Method for automatic stability detection of a controller cascade

The method employs energy operators and limit detectors to automatically assess controller cascade stability, addressing the challenge of unstable control loops and preventing mechanical damage by indicating instability through a traffic light system.

DE102022207368B4Active Publication Date: 2024-12-05LENZE SE
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
DE102022207368
Authority / Receiving Office
DE · DE
Patent Type
Patents
Current Assignee / Owner
Filing Date
2022-07-19
Publication Date
2024-12-05
Estimated Expiration
2042-07-19

AI Technical Summary

Technical Problem

Existing methods fail to reliably and automatically detect instabilities in controller cascades, which can lead to mechanical damage under production conditions due to unstable control loops that are not easily identifiable by non-experts.

Method used

A method using energy operators and limit detectors to monitor control errors and manipulated variables, generating a stability measure through a traffic light system to indicate stability or instability, allowing for automatic detection and identification of unstable control loops.

Benefits of technology

Enables reliable, automatic detection of controller cascade instabilities, preventing mechanical damage by providing timely alerts and recommendations for corrective actions.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 00000000_0000_ABST
    Figure 00000000_0000_ABST
Patent Text Reader

Abstract

Method for automatic stability detection of a controller cascade (100) with a number of cascaded controllers (1, 2, 3), wherein a control error (4, 5, 6) of a respective controller (1, 2, 3) is processed by the respective controller (1, 2, 3) and the processed control error (4, 5, 6) is output as a controller manipulated variable (7, 8, 9), wherein a reference variable (13, 14) for the subsequent controller (2, 3) in the controller cascade (100) is determined as a function of the controller manipulated variable (7, 8), wherein the method comprises the following steps: a) Determining the energy content of each control error (4, 5, 6), b) Determine whether the magnitude of a respective controller control variable (7, 8, 9) exceeds an associated limit value or not, and c) Generating a stability measure (10) for the controller cascade (100) depending on the energy contents of the respective control errors (4, 5, 6) and depending on whether the respective controller manipulated variables (7, 8, 9) exceed their associated limit values ​​or not.
Need to check novelty before this filing date? Find Prior Art

Description

[0001] US 6 004 017 A discloses a method and a system for predicting limiting vibrations comprising the steps of generating a signal based on a vibration behavior, statistically processing the signal to generate a generalized Teager measure, monitoring changes in the generalized Teager measure and generating a classification signal if the changes exceed a predetermined threshold for a predetermined period of time, wherein the classification signal enables a prediction of the limiting vibrations.

[0002] The invention is based on the object of providing a method for the automatic stability detection of a controller cascade, which can detect possible instabilities of the controller cascade as simply and reliably as possible.

[0003] The method is used for automatic stability detection of a controller cascade. The controller cascade typically comprises a number of cascaded controllers, for example, between 1 and 6. A control error of a respective controller is processed by the respective controller according to its control algorithm, and the processed control error is conventionally output by the respective controller as a controller manipulated variable. A reference variable for the subsequent controller in the controller cascade is determined as a function of the controller manipulated variable of the preceding controller in the controller cascade. A control error refers to the difference between the setpoint or reference variable and the actual value of the variable to be controlled. With regard to the basic structure of controller cascades, reference is also made to the relevant specialist literature.

[0004] The procedure includes the following steps: a) Determining the energy content of each control error, b) Determine whether an amount of a respective controller control variable exceeds an associated limit value or not, and c) Generating a stability measure for the controller cascade depending on the energy contents of the respective control errors and / or depending on whether the respective controller manipulated variables exceed their associated limit values ​​or not.

[0005] The stability measure represents the stability of the individual control loops of the controller cascade or the stability of the entire controller cascade. The stability measure can, for example, be embodied as a numerical value with a predefined range of values. The range of values ​​can, for example, extend between 0 and 10, where 0 corresponds to maximum stability and 10 corresponds to maximum instability. The stability measure can also be designed as a stability traffic light, where, for example, green corresponds to high stability, yellow corresponds to incipient instability, and red corresponds to instability, etc.

[0006] In one embodiment, the energy content of the respective control error is determined using energy operators.

[0007] In one embodiment, when the energy content of the respective control errors increases, the stability measure is generated or changed in such a way that it indicates an increasing instability of the controller cascade.

[0008] In one embodiment, in the event that the magnitudes of the respective controller manipulated variables exceed their associated limit values, the stability measure is generated or changed in such a way that it indicates an increasing instability of the controller cascade.

[0009] In one embodiment, steps a) to c) are repeated at a fixed time interval, for example every 100 ms.

[0010] In one embodiment, the controllers of the controller cascade are selected from the set of controllers consisting of: position controllers, speed controllers and current controllers.

[0011] In one embodiment, diagnostic and action information is generated depending on the energy contents of the respective control errors and depending on whether the magnitudes of the reference variables generated by the controllers exceed their associated limit values ​​or not.

[0012] In one embodiment, the stability measure is generated further depending on predeterminable properties of the controller cascade.

[0013] The invention is described in detail below with reference to the drawing. Fig. 1 shows an example of a controller cascade with a number of cascaded controllers, which is designed to carry out the method according to the invention.

[0014] Fig. 1 shows a controller cascade 100 with three cascaded controllers 1, 2, 3, wherein a control error 4, 5, or 6 of a respective controller 1, 2, or 3 is processed by the respective controller 1, 2, or 3, and the processed control error 4, 5, or 6 is output as controller manipulated variable 7, 8, or 9, respectively. A reference variable 13 or 14 for the subsequent controller 2 or 3 in the controller cascade 100 is determined as a function of the controller manipulated variable 7 or 8, respectively. In this respect, reference is also made to the relevant specialist literature.

[0015] According to the invention, an energy content of a respective control error is calculated.

[0016] Limiters 15, 16 and 17 determine whether an amount of a respective controller manipulated variable 7, 8 or 9 exceeds a corresponding limiting value or not, and if necessary limit the controller manipulated variable 7, 8, 9 to the corresponding limiting value.

[0017] Finally, a stability measure 10 is determined for the controller cascade 100 depending on the energy contents of the respective control errors 4, 5, 6 and whether the respective controller manipulated variables 7, 8, 9 exceed their corresponding limit values. The stability measure is symbolized here, for example, by a traffic light, where a green light symbolizes high stability, a yellow light symbolizes incipient instability, and a red light symbolizes instability.

[0018] The energy content of the respective control error 4, 5, 6 is determined using energy operators.

[0019] Fig.Figure 1 shows a typical control loop cascade with control loops or controllers comprising a current controller 3, a speed controller 2, and a position controller 1. The driven load is represented here as a simple single-mass system with an electrical path 18 and a mechanical path 19. Element 21 is a conventional differentiating element.

[0020] Each controller 1, 2, and 3 must be adapted to its control loop to achieve the desired behavior. For example, current controller 3 must be adapted to a motor's inductance and resistance, with a dead time, a non-controllable variable of a power output stage, and a measured value acquisition factor being included in the calculation.

[0021] Speed ​​controller 2, in turn, must be adapted to a mechanical system, the properties of which have a significant influence on the value range of the speed controller settings. For single-mass systems, there is still a simple relationship between the moment of inertia J and Vpn, but as soon as elasticity or backlash come into play, the relationships become nonlinear and difficult to predict.

[0022] The position controller 1, in turn, sees the subordinate speed control loop as its control system, so that its properties have at least as significant an impact on the setting of the position controller Vp.

[0023] Typically, an initial setting of controllers 1, 2, and 3 is made during commissioning and then not further adjusted. The initial setting is based on criteria such as response time to setpoint changes, the ability to compensate for disturbances, or the degree to which the controller follows the setpoint.

[0024] Finding the optimal setting can be challenging for complex mechanical systems. Elastic elements, such as timing belts, long racks, etc., often leave very little room for maneuver when it comes to finding a suitable parameter set. This is especially true when, for example, special filters are used to combat the problematic natural frequencies of the mechanical system, such as notch filters or similar. Often, a compromise in the controller setting must then be made, in which the original optimization goal takes a back seat and the aspect of a stable and robust controller setting becomes dominant. Typically, one then approaches the limits of stable behavior quite closely in order to at least partially achieve one's own optimization goals.

[0025] Once this compromise has been found, there is still an interaction between mechanical properties and the resulting setting limits. If these mechanical properties then change over time due to aging or wear, it can often happen that the selected controller setting exceeds the limit of unstable behavior.

[0026] If the controller cascade becomes unstable overall, mechanical damage can occur if this is not stopped immediately. Since the machine is usually in the field under production conditions, there may not necessarily be a specialist on-site to detect or correct the faulty behavior.

[0027] This is where the invention comes in.

[0028] An important parameter in control is the so-called control error, i.e., the value beyond the subtraction point of the setpoint and actual value. The control error passes through the respective controller and is then amplified and passed on to the next controller as its setpoint or reference variable. A (setpoint) limiter is typically installed at this transfer point to ensure that a permissible setpoint, such as a maximum permissible current, a maximum permissible speed, etc., is not exceeded.

[0029] The invention is based on the finding that unstable control loops can be detected by excessive signal changes, which manifest themselves in control errors 4, 5 and 6 of the respective control loops.

[0030] To distinguish whether a control process is normal or an unstable situation, the energy content of the respective control errors 4, 5, and 6 can be used. So-called Teager Energy Operators EO, such as those according to Kaiser, which are known from the field of speech recognition, can be used for observation.

[0031] Since energy operators (EO) are usually inaccurate or even unusable with constant or limited signals, the invention checks whether limiters 15, 16, and / or 17 are activated. For this purpose, so-called limit detectors (LDs) are used as additional information sources. The LDs determine whether the value of a respective controller manipulated variable (7, 8, 9) exceeds a corresponding limit value. In other words, the LDs determine whether limiters 15, 16, and / or 17 are activated.

[0032] The EOs can be made sensitive to specific frequency ranges, allowing them to be precisely tuned to the characteristics of the machine and control system combination.

[0033] After commissioning the control system, limit values ​​for the EOs can be determined and set for the good and bad cases. Furthermore, resonance frequencies can be determined here, to which the EOs are tuned. This information is fed to a monitoring logic 20. Furthermore, the monitoring logic 20 is supplied with rudimentary information 12 about the mechanical structure of the underlying system.

[0034] During operation, the monitoring logic 20 is now active, which, in order to carry out the method according to the invention, detects not only the values ​​of the EOs but also the impact of the limiters 15, 16 and 17 by means of the LDs.

[0035] Control systems are often implemented in a time-discrete manner, i.e. the control tasks are carried out on a computer in a fixed time frame and in a fixed sequence.

[0036] The monitoring logic 20 is also linked to this time frame, which records every control event and fault event.

[0037] By simultaneously recording all EOs and LDs, it is possible to identify in which control loop an oscillation occurs first. Furthermore, the energy content can also be used to determine whether an instability is emerging or whether it is already fully active. This can be used to output a message about the status of the control loops via the traffic light system 10.

[0038] Since the source of the instability can also be identified, targeted measures can also be recommended. For example, if the current control loop becomes unstable, something must have changed in the electrical properties of the motor or converter. This can be determined more precisely, for example, by measuring currents, although currents and voltages are typically known in a control system. In this case, either recalibrating the current control loop or checking the actuator (converter) would be recommended.

[0039] If the speed control loop becomes unstable, the connecting elements are usually the cause. If the mechanical structure is known, for example, due to a user query during commissioning, the combination of EO values ​​and LD events can be used to distinguish between, for example, a change in belt tension and a change in backlash, from which an appropriate recommendation for action can be derived.

Claims

[1] Method for the automatic stability detection of a controller cascade (100) with a number of cascaded controllers (1, 2, 3), wherein a control error (4, 5, 6) of a respective controller (1, 2, 3) is processed by the respective controller (1, 2, 3) and the processed control error (4, 5, 6) is output as a controller manipulated variable (7, 8, 9), wherein a reference variable (13, 14) for the respective subsequent controller (2, 3) in the controller cascade (100) is determined as a function of the controller manipulated variable (7, 8), the method comprising the steps of: a) Determining the energy content of a respective control error (4, 5, 6), b) determining whether an amount of a respective controller manipulated variable (7, 8, 9) exceeds an associated limit value or not, and c) generating a stability measure (10) for the controller cascade (100) as a function of the energy contents of the respective control errors (4, 5, 6) and as a function of whether the respective controller manipulated variables (7, 8, 9) exceed their associated limit values ​​or not. [2] Method according to claim 1, characterized by , that - the energy content of the respective control error (4, 5, 6) is determined using energy operators. [3] Method according to one of the preceding claims, characterized by , that - when the energy contents of the respective control errors (4, 5, 6) increase, the stability measure (10) is generated in such a way that it indicates an increasing instability of the controller cascade (100). [4] Method according to one of the preceding claims, characterized by , that - in the event that the amounts of the respective controller manipulated variables (7, 8, 9) exceed their associated limit values, the stability measure (10) is generated in such a way that it indicates an increasing instability of the controller cascade (100). [5] Method according to one of the preceding claims, characterized by , that - that steps a) to c) are repeated at a fixed time interval. [6] Method according to one of the preceding claims, characterized by , that - the controllers (1, 2, 3) of the controller cascade (100) are selected from the set of controllers consisting of: - position controllers (1), - speed controllers (2), and - Current regulators (3). [7] Method according to one of the preceding claims, characterized by the step: - generating diagnostic and action information (11) depending on the energy contents of the respective control errors (4, 5, 6) and depending on whether the amounts of the reference variables (7, 8, 9) generated by the controllers (1, 2, 3) exceed their associated limit values ​​or not. [8] Method according to one of the preceding claims, characterized by , that - the generation of the stability measure is further dependent on predeterminable properties (12) of the controller cascade (100).

Citation Information

Patent Citations

  • Teager-based method and system for predicting limit cycle oscillations and control method and system utilizing same

    US6004017A