Glossary

Signal level of the controller at 24 V DC. It is fed from its own power supply and is galvanically isolated from the 5 V signal logic of the signal bridge.

Occurs in Chapter 6 · System architecture; Chapter 7 · The test rig

See also Galvanic isolation· Converter stage

Closed-form mathematical solution of an equation. It exists for PT1 behaviour, which is why it serves in scenario 2 as the reference for the numerical computation of the model.

Occurs in Chapter 8 · The heated-tank model; Chapter 10 · Five test scenarios

See also PT1 element· Euler method (explicit)

Equation with which the model calculates the saturation pressure from the temperature. If the saturation pressure exceeds the pressure in the tank, the water boils (Fig. 10).

Occurs in Chapter 8 · The heated-tank model

See also Saturation pressure

Middle one of the three software layers: orchestrator.py on port 8050. It clocks the simulation, loads models at runtime (hot swap) and distributes the results.

Occurs in Chapter 6 · System architecture; Chapter 9 · The simulation cycle

See also Orchestrator· Presentation layer· Communication layer

Microcontroller board that works as the signal bridge in the test rig. It reads and sets the 5 V signals of the converter level and transmits the state serially at 115200 baud to the gateway.

Occurs in Chapter 6 · System architecture; Chapter 7 · The test rig

See also Signal bridge

Initial state before the thesis: solely the exam wall with simulation boards without internal states. There was no software at the beginning.

Occurs in Chapter 4 · Requirements

See also IHK simulation board

Central message intermediary through which all components communicate in MQTT. It would have had to be installed, configured and operated, and was not chosen in the concept decision.

Occurs in Chapter 5 · Solution approaches and concept decision

See also MQTT· Topic

Drawing interface of the browser. The presentation layer uses it for the animation of the plant.

Occurs in Chapter 6 · System architecture

See also Presentation layer

Documents the electrical design of the test rig (Fig. 13). It is drawn in detached representation and contains the self-holding of the contactor -Q1 and the emergency stop.

Occurs in Chapter 7 · The test rig

See also Miniature circuit breaker (-F2)· Test rig

The outputs of the controller act on the simulation, and its computed reports return by the same route to the inputs of the controller. An unchanged control program thus runs against the model.

Occurs in Chapter 2 · Objective and approach

See also Research questions· Return path

Lowest of the three software layers: gateway.py on port 8100, the bridge between serial hardware and HTTP.

Occurs in Chapter 6 · System architecture

See also Gateway· Application layer· Presentation layer

The single interface function of every plant model: compute_step(inputs, dt) returns the physical state (plant), the sensor bits (sensors) and the logical outputs (outputs_logical). The model tests in Chapter 10 run directly against this function.

Occurs in Chapter 4 · Requirements; Chapter 6 · System architecture; Chapter 9 · The simulation cycle; Chapter 10 · Five test scenarios; Chapter 12 · Outlook

See also Plugin· Sensor bits

Signals with which the controller actuates the plant: the inlet valves Y1 and Y2, the drain valve Y3, the vent valve Y4 and the heating command E1. They are the inputs of the model.

Occurs in Chapter 8 · The heated-tank model

See also Sensor (B1–B4)

Assembly that transmits signals between the 24 V level of the controller and the 5 V level of the signal bridge. Each direction has its own: optically through an optocoupler on the forward path, through potential-free relay contacts on the return path.

Occurs in Chapter 6 · System architecture; Chapter 7 · The test rig; Chapter 11 · Result and limits

See also Forward path· Return path· Galvanic isolation

Duration of one simulation cycle. Requirement NFA-02 demands at most 100 ms; on the Windows development computer the mean was 55.78 ms (17.9 Hz instead of 20 Hz). The bound is met on average, but not assured.

Occurs in Chapter 4 · Requirements; Chapter 9 · The simulation cycle; Chapter 11 · Result and limits

See also Simulation clock· Real-time capability· Non-functional requirements (NFA)

Level of the distinction by data flow: the real plant and the model exist side by side, and data is transferred manually in both directions.

Occurs in Chapter 3 · State of the art

See also Digital shadow· Digital twin

Level of the distinction by data flow: the plant feeds the model automatically, but not automatically back.

Occurs in Chapter 3 · State of the art

See also Digital model· Digital twin

Level of the distinction by data flow: data flows automatically in both directions. The distinction is based solely on the degree of automation, not on the accuracy of the model; in this thesis the real controller is the physical coupling partner.

Occurs in Chapter 2 · Objective and approach; Chapter 3 · State of the art

See also Digital model· Digital shadow

Conversion of the coupled equations of the model into computation steps of fixed step size. The model uses the explicit Euler method with Δt = 50 ms for this.

Occurs in Chapter 8 · The heated-tank model; Chapter 10 · Five test scenarios

See also Euler method (explicit)· Step size (Δt)

Deviation that arises solely from the step size of the method. The deviations from the analytical solution in scenario 2 are compatible with it.

Occurs in Chapter 10 · Five test scenarios

See also Discretization· Analytical solution

Transition of a sensor from 0 to 1 or vice versa. Only such edges are written to the signal bridge.

Occurs in Chapter 9 · The simulation cycle

See also Sensor (B1–B4)· Simulation clock

Operating state at the edge of the intended range. In scenario 5 it is a heating command on the empty tank, which runs over 1200 steps without a runtime error.

Occurs in Chapter 10 · Five test scenarios

See also Test scenario

Sub-model of the heated-tank model: it adds the heating energy and subtracts the evaporation energy and the outflow loss. The result is the update of the internal energy of the water.

Occurs in Chapter 2 · Objective and approach; Chapter 8 · The heated-tank model

See also Mass balance· State vector

Simplest one-step method: x(k+1) = x(k) + f(x(k), u(k)) · Δt. It fits into a fixed cycle, can be followed step by step and can be checked against the analytical solution.

Occurs in Chapter 8 · The heated-tank model; Chapter 10 · Five test scenarios

See also Discretization· Stability criterion (Δt/τ < 2)· Step size (Δt)

Four criteria of the test scenarios: plausibility (direction and order of magnitude), consistency (state and sensor bit match exactly), stability (the computation stays stable over tens of thousands of cycles) and visualization fidelity (the user interface shows the model values correctly).

Occurs in Chapter 10 · Five test scenarios

See also Test scenario

Signal direction from the controller to the simulation: 24 V is converted, galvanically isolated, to 5 V for the signal bridge through an optocoupler (PC817) (Fig. 14a). Twelve channels lead in this direction.

Occurs in Chapter 7 · The test rig

See also Return path· Optocoupler

The six requirements FA-01 to FA-06: read in the signals of the controller, compute the model at a fixed clock rate, report threshold switches in the exact step, display the plant state, be able to add further plants as model plugins and make errors recognizable.

Occurs in Chapter 4 · Requirements; Chapter 11 · Result and limits

See also Non-functional requirements (NFA)

Transmission of a signal without a conductive connection between two levels. On the forward path this is done optically through an optocoupler, on the return path through a potential-free relay contact.

Occurs in Chapter 6 · System architecture; Chapter 7 · The test rig

See also Optocoupler· Potential-free contact· Converter stage

Software component gateway.py on port 8100, the communication layer. It is the bridge between serial hardware and HTTP and works with its own read and write thread and automatic reconnection.

Occurs in Chapter 5 · Solution approaches and concept decision; Chapter 6 · System architecture; Chapter 9 · The simulation cycle; Chapter 11 · Result and limits

See also Communication layer· Orchestrator· Signal bridge

Gas space above the water in the tank, filled with steam and air; also called steam space. Its pressure follows from the steam mass and the air mass.

Occurs in Chapter 8 · The heated-tank model; Chapter 10 · Five test scenarios

See also Partial pressure (Dalton)· Heated-tank model

Regular sign of life of the coupling, running every 0.5 s. If it fails to appear, the watchdog takes effect.

Occurs in Chapter 9 · The simulation cycle

See also Watchdog

Computational model of a pressure vessel with a volume of 0.5 m³, containing water, above it a headspace of steam and air, heated with up to 150 kW. It is described by five state variables and four sub-models.

Occurs in Chapter 6 · System architecture; Chapter 8 · The heated-tank model; Chapter 10 · Five test scenarios; Chapter 11 · Result and limits

See also State vector· Plant model

Loading of plant models at runtime by the orchestrator.

Occurs in Chapter 6 · System architecture

See also Orchestrator· Plugin

Relation according to which the model derives the pressure in the headspace from the steam mass and the air mass with their gas constants. The partial pressures are then added according to Dalton.

Occurs in Chapter 8 · The heated-tank model

See also Partial pressure (Dalton)· Headspace

Exam board of the Chamber of Industry and Commerce, connected via a 32-pin plug-in card; around 30 are in use. The plant exists on it only as a printed silhouette, feedback is given exclusively by LEDs, and there is no internal plant state.

Occurs in Chapter 1 · Initial situation; Chapter 4 · Requirements; Chapter 8 · The heated-tank model; Chapter 13 · Preliminary projects

See also As-is state

The model computes what the equations demand. The tests of the thesis show no more than that: without measured data of a real tank, validation is limited to internal consistency.

Occurs in Chapter 10 · Five test scenarios; Chapter 11 · Result and limits; Chapter 12 · Outlook

See also Model fidelity· Test scenario

Layered model (Open Systems Interconnection) in which the thesis classifies the coupling (Fig. 12): browser, REST-HTTP and WebSocket at layer 7, JSON at layer 6, below that TCP (4), IP (3), UART byte framing (2) and USB and signal levels (1). There is no separate session layer; the physics model sits above layer 7.

Occurs in Chapter 3 · State of the art

See also REST-HTTP· WebSocket· JSON

Data format of the coupling; classified at layer 6 of the ISO/OSI model.

Occurs in Chapter 3 · State of the art

See also ISO/OSI reference model

Sub-model of the heated-tank model: the water mass grows via the inlet valves and decreases via the outflow and the evaporation. The outflow rate depends on the square root of the pressure.

Occurs in Chapter 2 · Objective and approach; Chapter 8 · The heated-tank model

See also Energy balance· State vector

Protective device in the circuit diagram of the test rig. -F2 is to be replaced with a two-pole one; after that the initial inspection by a qualified electrician is due, both before mains operation.

Occurs in Chapter 7 · The test rig; Chapter 11 · Result and limits; Chapter 12 · Outlook

See also Circuit diagram

Degree to which the equations match a specific real tank. It is not demonstrated in the thesis; a comparison with a real tank would raise the validation from internal consistency to model fidelity.

Occurs in Chapter 12 · Outlook

See also Internal consistency

Message Queuing Telemetry Transport: protocol in which all components communicate through a central broker, publishing to and subscribing to topics. It is an industry standard and was approach 1 of the concept decision; the in-house development of gateway and orchestrator was chosen instead.

Occurs in Chapter 5 · Solution approaches and concept decision

See also Broker· Topic

Practice partner in Oberlungwitz, where electronics technicians for industrial engineering and industrial electricians are trained.

Occurs in Chapter 1 · Initial situation

See also IHK simulation board

The three requirements NFA-01 to NFA-03: operation without external services, a cycle time of at most 100 ms and a model that can be tested without hardware and user interface. NFA-02 is only partly met.

Occurs in Chapter 4 · Requirements; Chapter 9 · The simulation cycle; Chapter 10 · Five test scenarios; Chapter 11 · Result and limits

See also Functional requirements (FA)· Cycle time

Panel on the front of the test rig with twelve push buttons and indicator lamps. It is connected in parallel with the controller, so that a push button and a relay contact at the input have the same effect.

Occurs in Chapter 4 · Requirements; Chapter 7 · The test rig

See also Test rig

Component that transmits a signal optically. In the test rig, the outputs of the controller switch the emitting diodes of an optocoupler assembly (PC817); the phototransistor in turn delivers 5 V to the signal bridge.

Occurs in Chapter 6 · System architecture; Chapter 7 · The test rig

See also Galvanic isolation· Forward path

Software component orchestrator.py on port 8050, the application layer. It calls the model in every cycle, fetches the process image from the gateway, loads models at runtime and supplies the browsers via WebSocket.

Occurs in Chapter 5 · Solution approaches and concept decision; Chapter 6 · System architecture; Chapter 9 · The simulation cycle; Chapter 11 · Result and limits

See also Application layer· Gateway· Simulation clock

Partial pressure that steam or air contributes in the headspace. According to Dalton, the partial pressures are added to give the total pressure.

Occurs in Chapter 8 · The heated-tank model

See also Ideal gas equation· Headspace

Computational model of a plant that implements the uniform interface compute_step and is loaded as a plugin. 13 plant models run on the platform; the heated-tank model is one of them.

Occurs in Chapter 1 · Initial situation; Chapter 2 · Objective and approach; Chapter 6 · System architecture; Chapter 11 · Result and limits

See also Plugin· compute_step

The software as a whole: gateway, orchestrator, plant models and browser interface. 13 plant models run on it without any change to the orchestrator.

Occurs in Chapter 2 · Objective and approach; Chapter 6 · System architecture; Chapter 7 · The test rig; Chapter 9 · The simulation cycle; Chapter 12 · Outlook

See also Test rig· Plant model

Industrial controller that is an integral part of the training. The standard IEC 61131 stands for the programmable logic controller itself; in the thesis, “controller” is the short form for the real PLC.

Occurs in Chapter 1 · Initial situation; Chapter 2 · Objective and approach; Chapter 3 · State of the art; Chapter 4 · Requirements

See also Siemens LOGO!

Exchangeable plant model with the same structure that the orchestrator knows through a single call (compute_step). The interface is only agreed as a convention, not formally specified.

Occurs in Chapter 4 · Requirements; Chapter 6 · System architecture; Chapter 11 · Result and limits; Chapter 12 · Outlook

See also compute_step· Plant model· Prototypical

Switching contact of the relay assembly on the return path. It applies 24 V to an input of the controller without a conductive connection to the 5 V side.

Occurs in Chapter 7 · The test rig

See also Relay assembly· Return path· Galvanic isolation

Top one of the three software layers: the browser with Canvas 2D animation, control panel and error log. It is supplied with the state via WebSocket.

Occurs in Chapter 6 · System architecture

See also Application layer· WebSocket

Current state of the inputs and outputs, which the orchestrator fetches from the gateway at the beginning of every cycle (GET /api/state).

Occurs in Chapter 9 · The simulation cycle

See also Orchestrator· Gateway

Limits in the model that intercept unphysical intermediate values. They apply only within the intended operating range.

Occurs in Chapter 11 · Result and limits

See also Internal consistency

Functional, but not formally specified. This is how the plugin interface is regarded: it carries 13 models, but is only agreed as a convention.

Occurs in Chapter 11 · Result and limits

See also Plugin

First-order lag element. The heating element of the model is a PT1 element with a time constant of τ = 12 s; its heating power follows the setpoint with a delay.

Occurs in Chapter 8 · The heated-tank model; Chapter 10 · Five test scenarios

See also Time constant (τ)· Step response

Computer on which the simulation runs and which is the target platform of the plant. The clock measurement, by contrast, comes from a Windows development computer.

Occurs in Chapter 6 · System architecture; Chapter 9 · The simulation cycle

See also Platform

Ability to assure a clock. A fixed clock is something different from a guaranteed clock: a general-purpose operating system gives no real-time guarantee, and the thesis does not claim otherwise.

Occurs in Chapter 3 · State of the art; Chapter 9 · The simulation cycle; Chapter 11 · Result and limits; Chapter 12 · Outlook

See also Cycle time· Simulation clock

Common reference point of a voltage level. In the test rig the reference potentials of the 24 V and 5 V levels are deliberately kept separate; every signal therefore has to be transmitted without a conductive connection.

Occurs in Chapter 7 · The test rig

See also Galvanic isolation

Assembly of the return path: the Arduino switches it through a driver stage, and its relays (coils at 12 V) apply 24 V to the inputs of the controller through potential-free contacts.

Occurs in Chapter 7 · The test rig

See also Potential-free contact· Return path

Three questions that run through the thesis: coupling (can a plant model be coupled to a compact controller so that the loop is closed?), model quality (is a model built from mass and energy balance sufficient?) and architecture (does an architecture without external services also hold with several models?).

Occurs in Chapter 2 · Objective and approach

See also Closed loop

Application protocol between gateway and orchestrator, at layer 7 of the ISO/OSI model. It requires no external middleware.

Occurs in Chapter 3 · State of the art; Chapter 5 · Solution approaches and concept decision

See also WebSocket· ISO/OSI reference model

Signal direction from the simulation to the controller: the Arduino switches a relay assembly whose potential-free contact applies 24 V to the input of the controller (Fig. 14b). Ten channels lead in this direction.

Occurs in Chapter 7 · The test rig

See also Forward path· Relay assembly

Pressure that the model calculates from the temperature to decide whether water evaporates: if it exceeds the pressure in the tank, the water boils. It rises steeply with temperature; at 100 °C it is about 1.01 bar.

Occurs in Chapter 8 · The heated-tank model

See also Antoine equation

Digital reports derived from the state of the model: B1 reports the upper fill level from 450 l, B2 the lower fill level at 50 l or less, B3 the temperature from 95 °C and B4 the pressure from 2.8 bar. Changed sensors travel back to the controller via relay contacts.

Occurs in Chapter 6 · System architecture; Chapter 8 · The heated-tank model; Chapter 9 · The simulation cycle

See also Sensor bits· Threshold switch· Control signals (Y1–Y4, E1)

Digital report states that the model derives from the physical state. They are the return value sensors of compute_step.

Occurs in Chapter 6 · System architecture; Chapter 10 · Five test scenarios

See also Sensor (B1–B4)· compute_step

Compact controller, used predominantly in the training operation. In the test rig it is the real controller that works with an unchanged program against the plant model.

Occurs in Chapter 1 · Initial situation; Chapter 2 · Objective and approach; Chapter 3 · State of the art; Chapter 6 · System architecture; Chapter 7 · The test rig

See also PLC (programmable logic controller)

Arduino Mega with firmware that reads and sets the 5 V signals of the converter level and transmits the state serially at 115200 baud to the gateway. If the sign of life fails to appear, the firmware switches all outputs off.

Occurs in Chapter 3 · State of the art; Chapter 6 · System architecture; Chapter 7 · The test rig; Chapter 9 · The simulation cycle; Chapter 11 · Result and limits

See also Arduino Mega· Gateway· Watchdog

Fixed cycle of 50 ms (20 Hz) in which the orchestrator calls the model. The simulation runs time-driven, not event-driven; the sequence of one cycle is shown in Fig. 9.

Occurs in Chapter 2 · Objective and approach; Chapter 3 · State of the art; Chapter 5 · Solution approaches and concept decision; Chapter 7 · The test rig; Chapter 8 · The heated-tank model; Chapter 9 · The simulation cycle; Chapter 10 · Five test scenarios; Chapter 11 · Result and limits

See also Cycle time· Real-time capability

Condition under which the explicit Euler method is stable for a PT1 element: Δt/τ < 2. Here the ratio is about 0.004; the statement concerns only the heating element.

Occurs in Chapter 8 · The heated-tank model

See also Euler method (explicit)· PT1 element

Groups whose needs give rise to the requirements: apprentices (testing a PLC program on a plant that reacts), trainers (a robust environment that starts quickly) and exam preparation (compatibility with the existing controllers).

Occurs in Chapter 4 · Requirements

See also Functional requirements (FA)

The five state variables x(k) of the heated-tank model: water mass, internal energy of the water, steam mass, air mass and currently effective heating power. Everything else, such as temperature and pressure, is calculated from them.

Occurs in Chapter 8 · The heated-tank model

See also Mass balance· Energy balance

Course of the heating power after a sudden setpoint step. For the PT1 element 63.2 % of the final value is reached after one time constant, about 95 % after three, 99.3 % after five (Fig. 8).

Occurs in Chapter 8 · The heated-tank model

See also PT1 element· Time constant (τ)

Time step of the simulation: Δt = 0.05 s, i.e. 50 ms. It is used both in the simulation clock and in the test scenarios.

Occurs in Chapter 6 · System architecture; Chapter 8 · The heated-tank model

See also Euler method (explicit)· Simulation clock

Delimitation of what belongs to the system: the simulation system that replicates the plant and displays its state. Controller, operator panel, monitor and operator adjoin it; only binary signals are exchanged (Fig. 17, Fig. 18).

Occurs in Chapter 4 · Requirements; Chapter 8 · The heated-tank model

See also Functional requirements (FA)

The complete wired setup as a mounting wall: on the front the controller, operator panel, indicator lamps, converter assemblies and screen of the running simulation, at the back power supplies and the wiring of the signal level. On it the real controller works against the plant model.

Occurs in Chapter 5 · Solution approaches and concept decision; Chapter 7 · The test rig; Chapter 11 · Result and limits

See also Platform· Operator panel

Defined run of the heated-tank model whose measurement is matched by an expectation. The thesis tests five scenarios directly at the model function: filling, heating, pressure buildup, emptying and an edge case with an empty tank.

Occurs in Chapter 2 · Objective and approach; Chapter 4 · Requirements; Chapter 5 · Solution approaches and concept decision; Chapter 10 · Five test scenarios; Chapter 11 · Result and limits

See also Evaluation criteria· Edge case· Internal consistency

Binary sensor that switches when a limit value is reached. Requirement FA-03 demands that it be reported in the exact step.

Occurs in Chapter 4 · Requirements

See also Sensor (B1–B4)· Functional requirements (FA)

Characteristic of a PT1 element: after one time constant the output has reached 63.2 % of its final value. The heating element of the model has τ = 12 s.

Occurs in Chapter 8 · The heated-tank model; Chapter 10 · Five test scenarios

See also PT1 element· Step response

Subject channel in MQTT: one component publishes to it, others subscribe to it.

Occurs in Chapter 5 · Solution approaches and concept decision

See also MQTT· Broker

Development methodology for mechatronic systems that the thesis follows. Every design step on the left is assigned a verification at the same level on the right: requirements to evaluation, concept design to the scenario run, system design to the interface test and domain design to the checking of the equations (module tests).

Occurs in Chapter 2 · Objective and approach; Chapter 3 · State of the art

See also Test scenario

Monitoring timer of the coupling, three-stage with a 2 s limit. If the sign of life fails to appear, the firmware of the signal bridge autonomously establishes a defined state (all outputs off) instead of freezing the last state.

Occurs in Chapter 9 · The simulation cycle

See also Heartbeat· Signal bridge

Application protocol between orchestrator and browser, at layer 7 of the ISO/OSI model. The state goes through it to all connected browsers in every cycle.

Occurs in Chapter 3 · State of the art; Chapter 5 · Solution approaches and concept decision; Chapter 6 · System architecture; Chapter 9 · The simulation cycle

See also REST-HTTP· Presentation layer