Process Schema

TEP_SCHEMA turns legacy arrays into named process quantities.

from tep_studio import TEP_SCHEMA

This page documents the in-code schema (the runtime name↔index map). It is distinct from the standalone JSON Schema deliverable under research/schema/ (research/schema/process_description.schema.json + research/schema/examples/cstr_minimal.json), which is the machine-readable process-description format described by the design principles.

Schema counts

Group	Count	Use
States	50	Dynamic state vector.
Standard measurements	41	Main observation vector.
Manipulated variables	12	Direct action vector.
Disturbances	28	Disturbance activation vector.
Additional measurements	32	Modified-TEP additional measurement array.
Disturbance monitors	21	Noise-free disturbance monitor outputs.
Process monitors	62	Internal process monitor outputs, including operating-cost entries.
Concentration monitors	96	Component concentration monitor outputs.

Variable metadata

Each variable stores:

Field	Meaning
name	Python-friendly variable name.
unit	Reported unit.
role	Variable group.
index	Zero-based Python index.
description	Process description.
lower, upper, nominal	Optional bound and nominal metadata.
available_online	Whether the variable is available online according to the schema.
legacy_symbol, legacy_index	Original legacy identifier when available.
stream, component, physical_type	Process context when available.
measurement_noise, sample_period	Measurement annotations when available.

Manipulated variables

Index	Name	Description
0	d_feed_valve	Valve position for feed component D, stream 2.
1	e_feed_valve	Valve position for feed component E, stream 3.
2	a_feed_valve	Valve position for feed component A, stream 1.
3	ac_feed_valve	Valve position for feed components A and C, stream 4.
4	compressor_recycle_valve	Compressor recycle valve position.
5	purge_valve	Purge valve position, stream 9.
6	separator_underflow_valve	Separator underflow valve position, stream 10.
7	stripper_underflow_valve	Stripper underflow valve position, stream 11.
8	stripper_steam_valve	Stripper steam valve position.
9	reactor_cooling_water_valve	Reactor cooling-water outlet valve position.
10	separator_cooling_water_valve	Separator cooling-water outlet valve position.
11	reactor_agitator_speed	Reactor agitator speed.

Standard measurements

Index	Name	Unit
0	feed_A_flow	kscmh
1	feed_D_flow	kg/h
2	feed_E_flow	kg/h
3	feed_AC_flow	kscmh
4	recycle_flow	kscmh
5	reactor_feed_flow	kscmh
6	reactor_pressure	kPa gauge
7	reactor_level	%
8	reactor_temperature	degC
9	purge_flow	kscmh
10	separator_temperature	degC
11	separator_level	%
12	separator_pressure	kPa gauge
13	separator_underflow	m3/h
14	stripper_level	%
15	stripper_pressure	kPa gauge
16	stripper_underflow	m3/h
17	stripper_temperature	degC
18	stripper_steam_flow	kg/h
19	compressor_work	kW
20	reactor_cooling_water_outlet_temperature_meas	degC
21	condenser_cooling_water_outlet_temperature	degC
22-27	reactor_feed_*_concentration	mol %
28-35	purge_*_concentration	mol %
36-40	stripper_underflow_*_concentration	mol %

Disturbances

The 28 legacy IDV entries are binary, latched activations. The public scenario schema accepts only 0 (off) or 1 (on). It does not expose intermediate values as a continuous severity scale. Disturbances whose physical root cause is not specified by the legacy model retain that status explicitly.

Index	Name	Description
0	idv_01	A/C ratio of stream 4, B composition constant.
1	idv_02	B composition of stream 4, A/C ratio constant.
2	idv_03	D feed temperature.
3	idv_04	Reactor cooling-water inlet temperature.
4	idv_05	Separator cooling-water inlet temperature.
5	idv_06	A feed loss.
6	idv_07	C header pressure loss.
7	idv_08	Random A/B/C composition of stream 4.
8	idv_09	Random D feed temperature.
9	idv_10	Random C feed temperature.
10	idv_11	Random reactor cooling-water inlet temperature.
11	idv_12	Random separator cooling-water inlet temperature.
12	idv_13	Reaction kinetics drift.
13	idv_14	Reactor cooling-water valve stiction.
14	idv_15	Separator cooling-water valve stiction.
15	idv_16	Random stripper heat-transfer deviation.
16	idv_17	Random reactor heat-transfer deviation.
17	idv_18	Random condenser heat-transfer deviation.
18	idv_19	Valve stiction group.
19	idv_20	Unknown random disturbance.
20	idv_21	Random A feed temperature.
21	idv_22	Random E feed temperature.
22	idv_23	Random A feed pressure/flow.
23	idv_24	Random D feed pressure/flow.
24	idv_25	Random E feed pressure/flow.
25	idv_26	Random A and C feed pressure/flow.
26	idv_27	Random reactor cooling-water pressure/flow.
27	idv_28	Random condenser cooling-water pressure/flow.

Published constraints

TEP_SCHEMA.constraints names the eight terminal limits used by the simulator. They cover high reactor pressure, high and low reactor level, high reactor temperature, high and low separator level, and high and low stripper level. TEP_SCHEMA.constraint_margins(measurements) returns a positive value inside each limit, zero at the limit, and a negative value after violation.

Print schema names

Use this when you need exact column names:

for variable in TEP_SCHEMA.measurements:
    print(variable.index, variable.name, variable.unit)

Named vector helpers

Use the schema helpers to avoid hard-coded array positions:

reactor_pressure_index = TEP_SCHEMA.index("measurements", "reactor_pressure")
reactor_pressure = result.measurements[reactor_pressure_index]

measurements = TEP_SCHEMA.to_dict("measurements", result.measurements)
print(measurements["reactor_pressure"])

Build a direct manipulated-variable vector by name:

action = TEP_SCHEMA.vector(
    "mvs",
    {
        "d_feed_valve": 63.053,
        "e_feed_valve": 53.98,
        "a_feed_valve": 24.644,
        "ac_feed_valve": 61.302,
        "compressor_recycle_valve": 22.21,
        "purge_valve": 40.064,
        "separator_underflow_valve": 38.10,
        "stripper_underflow_valve": 46.534,
        "stripper_steam_valve": 47.446,
        "reactor_cooling_water_valve": 41.106,
        "separator_cooling_water_valve": 18.114,
        "reactor_agitator_speed": 50.0,
    },
)

Update selected entries while preserving the rest of a vector:

next_action = TEP_SCHEMA.update_vector(
    "mvs",
    action,
    {"reactor_cooling_water_valve": 38.0},
)

The helpers only translate names and array positions. They do not implement feedback control or change the simulator clipping behavior.