SOG, rapid update	4 Hz	single frame
127488	Engine parameters, rapid	100 ms	single frame
127489	Engine parameters, dynamic	500 ms	fast packet
129029	GNSS position data	1 Hz, 43 B	fast packet
129038	AIS Class A position	event	fast packet

Satellite and shore links. Basis: sourced where named in the research (Iridium blog, AWI, boatnews, UNOLS); other rows estimate.
Link	Down / up	Round trip	Outage and behaviour
GEO Ku/Ka VSAT	typically 2-20 Mbps per vessel; research ship about 3 down / 1-1.5 up, peaks 10 / 2	about 500-800 ms	Availability 95-99.7%; mast or funnel blockage dropouts 1-30 min; rain fade 5-30 min with throughput 10-30%; beam handover 10-60 s; GEO gap above about 70-75 N/S
LEO (Starlink maritime)	vendor 100-350 Mbps	25-60 ms	Availability 97-99.5%; obstruction drops 1-15 s; polar gaps; re-acquisition 1-5 min (estimate)
Iridium Certus	about 700 kbps observed (704/352 kbps nominal for Certus 700)	500-600 ms typical	99%+ low rate; works at the poles
Inmarsat FleetBroadband	150-432 kbps	700-900 ms	Estimate only
Coastal cellular and port Wi-Fi	4G 5-50 Mbps up	30-60 ms	Cellular to about 20-40 km offshore, switchover to satellite 5-60 s; port Wi-Fi used for backfill

Shore reporting pattern (estimate): AIS-based position every 6 min, noon report daily, engine trend batched every 1-15 min via MQTT or file. Planning volume for a ship is 1 publish per 5-60 s over satellite, 2-30 kB per batch, 50 MB-2 GB raw per day, 5-100 MB reduced, under 5 MB on Iridium (estimate).

4 Voyage profiles

Voyage state machine. Basis: estimate.
State	Duration	SOG	Main engine load	Notes
Sea passage (open sea transit)	1-25 days	container 14-22 kn; tanker 11-15; bulker 10-14; ferry 18-30	60-85%; slow steaming 25-50% at 10-14 kn	Heading steady +-1-2 deg; ROT small; AIS 6-10 s steady; two-stroke at 70-105 rpm
Manoeuvring (unmooring, pilotage, departure, arrival)	1-3 h	0-10 kn	Bell orders: engine rpm steps to 20-60 rpm (Dead slow, Slow, Half, Full) for 10-30 min	AIS 3 1/3 s while turning; rudder activity high
River or channel	1-6 h	6-12 kn	30-50%	Between berth and open sea
Port stay (at berth)	12-48 h (container 18-36 h); ferry turnaround 30-90 min	0	Main engine off; 2-3 gensets; cargo gear and reefer power load gensets 20-40%	AIS 3 min, status 5 (moored); reefer plugs powered
Anchorage wait	hours to days	0; drift circle 50-200 m	Main engine off	AIS status 1 (at anchor), 3 min interval
Ferry operations	2-8 crossings per day	15-25 kn	70-85%, 2 of 4 engines	Turnaround between crossings; N2K bus common
Weather deviation	hours	reduced 2-8 kn	Voluntary speed reduction	Added resistance 10-40% in heavy weather

Typical sequence for a cargo ship: berth, unmooring and departure, river or channel, open sea (with possible slow steaming or weather deviation), channel, arrival manoeuvring, berth or anchorage wait. Inference: a state machine with these states and the dwell ranges above is sufficient to drive all signals; the notes do not give transition probabilities.

5 Faults and anomaly signatures

Anomaly scenarios and injection frequency. Basis: estimate; frequencies are not from a fetched source.
Scenario	Signature	Frequency
Single-cylinder exhaust T drift	+30..+80 K over hours to days (injector fouling or valve leakage)	0.5-1 events per engine-month
Scavenge fire	Exhaust T spike +100-200 K; scavenge air T over 100-150 degC	about 0.01 per engine-month
Turbocharger fouling	Boost down 5-10%; exhaust T rising over weeks	gradual
Lube oil filter clogging	Lube oil pressure falls 0.5 bar over days	low alarm about 0.05 per month
High ambient engine room	Jacket water +5 K at 35-45 degC engine room	seasonal
Sensor failures	Stuck value, flatline, spike to full scale, wire break (4-20 mA below 3.8 mA)	0.5-2% of sensors per month
GNSS jamming or spoofing	Position jumps 10-100+ km, SOG over 50 kn, HDOP up, circle spoofing	0.1-1% of time in hot regions
AIS gap	No message for over 15 min at sea	1-3% of voyages
Gyro or compass failure	Heading frozen or offset 5-30 deg	rare
Blackout	Generator frequency collapses; electrical load to 0 for 1-5 min	about 1 per year across the fleet
Bilge flooding	Bilge level rising with pumps stuck on	about 0.1 per year
Reefer failure	Supply air T rising about 0.5 K/min; temperature excursion	about 1-3% of containers per voyage have at least one alarm; 0.5-1% of reefers per day alarm on power loss
Heavy weather	Roll and hull stress up, speed down 2-8 kn	5-10% of time on North Atlantic and Pacific winter routes
Abnormal fuel consumption or leak	+10-20% kg per nm against baseline	occasional
Hull fouling	+0.5-1% fuel per nm per month	gradual
Man overboard, fire	Event	very rare

Bus, gateway and link failures that affect ship data. Basis: estimate (bridge notes).
Failure	Signature	Rate
CAN bus-off	All signals vanish 0.1-5 s or until power cycle	0.05-0.5 episodes per vehicle per day
Fast-packet corruption	PGN such as 129029 or 127489 missing or stale	0.5-2%
Modbus retries	Delayed or repeated polls	3-10% on noisy power
Not-available sentinels and scaling errors	Impossible values when sensor absent	0.1-1% of records
Frozen value	Flatline with valid timestamp	0.1-1% of signals
NMEA 0183 gateway overload	Dropped sentences when full conversion at 4800 baud	config dependent
Link outage	Satellite drop; data buffered then replayed out of order	section 3 outage patterns
Store-and-forward overflow	Records dropped after 6-72 h offline at 1 Hz; sequence gap	0.5-2% of outages over 6 h

6 Alarms, events and thresholds

Alarm classes. Basis: estimate.
Class	Meaning	Examples	Typical share of alarm records
Emergency alarm	General alarm, fire, CO2 release, flooding	Fire detector zone, fire pump start fail	under 0.1%
Critical alarm / shutdown	Safety trips	Lube oil low-low, overspeed, oil mist	about 0.5%
Warning / alarm	Needs action	High exhaust T, low pressure	5-10%
Caution / advisory	Information	Filter differential pressure, sensor deviation	about 20%
Event	State change	Pump start or stop	majority
Sensor fault	Instrument failure	Wire break	about 3%

Alarm thresholds. Basis: estimate.
Signal	Alarm	Shutdown / slowdown
Engine speed	Above 105-108% MCR	Overspeed trip at 110-115%
Exhaust T per cylinder	High 470-480 degC; deviation from mean beyond +-40-50 K	-
Turbo inlet exhaust T	520 degC	-
Scavenge air T	High 60 degC	-
Jacket cooling water	High 90-95 degC; low 60 degC	Slowdown 95-97 degC
Jacket water pressure	Low 1.0-1.5 bar	-
Lube oil pressure	Low 2.5 bar	Low-low 1.0-2.0 bar
Lube oil T	High 65 / 90 degC (two-stroke / four-stroke)	-
Oil mist	2.5 mg/l	Shutdown on high
Main bearing T	High 80-95 degC	-
Fuel pressure	Low 5 bar	-
Starting air / control air	Low 15-18 bar / low 5 bar	-
Engine vibration	High 11 mm/s (ISO 10816)	-
Generator V / Hz	+-10% / +-5%	Blackout detect
Generator winding T	130-140 degC	-
Stern tube bearing T	70 degC	80 degC
Fuel tank level	Low 10-15%; high-high 95%	-
Ballast tank level	High 90-98%	-
Bilge level	150-300 mm; high-high 500 mm	-
Oily water separator	Above 15 ppm	Stops overboard discharge
Cargo tank level / IG oxygen	High level 95%; O2 above 8%	Overfill 98%
Gas detection	10-20% LEL	-
Reefer supply air	Deviation over 2-3 K for over 30-60 min	-
Battery (24 V)	Outside 22-29 V	-

Bridge alert management (IEC 62923) priorities are emergency alarm, alarm, warning, caution. Alarms need acknowledgement; a warning escalates to an alarm if unacknowledged for about 2 min (set by the ship). With an unmanned machinery space, an engineer call and extension alarm raise a watch call if there is no acknowledgement in 2-5 min. A healthy cargo ship produces 20-150 AMS alarms per day, about half nuisance, and about 10 events per hour in port (estimate).

7 Generator guidance for the simulator

Generator block A: truth signals at native rates

Drive every signal from the voyage state (section 4). Sample at native rates, then pass through the bus and link stages. Suggested parameters, all inference from the ranges above:

Position: integrate SOG and COG; add Gaussian noise with sigma about 1.5 m; HDOP uniform 0.7-1.5; satellites 8-14.
Heading: base course plus AR(1) yaw term with amplitude +-0.5 deg calm to +-2 deg in seaway; gyro drift 0.1 deg/h.
Water speed: SOG minus current component, where current is uniform 0-3 kn.
Depth: report only below about 100-300 m; drop out 1-3% of samples.
Roll: sinusoid with period uniform 12-25 s (container) amplitude 1-3 deg calm scaling to 10-30 deg heavy weather; pitch period 6-10 s amplitude 0.5-1.5 deg scaling to 3-8 deg.
Apparent wind: compute from true wind plus the ship velocity vector.

Generator block B: engine as a function of load

Let L be the load fraction of MCR. Inference-based interpolation within the typical ranges:

rpm            = 70 + (105 - 70) * clamp((L - 0.1) / 0.75, 0, 1)
exhaust_T_cyl  = 340 + (440 - 340) * clamp(L / 0.85, 0, 1) + N(0, 4)
scav_P_bar     = 0.4 + (3.0 - 0.4) * clamp(L, 0, 1)
fuel_kg_h      = SFOC_g_per_kWh * P_kW / 1000        # SFOC in 165-185 for 2-stroke
P_kW           = L * MCR_kW
jacket_out_C   = 85 + N(0, 1.5) + 5 * hot_engine_room
lube_P_bar     = 4.2 + N(0, 0.15)

Hold the per-cylinder exhaust spread within +-15 K of the mean in healthy state so that the alarm deviation of +-40-50 K only appears on injected drift faults. Alarm limits come from section 6.

Generator block C: state machine and dwell times

Sample dwell uniformly within the section 4 ranges; sea passage days sampled from 1-25.
Port stay 18-36 h for container, 12-48 h otherwise, ferry 30-90 min per turnaround.
Manoeuvring bell sequence: rpm steps to 20-60 rpm for 10-30 min total, SOG 0-10 kn.
AIS cadence follows the Class A table keyed on SOG and whether the course is changing; anchored or moored at 3 kn or less uses 3 min.
Heavy weather as a modifier: 5-10% of time for winter ocean routes; speed loss 2-8 kn; roll and hull stress raised per section 2.4.

Generator block D: per-tenant variant knobs

Variant knobs by tenant type. Basis: inference built on the research ranges.
Knob	Container line	Tanker	Ferry
Service speed	14-22 kn	11-15 kn	18-30 kn
Main engine	2-stroke, rpm 70-105	2-stroke, rpm 70-105	4-stroke medium speed, 500-1000 rpm, 2 of 4 engines
Port stay	18-36 h	12-48 h	30-90 min turnaround, 2-8 crossings per day
Extra signal groups	Reefer block (hundreds to 2000 plugs), GM 1.0-3.5, roll period 12-25 s	Cargo tank level, T, pressure, inert gas O2, gas detection, LEL	Passenger-vessel items not in the research (inference: none generated)
Primary bus	AMS Modbus plus 0183 / 61162-450	AMS Modbus plus 0183	NMEA 2000 plus Modbus
Daily fuel	80-200 t/day (8000 TEU)	45-70 t/day (VLCC at 14 kn)	30-100 t/day
Draft range	6-14 m	10-22 m (VLCC)	not in research
Link	VSAT or Starlink; port Wi-Fi backfill	VSAT or Iridium on low-rate plans	Cellular near coast, short crossings
Satellite report interval	10-60 s batches, 1 s on port Wi-Fi	30-300 s on low-bandwidth plans	10 s

Generator block E: imperfection pipeline

Truth signals at native bus rates.
Bus layer: frame loss, fast-packet corruption 0.5-2%, sentinels, bus-off windows 0.1-5 s.
Gateway: finite buffer, deadband (inference: speed 0.1-1 kn, heading 1-2 deg, temperature 1 degC), batch windows of 10-300 s.
Link state machine per link (up, degraded, down) using the outage ranges in section 3, with Markov dwell times; store and forward with out-of-order replay.
Cloud record with received_at, sequence number, boot id and tenant; label each injected imperfection as ground truth.

Inject fault classes at the section 5 frequencies; sensor failures at 0.5-2% of sensors per month; AIS gaps in 1-3% of voyages; GNSS spoofing windows 0.1-1% of time only when the route is in a flagged region.

8 Uncertainty and assumptions

Treat as tunable defaults

Items sourced from search summaries are not full-page reads. Most numeric ranges for engines, tanks, reefers, hull motion and all fault frequencies are domain-knowledge estimates, not specification quotes.
No source confirmed Inmarsat throughput or per-ship fault frequencies.
Satellite figures vary by source (VSAT RTT 500-800 ms, throughput depends on elevation, position and weather).
The simulator formulas in section 7 are inference: interpolation between quoted ranges, not published engine curves.
No ferry passenger-specific signals, ship-wide electrical load models or transition probabilities for the voyage state machine are present in the research.
AIS reception loss and dark periods (deliberate gaps below 0.5% of voyages) come from the notes as estimates.

9 Sources

https://docs.raymarine.com/87326/en-US/latest/Reference_ais500_reporting_interva-01E02213.html (AIS reporting intervals)
https://itu.int/dms_pubrec/itu-r/rec/m/R-REC-M.1371-2-200603-S!!MSW-E.doc (ITU-R M.1371)
https://www8.garmin.com/manuals/webhelp/echomapchirp/EN-US/GUID-46195FE3-A349-4BAD-A3DB-672B7E7D894E.html (NMEA 2000 PGN list)
https://www.isobus.net/isobus/pGNAndSPN/1856?type=PGN (J1939 PGN and SPN)
https://saemobilus.sae.org/content/J1939DA_201505 (SAE J1939DA)
https://www.iridium.com/blog/maritime-monday-low-latency (satellite latency)
https://spaces.awi.de/x/7zofGw (research ship satellite throughput)
https://newspaceeconomy.ca/2026/02/24/maritime-satellite-services-market-analysis-2026/ (maritime satellite market)
https://arxiv.org/pdf/1903.11814 (satellite latency and throughput)
https://spaces.awi.de/x/5VhaFQ (GEO VSAT throughput)
https://www.boatnews.com/news/52449/satellite-internet-for-boats-what-networks-are-behind-starlink-oneweb-and-inmarsat (VSAT and Starlink maritime)
https://www.unols.org/sites/default/files/202211rvt_apIX.pdf (Starlink polar test)
https://satellitephonestore.com/uploads/documentation/5192/Iridium-Certus-100-Fact-Sheet.pdf (Iridium Certus)
https://www.picotech.com/library/knowledge-bases/oscilloscopes/how-to-use-marine-network-decoding-nmea-2000-with-picoscope-oscilloscopes (NMEA 2000 fast packet)
https://pkg.go.dev/github.com/open-ships/n2k@v0.3.0/internal/framer (NMEA 2000 framer)
https://actisense.com/knowledge-base/hottest-articles/choosing-the-right-ngw-1/?seq_no=2 (NMEA 0183 gateway)
https://panbo.com/amec-nk-80-0183n2k-adaptor-and-milltechs-7th-year-of-ais/ (NMEA 0183 rates)