DE102022113515A1 - Method and device for analyzing the operational behavior of technical systems - Google Patents

Abstract

The invention relates to a method and a device for analyzing the operational behavior of technical systems and in particular relates to such a method and such a device for analysis by means of statistical evaluation. An analysis method for technical systems is disclosed, which has the steps: determining the reaction of the technical system using a Monte Carlo analysis with several samples, whereby a deviation of a desired target state from an actual target state of the technical system is determined by changing a manipulated variable (including (i;n)) are minimized in a control loop (400), and a disturbance variable (d(i)) of the control loop (400) and / or a control quality of a controller (402) are randomly determined in the samples.

Description

The invention relates to a method and a device for analyzing the operational behavior of technical systems and in particular relates to such a method and such a device for analysis by means of statistical evaluation.

In order to determine the behavior of technical systems, for example in a risk assessment, for example as part of an ETCS (European Train Control System) risk analysis or in an ETCS braking curve determination, it is necessary to analyze and evaluate the behavior of technical systems, for example a braking system of a rail vehicle .

For example, from the standard UIC B 127 / DT 414:2006 it is known to use Monte Carlo analyzes for the analysis of the braking ability of rail vehicles, based on a hierarchy model that mathematically models a technical system, for example a braking system of a rail vehicle. The hierarchy model is a mathematical modeling of the system architecture, here a braking system for rail vehicles, and the properties of the individual system components with regard to the stochastic spread of their characteristics, their reliability and failure probabilities and their interdependencies to fulfill the function of the overall system. Applied to technical systems, Monte Carlo analyzes are able to describe the statistical distribution of functionally relevant properties and take into account the failure and deviation probabilities that influence the desired function, here the braking function.

The disadvantage of the prior art is that the methods taught there are suitable for analyzing controlled systems with fixed nominal values of the system parameters without readjusting in the event of deviations in the reaction of the overall system, but are not suitable for compensating for undesirable functional deviations take into account. In particular, it is not possible to take the behavior of a control loop integrated into the system into account accordingly.

It is therefore the object of the invention to eliminate the disadvantages present in the prior art and, in particular, to provide a method and a device which allow a more precise analysis of technical systems with an integrated control loop.

This task is solved by the subordinate claims. Advantageous developments of the invention are the subject of the subclaims.

An analysis method for technical systems is disclosed, which has the steps: determining the reaction of the technical system using a Monte Carlo analysis with several samples, whereby a deviation of a desired target state from an actual target state of the technical system is determined by changing a manipulated variable in one Control loop is minimized, and a disturbance variable of the control loop and / or a control quality of a controller are randomly determined in the samples.

It is further disclosed that determining the reaction of the technical system comprises the steps: a) initiating an analysis cycle for each sample, whereby the disturbance variable of the control loop and/or a control quality of the controller are randomly determined in the samples, b) carrying out the analysis cycle for each sample, the implementation comprising the steps: ba) carrying out at least one control cycle, wherein the deviation of the desired target state from the actual target state is minimized by changing the manipulated variable, and bb) storing the controlled variable as a result of the analysis cycle for the sample.

It is further disclosed that step ba) further comprises at least one of the steps: carrying out a further control cycle if the controlled variable is outside a predetermined value range, carrying out a further control cycle if a predetermined number of control cycles has not been exceeded in the analysis cycle, applying a control quality the controller, or modifying the manipulated variable according to the control quality of the controller.

It is further disclosed that step bb) further comprises at least one of the steps: determining the controlled variable based on a modified manipulated variable.

It is further disclosed that step a) further comprises at least one of the steps: determining the control quality of the controller from a stochastic distribution function and the associated deviation values, determining the disturbance variable of the controlled system from a stochastic distribution function and the associated deviation values.

It is further disclosed that the controller is optionally configured with control limits or without control limits, the controller is optionally configured as a continuous controller or as a step controller, or the controller is optionally configured as a two-sided controller or as a one-sided controller.

It is also revealed that the disturbance variable represents partial failures and deviations of the system components for a specific analysis cycle.

It is also revealed that a controlled system of the control loop contains a hierarchy model that mathematically models the technical system.

Furthermore, an analysis system for technical systems is disclosed, which has an electronic data processing device which is set up to carry out a method according to one of the preceding method claims.

Furthermore, an analysis computer program for technical systems is disclosed, which is set up to allow an electronic data processing device to carry out a method according to one of the preceding method claims when the analysis computer program is executed on the electronic data processing device.

• 1 zeigt einen Schienenfahrzeugverbund mit Bremssystemen.
• 2 zeigt das Bremssystem des Schienenfahrzeugs.
• 3 zeigt ein Blockschaltbild eines Regelkreises.
• 4 zeigt ein Blockschaltbild einer Monte-Carlo-Simulation gemäß der Erfindung.
• 5 zeigt ein Ablauf-Diagramm gemäß einem Ausführungsbeispiel der Erfindung.
• 6 zeigt ein Ablauf-Diagramm gemäß einem weiteren Ausführungsbeispiel der Erfindung.
• 7 zeigt ein Diagramm zur ETCS Gefährdungswahrscheinlichkeit.

Exemplary embodiments of the invention are described using the figures.

• 1 shows a rail vehicle network with braking systems.
• 2 shows the braking system of the rail vehicle.
• 3 shows a block diagram of a control loop.
• 4 shows a block diagram of a Monte Carlo simulation according to the invention.
• 5 shows a flow chart according to an exemplary embodiment of the invention.
• 6 shows a flow chart according to a further exemplary embodiment of the invention.
• 7 shows a diagram of the ETCS threat probability.

Based 1 A first exemplary embodiment of the invention is described.

1 discloses a rail vehicle group 1 with rail vehicles 110, 150, 160. The rail vehicle group has a driven car 110, in which a drive means 112, which in the exemplary embodiment is implemented by an electric motor, is provided. The driving force provided by the drive means 112 is transmitted to drive wheels 116 via driven wheel axles 114. The drive wheels 116 run on rails 118.

Furthermore, the driven car 110 has a pantograph 120, via which the train combination 1 is supplied with electrical energy via an overhead line 122 and the rail 118. Electrical energy is temporarily stored via an accumulator 123.

The driven car 110 has a driver's cab 124. In this, the train combination can be controlled by a train driver (not shown).

A braking system 200, via which the train combination 1 can be decelerated, will be discussed later using 2 described.

The driven carriage 110 is connected to one or more pulled carriages 150 via a coupling 140. If there are several towed cars 150, these are also connected to one another with a coupling 140. The towed carriage 150 has non-driven wheel axles 154 with non-driven wheels 156.

The towed car 150 has a braking system 152, via which the train combination 1 can be slowed down. In the exemplary embodiment, the brake system 152 of the towed car is constructed identically to the brake system 200 of the driven car. For its description, reference is made to the description of the braking system 200 of the driven car.

A last pulled car 160 in the train combination 1 is configured as the end car. It is constructed like the towed car 150, but has its own driver's cab 162 for reversing.

Tensile and compressive forces are transmitted between the respective cars 110, 150, 160 of the train combination via the couplings 140. Furthermore, all wagons in the train combination are supplied with electrical energy via the couplings 140 via an electrical supply line 126 via the driven wagon 110 and the pantograph 120. Furthermore, all cars of the train group 1 are connected via the couplings 140 via a control line 128, so that data communication can take place between the cars 110, 150, 160 of the train group 1. In the exemplary embodiment, the control line is designed as a CAN bus, via which components in the carriages 110, 150, 160 can exchange sensor and control data with one another.

Based on 2 the braking system 200 is described. The braking system 200 is a mechanical braking system, for example a pneumatic or electromechanical braking system. A pneumatic braking system is a braking system that has the ability to generate a deceleration/braking force of a vehicle using compressed air-operated components. An electromechanical braking system is a braking system that provides the ability to generate deceleration/braking force of a vehicle using electronic, electrical and mechanical components. The brake system 200 has a power supply interface 202, which is connected to the electrical supply line 126 and via which the brake system is supplied with operating energy. Furthermore, the brake system 200 has a control line interface 204 which is connected to the control line 128. The brake system 200 has a moving, in particular rotary, friction element in the form of a brake disk 206, which is connected in a rotationally fixed manner to one of the axles 114, 154 so that torque can be transmitted between the wheel axle 114, 154 and the brake disk 206. A brake caliper 208 is mounted in the car so that it does not move relative to the car and has static friction elements in the form of brake shoes 210 which are mounted in the brake caliper 208 so that they can be moved back and forth towards the brake disc 206. Via friction element drive means in the form of a brake actuator motor 212 and a spindle drive 214, the brake shoes 210 can be moved with a movement component perpendicular to the axis of rotation of the brake disk 206 and friction can be generated between the brake shoes 210 and the brake disk 206, a torque can be transmitted between the brake caliper 208 and the brake disk and the car will be delayed.

Based on 3 a control loop 300 for the brake system 200 is described.

The control loop 300 refers to the dynamic sequence of actions between a controller 300 and a controlled system 302 for influencing a controlled variable y(t) in a closed system, in which the controlled variable y(t) is continuously measured and compared with the reference variable w(t). What is essential here is the feedback (negative feedback) of the current value of the controlled variable y(t) to the controller as a control deviation e(t) = w(t) - y(t), which continuously counteracts a deviation from the reference variable. A reference variable is called a setpoint. The control deviation e(t) is transmitted to a controller (300), which determines a manipulated variable u(t) and applies it to a controlled system that is exposed to disturbance variables d(t). There may be control limits within which the manipulated variable u(t) must move. If these are not the case, the controller 300 is an ideal controller that can compensate for any control deviation e(t).

In the present example of a braking system for rail vehicles, the reference variable is, for example, a predetermined deceleration of, for example, -2m/s ² . When braking is initiated, this value is transmitted as a control deviation e(t) to the brake controller 300, which uses it to determine a manipulated variable u(t), for example a force acting on the brake shoe 210 with which it is pressed against the brake disc 206 and so on Deceleration of the rail vehicle 1 is effected. The controlled variable y(t) resulting from this control in the form of a measured deceleration of the rail vehicle 1 is dependent on disturbance variables, such as deviations in the friction coefficient μ of the brake shoes 210 and the brake disc 208, the gradient of the route or partial failures of individual elements of the brake system such as failures of the Braking system on one or more axles. The measured controlled variable y(t) is fed back and taken into account in the control deviation e(t) = w(t) - y(t).

Based on 4 A Monte Carlo analysis according to the invention of the operational behavior of a controlled system in general and of a braking system for rail vehicles in particular is described.

A Monte Carlo analysis or Monte Carlo simulation is a method from stochastics or probability theory in which random samples of a distribution are repeatedly taken using random experiments. The aim is to numerically solve problems that cannot be solved analytically or can only be solved with great difficulty using the samples taken. The random experiments can either be carried out in real life or simulated in computer calculations. The latter is the case in the present exemplary embodiment.

a

stochastisch streuende Verzögerung

a0

nominelle Verzögerung

i

laufende Nummer der Monte-Carlo-Analysezyklen

Kdry

Verhältnis der durch einen Monte-Carlo-Analysezyklus bestimmten stochastisch streuenden Verzögerung a und nomineller Verzögerung a0 bei ausreichend verfügbarem Kraftschluss

n

laufende Nummer des Regelzyklus des Regelkreises für einen Monte-Carlo-Berechnungszyklus

In 4 are:

a

stochastically scattering delay

a0

nominal delay

i

Consecutive number of the Monte Carlo analysis cycles

Kdry

Ratio of the stochastically scattering delay a determined by a Monte Carlo analysis cycle and the nominal delay a0 with sufficient available adhesion

n

Consecutive number of the control cycle of the control loop for a Monte Carlo calculation cycle

For the analysis, the technical system, here the braking system 200, is considered activated. Such activation can be triggered by activation criteria. Such activation criteria can, for example, be below or above progression of predetermined target values, such as exceeding a certain speed of the rail vehicle, the state of a system or subsystem, for example the overheating of a service brake system, an error reaction of the system or subsystem, for example the failure of a service brake system, or the like.

The control loop is characterized by the activation criteria, control limits and control quality.

The response of the system to be analyzed is determined using a Monte Carlo analysis and its nominal parameter states. The state resulting from the control, i.e. the reaction of the system, is determined under the influence of subsystem errors and tolerances of the subfunctions.

The deviation of the resulting state, ie from the desired state, is determined by the control loop 4 balanced by varying the manipulated variables. The desired reaction of the system is the reference variable and the resulting state is the controlled variable.

A reference variable w(i) is the setpoint of the quotient of a stochastically controlling delay and a nominal delay: $$a\text{/}{a}_0=K_{dry}=1$$

A controller 402 in the control loop 400 receives the control deviation e(i;n) as input, i.e. a control deviation depending on the current Monte Carlo analysis cycle i and current control cycle n and outputs a manipulated variable u(i;n). The controller 402 can have control limits, i.e. the output values of the manipulated variable u(i;n) cannot exceed or fall below predetermined threshold values, or cannot have such control limits. Furthermore, the controller 402 can be a continuously regulating controller, which can continuously increase or decrease the manipulated variable u(i;n), or a graduated controller, which can only output certain discrete values as manipulated variables u(i;n). Furthermore, the controller can be a two-way controller that can increase or decrease the manipulated variable u(i;n), or a one-way controller that can only manipulate the manipulated variable u(i;n) in one direction, or a combination thereof .

A controlled system 404 models a hierarchy model of the system functions/components. The hierarchy model is a mathematical modeling of the system architecture, here a braking system for rail vehicles, and the properties of the individual system components with regard to the stochastic spread of their characteristics, their reliability and failure probabilities and their interdependencies to fulfill the function of the overall system.

Based on 5 The process of a first exemplary embodiment of the analysis method according to the invention is described. In this first exemplary embodiment, complete control quality is assumed in the initial state of the control loop 400.

In a step S101 (initialization step), complete control quality is assumed in the initial state of the control loop and a/a ₀ = K _dry = 1 is set as the Monte Carlo analysis cycle-dependent reference variable w(1). The controller 402 is configured either with control limits or without control limits. Furthermore, the controller 402 is configured either as a continuous controller or as a step controller, or as a two-sided controller or as a single-sided controller. Furthermore, the controlled system 404 is defined as a hierarchy model and configured with predetermined disturbance variables d(1) from (partial) failures and tolerance deviations of system components.

In a step S102 (control step), the control cycle of the first Monte Carlo analysis cycle is started. From the control deviation e(1;1) = 0, the controller 402 determines a manipulated variable u(1;1), which is intended to achieve a/a ₀ = K _dry = 1 if a/a ₀ ≠ 1. This happens if necessary, taking existing control limits into account. The manipulated variable u(1;1) is applied to the controlled system 404 and is exposed there to the preconfigured disturbance variables d(1).

If the resulting controlled variable y(1;1) moves within predetermined limits, or if another termination criterion such as running through a predetermined number of control cycles n is reached, the process continues with step S 103, otherwise the control loop 400 is run through again at step S102.

In step S103 (evaluation step), the manipulated variable u(i;n) from the last control pass, which corresponds to an ideal manipulated variable, is modified by a predetermined control quality, the control quality being described by a stochastic distribution function and the associated deviation values. The modified manipulated variable u(i;n) is applied to the controlled system and the resulting controlled variable K _dry;i is saved as the result of the sample of the Monte Carlo analysis cycle.

If a sufficiently large sample quantity is reached, the analysis process is ended, otherwise the process continues from step S101 Changed disturbance variables d(i) run through again in the next analysis cycle.

Based on 6 The process of a second exemplary embodiment of the analysis method according to the invention is described. In this second exemplary embodiment, random control quality is assumed in the initial state of the control loop 400.

In a step S201 (initialization step), a control quality with a Monte Carlo analysis cycle-specific random control accuracy is assumed in the initial state of the control loop. The control quality is determined by a stochastic distribution function, for example a normal distribution with associated deviation values. In addition, a/a ₀ = K _dry = 1 is set as a Monte Carlo analysis cycle-dependent reference variable w(1). The controller 402 is configured either with control limits or without control limits. Furthermore, the controller 402 is configured either as a continuous controller or as a step controller, or as a two-sided controller or as a single-sided controller. Furthermore, the controlled system 404 is defined as a hierarchy model and configured with predetermined disturbance variables d(1) from (partial) failures and tolerance deviations of system components.

In a step S202 (control step), the control cycle of the first Monte Carlo analysis cycle is started. Taking into account the predetermined control accuracy, the controller 402 determines a manipulated variable u(1;1), which is intended to achieve a/a ₀ = K _dry = 1 if a/a ₀ ≠ 1. This happens, if necessary, taking existing control limits into account. The manipulated variable u(1;1) is applied to the controlled system 404 and is exposed there to the disturbance variables d(1) determined in step S201.

If the resulting controlled variable y(1;1) moves within predetermined limits, or if another termination criterion such as running through a predetermined number of control cycles n is reached, the process continues with step S203, otherwise the control loop 400 is run through again at step S202.

In step S203 (evaluation step), the system reaction is determined using the manipulated variables u(i;n) determined in step S202, ie the Monte Carlo analysis cycle-specific controlled variable K _dry;i , stored as the result of a Monte Carlo analysis cycle and thus the ongoing Monte Carlo analysis cycle ended.

If a sufficiently large sample quantity is reached, the analysis process is ended, otherwise the process is run through again from step S201 with changed control accuracy and changed disturbance variables d(i) in the next analysis cycle.

A system for Monte Carlo analysis is disclosed. The system has an electronic data processing device that is set up to carry out the analysis methods described above. In particular, the electronic data processing device has programming by a computer program that is set up so that the electronic data processing device can carry out one of the analysis methods.

7 shows a diagram of the ETCS threat probability. 7 shows on the ordinate the hazard probability P(k≥K _{dry_rst} ) used for the ETCS and on the abscissa the ETCS correction factor K _{dry_rst} . How out 7 As can be seen, the consideration of adaptive systems made possible by this invention allows a larger correction factor with the same probability of danger. The correction factor represents the proportion of the nominally expected delay that cannot be expected to be smaller given the hazard probability assumed. This means that larger correction factors lead to higher delays than smaller correction factors.

• Systementwicklern ist mit den Monte-Carlo-generierten K_{dry_rst}-Werten eine modulare Auslegungsunterstützung verfügbar, die die bisherige Beschränkung auf eine bestimmte definierte Bremssystemarchitektur aufhebt.

The teaching disclosed here, which also allows the interactions between the individual brake system components to be taken into account, generates advantages over the prior art in at least one of the following points:

• The Monte Carlo-generated K _{dry_rst} values provide system developers with modular design support that removes the previous limitation to a specific, defined brake system architecture.

Monte Carlo analyzes can support system developers in calculating the dynamic braking behavior and thus also specifically quantify the extent to which tolerances or the assumed failure of any system component can be reliably compensated for by a brake control system by another braking function available in the train.

The invention was described using exemplary embodiments. The embodiments are merely illustrative in nature and do not limit the invention as defined by the claims. As will be apparent to those skilled in the art, deviations from the exemplary embodiment are possible without departing from the scope of protection of the claims.

The analysis of the technical system was described using an electrically operated rail vehicle and vehicle network. However, the analysis of the technical system is not only applicable to such vehicles, but to any type of vehicle or, even further, to any technical system.

Furthermore, the analysis method was described using a brake system. However, the analysis method is suitable for other technical systems, for example systems for propulsion control of vehicles, in particular rail vehicles.

In particular, it is possible to combine features from different exemplary embodiments.

The junctions ... "and", "or" and "either ... or" are used in the meaning reminiscent of the logical conjunction (logical AND), the logical adjunction (logical OR, often "and/or") , or the logical contravalence (logical exclusive OR). In particular, in contrast to “either ... or,” the junctor “or” can contain the joint presence of both operands.

A list of process steps in the description and the claims only has the function of listing the required process steps. It does not imply any necessary order or sequence of the procedural steps, unless such order or sequence is explicitly stated or is obvious to the person skilled in the art. Furthermore, such a list does not mean that it is complete.

The term “having” does not imply an exhaustive list in the claims; the presence of additional elements and steps is possible.

The use of the indefinite article "a" or "an" does not exclude the presence of a plural, but is to be understood as "at least one" or "at least one", unless it is translated as "exactly one" or "exactly one". " restricted.

Furthermore, within the scope of this invention, the following terms are understood in the meaning given below.

Deceleration describes the relative reduction in speed of a vehicle in the form of a negative acceleration by applying a force that is directed against the direction of travel of the vehicle and is usually generated by increasing frictional resistance on the wheels or axles.

Control quality describes the permanent control deviation. With complete control quality, the remaining control deviation is zero. To describe the control quality, the remaining control deviation can be specified or as a stochastic distribution function for random experiments.

Monte Carlo analysis / Monte Carlo simulation describes methods from stochastics or probability theory in which random samples of a distribution are repeatedly taken using random experiments. The aim is to numerically solve problems that cannot be solved analytically or can only be solved with great difficulty using the samples taken. The random experiments can either be carried out in real life or simulated in computer calculations. In particular, the Monte Carlo analysis consists of several Monte Carlo analysis cycles to process a sufficiently large number of samples.

Monte Carlo analysis cycle / analysis cycle describes the processing of a sample to determine a result related to the sample. Here it consists of one or more control cycles.

Control cycle describes a passage through the control loop.

REFERENCE SYMBOL LIST

1

Rail vehicle network

110

powered car

112

Drive means, electric motor

114

driven wheel axle

116

drive wheel

118

rail

120

pantograph

122

Overhead line

123

accumulator

124

Driver's cab of the driven car

126

electrical supply line

128

Control line

140

coupling

150

pulled wagon

154

non-driven axles

156

non-driven wheels

160

End car

162

Driver's cab of the final car

200

Brake system (first embodiment)

202

Power supply interface

204

Control line interface

206

moving friction element, rotary friction element, brake disc

208

caliper

210

static friction element, brake shoes

212

Friction element drive means, brake actuator motor

214

Spindle drive

300

Control circuit (brake system)

302

Regulator/brake controller (brake system)

304

Controlled system (braking system)

400

Control loop (system analysis)

402

Controller/brake controller (system analysis)

404

Controlled system (system analysis)

Claims (10)

An analysis method for technical systems that has the steps: Determine the response of the technical system using a multi-sample Monte Carlo analysis, where a deviation of a desired target state from an actual target state of the technical system is minimized by changing a manipulated variable (u(i;n)) in a control loop (400), and a disturbance variable (d(i)) of the control loop (400) and/or a control quality of a controller (402) are randomly determined in the samples. The procedure according to Claim 1 , wherein determining the reaction of the technical system comprises the steps: a) initiating an analysis cycle for each sample, the disturbance variable (d(i)) of the control loop (400) and/or a control quality of the controller (402) being random in the samples b) carrying out the analysis cycle for each sample, the implementation having the steps: ba) carrying out at least one control cycle, the deviation of the desired target state from the actual target state being determined by changing the manipulated variable (u(i;n)). is minimized, and bb) storing the controlled variable (K _dry;i ) as a result of the analysis cycle for the sample. The method according to the preceding claim, wherein: step ba) further comprises at least one of the steps: Carrying out another control cycle if the controlled variable is outside a predetermined value range, Carrying out another control cycle if a predetermined number of control cycles has not been exceeded in the analysis cycle, Applying a control quality to the controller (402), or Modifying the manipulated variable (u(i;n)) according to the control quality of the controller (400). The procedure according to one of the Claims 2 or 3 , where: step bb) further comprises at least one of the steps: determining the controlled variable (K _dry;i ) based on a modified manipulated variable u(i;n). The procedure according to one of the Claims 2 or 3 , wherein: step a) further comprises at least one of the steps: determining the control quality of the controller (402) from a stochastic distribution function and the associated deviation values, determining the disturbance variable (d (i)) of the controlled system (404) from a stochastic distribution function and the associated deviation values. The method according to any one of the preceding claims, wherein the controller (402) is configured either with control limits or without control limits, the controller (402) is configured either as a continuous controller or as a step controller, or The controller (402) can be configured either as a two-sided controller or as a one-sided controller. The method according to one of the preceding claims, wherein the disturbance variable (u(i,n)) represents partial failures and deviations of the system components for a specific analysis cycle. The method according to one of the preceding claims, wherein a controlled system (404) of the control loop (400) contains a hierarchy model that mathematically models the technical system. An analysis system for technical systems, which has an electronic data processing device that is set up, a method to carry out according to one of the preceding method claims. An analysis computer program for technical systems, which is set up to allow an electronic data processing device to carry out a method according to one of the preceding method claims when the analysis computer program is executed on the electronic data processing device.

Images