Reference model switching method and device based on model estimation error trend

Abstract

The invention discloses a reference model switching method and device based on a model estimation error trend, and the method comprises the steps: constructing a segmented model of a controlled object system, and collecting the system state information of the controlled object system through a sensor; extracting a first model and a second model which are adjacent from the model set according to the system state information, and respectively estimating a first system state value and a second system state value of the first model and the second model; calculating errors between the system state value, collected by the sensor, of the controlled object system and the first system state value and the second system state value to obtain a first deviation value and a second deviation value, calculating a first error trend set based on the first deviation value, and calculating a second error trend set based on the second deviation value; and respectively calculating a first sub-control law and a second sub-control law according to the first model and the second model, and outputting a target control law in combination with the first deviation value, the second deviation value, the first error trend set and the second error trend set.

Description

Technical Field

[0001] The present application relates to the field of ship power system control, and more specifically, to a reference model switching method and device based on model estimation error trend. Background Art

[0002] At present, the ship power controller generally adopts a multi-stage control parameter set to match the dynamic process requirements under all working conditions. A single linearized model cannot match the control requirements of all working conditions. Therefore, different reference models are generally selected for the controller according to the working point during operation. The switching timing of the reference model has an important impact on the stable control of the system. The wrong switching timing may even cause the control law to diverge, causing the system working condition switching to fail and enter an unsteady state.

[0003] The existing direct model switching method based on preset parameters will cause problems such as sudden changes in the control law output and large fluctuations in the system state parameters during the switching process. Summary of the invention

[0004] In response to at least one defect or improvement need in the prior art, the present invention provides a reference model switching method and device based on the model estimation error trend, which solves the technical problem that the switching process may cause sudden changes in the control law output and large fluctuations in the system state parameters, achieves smooth switching of the reference model, and alleviates equipment deterioration caused by drastic changes in control instructions.

[0005] To achieve the above-mentioned purpose, according to the first aspect of the present invention, a reference model switching method based on model estimation error trend is provided, the method comprising: constructing a segmented model of a controlled object system, collecting system state information of the controlled object system through a sensor, wherein the segmented model comprises an initial zero dynamic model, and the system state information comprises the current working condition of the controlled object system; extracting adjacent first models and second models from a model set according to the system state information, and estimating the first system state value and the second system state value of the first model and the second model respectively; calculating the error between the system state value of the controlled object system collected by the sensor and the first system state value and the second system state value, respectively obtaining a first deviation value and a second deviation value, calculating a first error trend set based on the first deviation value, and calculating a second error trend set based on the second deviation value; calculating a first sub-control law and a second sub-control law according to the first model and the second model, respectively, and outputting a target control law by combining the first deviation value and the second deviation value, and the first error trend set and the second error trend set.

[0006] In an exemplary embodiment, after constructing a segmented model of the controlled object system and collecting system state information of the controlled object system through sensors, the method also includes: a supervision rule selects adjacent first and second models from a model set including multiple fixed models at each calculation step according to system operating parameters in the system state information; the first model serves as the current initial state model of the adaptive reference model, and the second model serves as an alternative state model of the adaptive reference model.

[0007] In an exemplary embodiment, after calculating the errors between the system state value of the controlled object system acquired by the sensor and the first system state value and the second system state value to obtain the first deviation value and the second deviation value respectively, the method further includes: constructing a first initial error set based on the first deviation value, wherein the first initial error set rollingly records the errors of ten consecutive calculation steps of the first model when the algorithm starts; constructing a second initial error set based on the second deviation value, wherein the second initial error set rollingly records the errors of ten consecutive calculation steps of the second model when the algorithm starts.

[0008] In an exemplary embodiment, after constructing a second initial error set based on the second deviation value, the method further includes: calculating a first error trend set based on the first initial error set, and counting the first number of negative values, the first sum of positive values, and the first sum of negative values ​​in the first error trend set; calculating a second error trend set based on the second initial error set, and counting the second number of negative values, the second sum of positive values, and the second sum of negative values ​​in the second error trend set.

[0009] In an exemplary embodiment, after calculating the second error trend set based on the second initial error set and counting the second negative value number, the second positive value sum and the second negative value sum in the second error trend set, the method also includes: when the first negative value number is greater than the second negative value number, the output target control law is the first sub-control law; when the first negative value number is less than the second negative value number, the output target control law is the second sub-control law; when the first negative value number is equal to the second negative value number, the target control law is output.

[0010] In an exemplary embodiment, when the first negative value is equal to the second negative value, the output target control law includes: determining the first weight and the second weight based on the first negative value sum and the second negative value sum, and calculating the output weight by the first weight and the second weight; and determining the target control law by combining the first sub-control law and the second sub-control law using the output weight.

[0011] In an exemplary embodiment, after determining the target control law by combining the first sub-control law and the second sub-control law using the output weight, the method further includes: limiting the output change rate of the target control law based on the first calibration value and the second calibration value, wherein the output change rate is the difference between two adjacent outputs of the target control law.

[0012] According to the second aspect of the present invention, there is also provided a reference model switching device based on model estimation error trend, which includes: an acquisition unit, used to construct a segmented model of a controlled object system, and to acquire system state information of the controlled object system through a sensor, wherein the segmented model includes an initial zero dynamic model, and the system state information includes the current working condition of the controlled object system; an estimation unit, used to extract adjacent first models and second models from a model set according to the system state information, and estimate the first system state value and the second system state value of the first model and the second model respectively; a first calculation unit, used to calculate the error between the system state value of the controlled object system acquired by the sensor and the first system state value and the second system state value, respectively obtain a first deviation value and a second deviation value, calculate a first error trend set based on the first deviation value, and calculate a second error trend set based on the second deviation value; a first output unit, used to calculate a first sub-control law and a second sub-control law according to the first model and the second model, and output a target control law in combination with the first deviation value and the second deviation value, and the first error trend set and the second error trend set.

[0013] According to a third aspect of the present invention, there is further provided a computer-readable storage medium, in which a computer program is stored, wherein the computer program is configured to execute the above-mentioned reference model switching method based on the model estimation error trend when running.

[0014] According to a fourth aspect of the present invention, there is also provided an electronic device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor executes the reference model switching method based on the model estimation error trend through the computer program.

[0015] In general, the above technical solutions conceived by the present invention can achieve the following beneficial effects compared with the prior art:

[0016] (1) The present invention provides a reference model switching method based on the model estimation error trend. By simultaneously calculating the estimated outputs of two reference models with continuous states, the output values ​​of the reference models are compared with the sensor acquisition values. When the error value of the reference model exceeds the set threshold, the error change trend of each model is determined by the historical data of the estimated error, and the model with the error decreasing trend is selected. When the error trend difference exceeds the set threshold, the control law calculation reference model is switched. Through the error and error trend, the model intersection is generated to achieve smooth switching of the reference model, and alleviate the equipment deterioration caused by the drastic change of the control command.

[0017] (2) By adopting a reference model switching method based on the model estimation error trend provided by the present invention, the most appropriate model is dynamically selected in a multi-model control system according to the real-time working conditions (such as the reference model error and its changing trend). This method avoids the inadaptability that may be caused by model switching based on fixed rules and enhances the adaptive ability of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0019] Figure 1 A schematic flow chart of an optional reference model switching method based on a model estimation error trend provided in an embodiment of the present application;

[0020] Figure 2 An optional controlled object model set composition diagram provided in an embodiment of the present application;

[0021] Figure 3 A schematic diagram of a flow chart of an optional model switching method provided in an embodiment of the present application;

[0022] Figure 4 An algorithm logic block diagram of an optional reference model switching method based on a model estimation error trend provided in an embodiment of the present application;

[0023] Figure 5 A schematic flow chart of an optional reference model switching method based on a model estimation error trend provided in an embodiment of the present application;

[0024] Figure 6 A schematic diagram of the structure of an optional reference model switching device based on a model estimation error trend provided in an embodiment of the present application;

[0025] Figure 7 A schematic diagram of the structure of an optional electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0026] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0027] The terms "first", "second", "third", etc. in the specification and claims of this application and the above-mentioned drawings are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally includes steps or units that are not listed, or optionally includes other steps or units inherent to these processes, methods, products or devices.

[0028] According to one aspect of an embodiment of the present application, a reference model switching method based on a model estimation error trend is provided. Figure 1 The reference model switching method based on the model estimation error trend provided in an embodiment of the present application is described.

[0029] Figure 1 is a flow chart of an optional reference model switching method based on a model estimation error trend provided in an embodiment of the present application, such as Figure 1 As shown, the process of the method may include the following steps:

[0030] S102, constructing a segmented model of the controlled object system, and collecting system state information of the controlled object system through sensors, wherein the segmented model includes an initial zero dynamic model, and the system state information includes a current working condition of the controlled object system;

[0031] S104, extracting adjacent first and second models from the model set according to the system state information, and respectively estimating first system state values ​​and second system state values ​​of the first and second models;

[0032] S106, calculating the errors between the system state value of the controlled object system acquired by the sensor and the first system state value and the second system state value, respectively obtaining a first deviation value and a second deviation value, calculating a first error trend set based on the first deviation value, and calculating a second error trend set based on the second deviation value;

[0033] S108, respectively calculating the first sub-control law and the second sub-control law according to the first model and the second model, combining the first deviation value and the second deviation value, the first error trend set and the second error trend set, and outputting the target control law.

[0034] A reference model switching method based on a model estimation error trend provided in an embodiment of the present application can be applied to scenarios where a reference model needs to be switched to achieve high-performance index control at different operating points after an operating condition changes when a multi-model reference controller is applied.

[0035] In practical applications, the trend change of model estimation error is a complex and unpredictable problem. Traditional model switching methods often rely on preset thresholds or rules to decide when to switch models, which performs poorly in the face of dynamically changing error trends. Specifically, when the estimation error of the model gradually increases to a certain extent, the traditional method may trigger the switching mechanism too early, resulting in performance degradation; conversely, if the error increases to an excessively large extent before switching, it may have a significant negative impact on the system. In addition, the adaptability and robustness of different models also vary. In some cases, a model may show high accuracy in a specific time period, but may become inaccurate in another period of time. This uncertainty makes it difficult for model switching strategies based on fixed rules to effectively cope with dynamic changes in practical applications.

[0036] With the advancement and development of computer technology, the coupling degree of controlled systems has gradually increased, and the control system has gradually risen from single-component control to giant system control with multiple subsystems coupled, which has also put forward new and higher requirements for the design method of the controller. Traditional model-based control methods cannot adapt to the dynamic changes of wide-area working conditions, and the control of situations that deviate far from the ideal value of the system model. The robustness of the adaptive control algorithm is difficult to achieve transient control performance in the full range of working conditions to meet the control requirements. In order to improve the transient performance in all working conditions, the multi-model idea came into being. By constructing model sets, control sets, and supervision rules, parameter coverage and dynamic switching in all working conditions are achieved. Here, the supervision rules directly affect the transient control performance of the model switching, which restricts the overall performance of the system.

[0037] Multi-model switching control system is a typical means to solve the engineering application of advanced control theory. When designing the controller, the control effect is generally focused on while ignoring the supervision rules of the switching process. Generally, a weighted method is used to calculate multiple controllers in the control set at the same time. Different weights are selected according to the model error, and the calculated values ​​of the sub-controllers are integrated and output. This method requires a lot of experimental calibration for the weighted method of the local controller, and the stability and convergence of the system are difficult to prove. Another way is to realize the integration of the model set and the controller by switching the multi-model controller design method, which will lead to too many models and a large amount of calculation, which is not suitable for the engineering application of the controller. In order to solve the above problems, the present invention proposes a reference model switching method based on the trend of model estimation error. By selecting two adjacent models in a local area, the control law output is calculated synchronously, and then the control law output weight is assigned to each model according to the calculation error and error trend of each model. While reducing the amount of calculation, it also ensures the smoothness of the model switching process and optimizes the transient performance of the switching process.

[0038] Through the above steps S102 to S108, by constructing a segmented model of the controlled object system, the system state information of the controlled object system is collected through sensors, wherein the segmented model includes an initial zero dynamic model, and the system state information includes the current working condition of the controlled object system; the adjacent first model and the second model are extracted from the model set according to the system state information, and the first system state value and the second system state value of the first model and the second model are estimated respectively; the error between the system state value of the controlled object system collected by the sensor and the first system state value and the second system state value is calculated to obtain the first deviation value and the second deviation value respectively, and the first error trend set is calculated based on the first deviation value, and the second error trend set is calculated based on the second deviation value; the first sub-control law and the second sub-control law are calculated according to the first model and the second model respectively, and the target control law is output by combining the first deviation value and the second deviation value, the first error trend set and the second error trend set, so as to solve the technical problem that the switching process may cause a sudden change in the control law output and a large fluctuation in the system state parameters, realize the smooth switching of the reference model, and alleviate the equipment deterioration caused by the drastic change of the control instruction.

[0039] In an exemplary embodiment, after constructing the segmented model of the controlled object system and collecting the system state information of the controlled object system through sensors, the method further includes:

[0040] S11, the supervision rule selects adjacent first models and second models from a model set including multiple fixed models at each calculation step according to the system operation parameters in the system state information;

[0041] S12, the first model is used as the current initial state model of the adaptive reference model, and the second model is used as the candidate state model of the adaptive reference model.

[0042] In this embodiment, the historical error data of the model with a finite step length is used as the basis for model switching, which effectively avoids problems such as excessive calculation of the system model set and calibration of control weights. Figure 2 An optional controlled object model set composition diagram provided in an embodiment of the present application, such as Figure 2 As shown, the nonlinear dynamic process model set of the controlled object system can include n sub-models, namely model 1 to model 5. Figure 3 A flow chart of an optional model switching method provided in an embodiment of the present application is as follows: Figure 3As shown in the figure, the multi-model reference control method includes the reference model set Ω, supervision rule l, model error e and error trend estimation δ, sub-control law calculation u', control law integrated output u and other parts. During the strategy working process, the supervision rule extracts two adjacent models from the model set Ω through the system state information collected by the sensor, and then estimates the system state values ​​of the two models respectively, calculates the error between the sensor value and the estimated value, records the error and calculates the error deviation value. The sub-control law is calculated according to two different models respectively, and then the control law is integrated and output according to the error deviation value and the error change trend.

[0043] Exemplarily, the supervisory rule selects adjacent first and second models from a model set containing multiple fixed models at each calculation step according to the system operating parameters in the system state information, that is, after constructing a segmented model of the controlled object system (the typical number of models in a ship power system is 6, including an initial zero dynamic model). At each calculation step, the supervisory rule can select two adjacent models M1,k (e.g., the first model) and M2,k (e.g., the second model) from a model set Ω containing n fixed models according to the system operating parameters (e.g., speed and load) as the current initial state and alternative state of the adaptive reference model, and then convert them into two conventional model reference adaptive control control law calculation methods.

[0044] Through this embodiment, the dynamic switching process from model i to model i+1 is optimized, and the smooth transition of the control law calculation reference model is achieved according to the design of the supervision rule and the error estimation rule, thereby ensuring the robustness and adaptability of the system.

[0045] In an exemplary embodiment, after calculating the errors between the system state value of the controlled object system acquired by the sensor and the first system state value and the second system state value to obtain the first deviation value and the second deviation value respectively, the method further includes:

[0046] S21, constructing a first initial error set based on the first deviation value, wherein the first initial error set rollingly records the errors of ten consecutive calculation steps of the first model when the algorithm starts;

[0047] S22, constructing a second initial error set based on the second deviation value, wherein the second initial error set rollingly records the errors of ten consecutive calculation steps of the second model when the algorithm starts.

[0048] In this embodiment, Figure 4 An algorithm logic block diagram of an optional reference model switching method based on a model estimation error trend provided in an embodiment of the present application, Figure 5 A flowchart of an optional reference model switching method based on a model estimation error trend provided in an embodiment of the present application, combined with Figure 4 and Figure 5 As shown, according to the conventional model reference adaptive control method, M1 (first model) and M2 (second model) can be used to calculate the control law of the current control step, namely the first sub-control law u1'(k) and the second sub-control law u2'(k). According to the selected reference models M1,k and M2,k, the estimated value of the current model output state is calculated respectively. According to the model estimation value of the kth step and the sensor acquisition value xk (system state value), the model error ek (first deviation value, second deviation value) of the kth step is calculated. Construct the initial error set {0,…,0} respectively 1x10 , rolling record the errors of the two models for 10 consecutive steps from the beginning of the algorithm, and construct the error set e M1 , e M2 , that is, are the first initial error set and the second initial error set respectively.

[0049] Through this embodiment, by selecting two reference models corresponding to adjacent operating points, it is possible to significantly reduce the computational complexity and resource consumption while ensuring the breadth of coverage of operating parameters.

[0050] In an exemplary embodiment, after constructing the second initial error set based on the second deviation value, the method further includes:

[0051] S31, calculating a first error trend set according to the first initial error set, and counting the first negative value number, the first positive value sum, and the first negative value sum in the first error trend set;

[0052] S32, calculating a second error trend set according to the second initial error set, and counting the second number of negative values, the second sum of positive values, and the second sum of negative values ​​in the second error trend set.

[0053] In this embodiment, after constructing the second initial error set based on the second deviation value, the first error trend set can be calculated based on the first initial error set, and the first negative value number, the first positive value sum, and the first negative value sum in the first error trend set can be counted. Similarly, the second error trend set can be calculated based on the second initial error set, and the second negative value number, the second positive value sum, and the second negative value sum in the second error trend set can be counted.

[0054] Specifically, Figure 4 and Figure 5 As shown, the first initial error set e can be obtained according to the error set M1 , the second initial error set e M2 , calculate the error trend set, that is, the first error trend set A, the second error trend set B, the above two error trend sets can be expressed as {δ1,…,δ9}, where δ 1 =e k -e k-1 …,δ9＝ e k-9 -e k-10 .

[0055] Furthermore, the first negative value number N1 in the first error trend set A, the second negative value number N2 in the second error trend set B, the first positive value sum Ps1 in the first error trend set A, the second positive value sum Ps2 in the second error trend set B, the negative value sum Ns1 in the first error trend set A, and Ns2 in the second error trend set B can be counted.

[0056] Through this embodiment, when the working conditions change dynamically, the control law output control can be completed according to the reference model error and the error change trend, ensuring the robustness and adaptability of the control system when the multi-model reference control switches models. It is possible to dynamically determine the most suitable model for control based on the current operating status and change trend of the system, rather than statically presetting the model switching point, which enables the control strategy to better adapt to the uncertainty of the environment or working conditions. The adaptive model switching strategy based on the error trend can smoothly transition to the new control model and reduce the impact during switching.

[0057] In an exemplary embodiment, after calculating the second error trend set according to the second initial error set and counting the second number of negative values, the second sum of positive values, and the second sum of negative values ​​in the second error trend set, the method further includes:

[0058] S41, when the first negative value is greater than the second negative value, the output target control law is the first sub-control law;

[0059] S42, when the first negative value is less than the second negative value, the output target control law is the second sub-control law;

[0060] S43, outputting the target control law when the first negative number is equal to the second negative number.

[0061] In this embodiment, combined with Figure 4 and Figure 5 As shown, after calculating the second error trend set based on the second initial error set and counting the second negative value number, the second positive value sum and the second negative value sum in the second error trend set, the ratio of the first positive value sum to the second positive value sum and the ratio of the first negative value sum to the second negative value sum can be calculated.

[0062] Specifically, the ratio of the first positive sum to the second positive sum, i.e., Ps1 / Ps2, and the ratio of the first negative sum to the second negative sum, i.e., Ns1 / Ns2, can be calculated. If Th2 > Ps1 / Ps2 > Th1, then proceed to the next calculation step, where Th1 ∈ [0.01, 0.9], Th2 ∈ [1, 10], and the specific values can be adjusted according to the output result of the control law. If Ps1 / Ps2 < Th1, then directly output u1'(k) by abandoning the calculation result of u2'(k). If Ps1 / Ps2 > Th2, then abandon the calculation result of u1'(k) and output u2'(k).

[0063] Further, in the case where the number of first negative values is greater than the number of second negative values, the output target control law is the first sub-control law, that is, if N1 > N2, u k = u 1 '(k); in the case where the number of first negative values is less than the number of second negative values, the output target control law is the second sub-control law, that is, if N1 < N2, u k = u 2 '(k); in the case where the number of first negative values is equal to the number of second negative values, output the target control law, that is, u k = W1 * u 1 '(k) + W2 * u 2 '(k), and the weight coefficients W1 and W2 will be described in detail in the following embodiments.

[0064] Through this embodiment, the control weights of each reference model are dynamically determined in the multi-model control system. By considering the numerical magnitude of the error and its change trend, it can more accurately reflect the current control effect and future prediction, and then adjust the weights of the control model, avoiding the disadvantages such as large workload of weight parameter calibration and reducing the dependence on testing and experiments.

[0065] In an exemplary embodiment, in the case where the number of first negative values is equal to the number of second negative values, the output target control law includes:

[0066] S51, determining the first weight and the second weight based on the first negative sum and the second negative sum, and calculating the output weight through the first weight and the second weight;

[0067] S52, using the output weight, combining the first sub-control law and the second sub-control law to determine the target control law.

[0068] In this embodiment, when the first negative value is equal to the second negative value, that is, when the error decays simultaneously and has the same amplitude, the output target control law can be determined in the following manner. Optionally, the first weight and the second weight can be determined based on the first negative value sum and the second negative value sum, and the output weight can be calculated using the first weight and the second weight. Using the output weight, the target control law is determined by combining the first sub-control law and the second sub-control law.

[0069] Exemplarily, in combination with Figure 4 and Figure 5 as shown, the output weights W1 (the first weight) and W2 (the second weight) can be calculated, where W1 = |Ns1| / (|Ns1| + |Ns2|), W2 = 1 - W1. Combining u k = W1 * u 1 '(k) + W2 * u 2 '(k), the specific output target control law can be obtained when the first negative value is equal to the second negative value.

[0070] Through this embodiment, in the method for calculating the weight factor of the error statistical value with continuous step sizes, based on the statistical information, the weight factor of each reference model is calculated. The smaller the error statistical value, or the more favorable the error change trend is for achieving the control target, the higher the weight factor of this model, which means that the control strategy will be more inclined to adopt the model with better performance, and based on this, more stable model switching is achieved.

[0071] In an exemplary embodiment, after determining the target control law by using the output weight and combining the first sub-control law and the second sub-control law, the above method further includes:

[0072] S61, restricting the output change rate of the target control law based on the first calibration value and the second calibration value, where the output change rate is the difference between two adjacent target control laws of the output.

[0073] In the embodiments of the present application, after determining the target control law by using the output weight and combining the first sub-control law and the second sub-control law, in order to achieve a better model switching effect, it is necessary to restrict the output change law of the target control law.

[0074] Optionally, the output change rate of the target control law can be restricted based on the first calibration value (Th3) and the second calibration value (Th4). In combination with Figure 4 and Figure 5 as shown, the output change rate of u (the difference between two adjacent target control laws of the output, that is, Δu) can be restricted. First, Δu = u k - u k-1 .

[0075] If Th4 < Δu < Th3, then output uk; otherwise, if Δu > Th3, uk =u k-1 +Th3; if Δu <Th4,u k =u k-1 +Th4, where Th3>0 is the calibration value and Th4<0 is the calibration value. Then u can be output k , and u k Assign to u k-1 , waiting for the next calculation trigger.

[0076] Through this embodiment, a limited number of adjacent reference models are selected for simultaneous calculation, which can not only ensure the breadth of the model's coverage of the operating range, but also effectively reduce the computing resources consumed by the system state estimation during the control law calculation process, and effectively ensure the real-time and rapidity of the control law calculation. The adaptive model switching strategy based on the error trend can integrate the control laws calculated by the two models based on the error value, gradually reduce the output ratio of the model with large error, and smoothly transition to the new control model, reducing the impact during switching.

[0077] According to another aspect of an embodiment of the present application, a reference model switching device for implementing the above-mentioned reference model switching method based on model estimation error trend is also provided. Figure 6 is a schematic structural diagram of an optional reference model switching device based on a model estimation error trend according to an embodiment of the present application, such as Figure 6 As shown, the device may include:

[0078] The acquisition unit 602 is used to construct a segmented model of the controlled object system, and acquire system state information of the controlled object system through sensors, wherein the segmented model includes an initial zero dynamic model, and the system state information includes a current working condition of the controlled object system;

[0079] An estimating unit 604 is configured to extract adjacent first and second models from a model set according to the system state information, and estimate a first system state value and a second system state value of the first and second models, respectively;

[0080] A first calculation unit 606 is used to calculate the errors between the system state value of the controlled object system acquired by the sensor and the first system state value and the second system state value, respectively obtain a first deviation value and a second deviation value, calculate a first error trend set based on the first deviation value, and calculate a second error trend set based on the second deviation value;

[0081] The first output unit 608 is used to calculate the first sub-control law and the second sub-control law according to the first model and the second model respectively, and output the target control law in combination with the first deviation value and the second deviation value, the first error trend set and the second error trend set.

[0082] It should be noted that the acquisition unit 602 in this embodiment can be used to execute the above step S102, the estimation unit 604 in this embodiment can be used to execute the above step S104, the first calculation unit 606 in this embodiment can be used to execute the above step S106, and the first output unit 608 in this embodiment can be used to execute the above step S108.

[0083] Through the above modules, by constructing a segmented model of the controlled object system, the system state information of the controlled object system is collected through sensors, wherein the segmented model includes an initial zero dynamic model, and the system state information includes the current working condition of the controlled object system; according to the system state information, the adjacent first model and the second model are extracted from the model set, and the first system state value and the second system state value of the first model and the second model are estimated respectively; the error between the system state value of the controlled object system collected by the sensor and the first system state value and the second system state value is calculated to obtain the first deviation value and the second deviation value respectively, and the first error trend set is calculated based on the first deviation value, and the second error trend set is calculated based on the second deviation value; the first sub-control law and the second sub-control law are calculated according to the first model and the second model respectively, and the target control law is output by combining the first deviation value and the second deviation value, the first error trend set and the second error trend set, so as to solve the technical problem that the switching process will cause a sudden change in the control law output and a large fluctuation in the system state parameters, realize the smooth switching of the reference model, and alleviate the equipment deterioration caused by the drastic change of the control instructions.

[0084] In an exemplary embodiment, the apparatus further comprises:

[0085] A selection module, configured to select adjacent first and second models from the model set including a plurality of fixed models at each calculation step according to the system operation parameters in the system state information according to the supervision rule;

[0086] The first model serves as a current initial state model of the adaptive reference model, and the second model serves as a candidate state model of the adaptive reference model.

[0087] In an exemplary embodiment, the apparatus further comprises:

[0088] A first construction unit is used to construct a first initial error set based on the first deviation value, wherein the first initial error set rollingly records the errors of ten consecutive calculation steps of the first model when the algorithm starts;

[0089] The second construction unit is used to construct a second initial error set based on the second deviation value, wherein the second initial error set rollingly records the errors of ten consecutive calculation steps of the second model when the algorithm starts.

[0090] In an exemplary embodiment, the apparatus further comprises:

[0091] A second calculation unit is used to calculate the first error trend set according to the first initial error set, and count the first negative value number, the first positive value sum, and the first negative value sum in the first error trend set;

[0092] The third calculation unit is used to calculate the second error trend set according to the second initial error set, and count the second negative value number, the second positive value sum and the second negative value sum in the second error trend set.

[0093] In an exemplary embodiment, the apparatus further comprises:

[0094] a second output unit, configured to output the target control law as the first sub-control law when the first negative value is greater than the second negative value;

[0095] a third output unit, configured to output the target control law as the second sub-control law when the first negative value is smaller than the second negative value;

[0096] A fourth output unit is used to output the target control law when the first negative value is equal to the second negative value.

[0097] In an exemplary embodiment, the fourth output unit includes:

[0098] A first determining module, configured to determine a first weight and a second weight based on the first negative value sum and the second negative value sum, and calculate an output weight by using the first weight and the second weight;

[0099] The second determination module is used to determine the target control law by using the output weight and combining the first sub-control law and the second sub-control law.

[0100] In an exemplary embodiment, the apparatus further comprises:

[0101] A limiting unit is used to limit the output change rate of the target control law based on the first calibration value and the second calibration value, wherein the output change rate is the difference between two adjacent outputs of the target control law.

[0102] It should be noted here that the examples and scenarios implemented by the above-mentioned modules and corresponding steps are the same, but are not limited to the contents disclosed in the above-mentioned embodiments. It should be noted that the above-mentioned modules as part of the device can run in a hardware environment and can be implemented by software or hardware, wherein the hardware environment includes a network environment.

[0103] According to another aspect of the embodiments of the present application, a storage medium is further provided. Optionally, in this embodiment, the storage medium can be used to execute the program code of any of the reference model switching methods based on model estimation error trend in the embodiments of the present application.

[0104] Optionally, in this embodiment, the storage medium is configured to store program codes for executing the following steps:

[0105] S1, constructing a segmented model of the controlled object system, and collecting system state information of the controlled object system through sensors, wherein the segmented model includes an initial zero dynamic model, and the system state information includes the current working condition of the controlled object system;

[0106] S2, extracting adjacent first models and second models from the model set according to the system state information, and respectively estimating first system state values ​​and second system state values ​​of the first model and the second model;

[0107] S3, calculating the errors between the system state value of the controlled object system acquired by the sensor and the first system state value and the second system state value, respectively obtaining a first deviation value and a second deviation value, calculating a first error trend set based on the first deviation value, and calculating a second error trend set based on the second deviation value;

[0108] S4, respectively calculating the first sub-control law and the second sub-control law according to the first model and the second model, combining the first deviation value and the second deviation value, the first error trend set and the second error trend set, and outputting the target control law.

[0109] Optionally, the specific examples in this embodiment may refer to the examples described in the above embodiments, which will not be described in detail in this embodiment.

[0110] Among them, computer-readable storage media may include, but are not limited to, any type of disk, including floppy disks, optical disks, DVDs, CD-ROMs, microdrives and magneto-optical disks, ROMs, RAMs, EPROMs, EEPROMs, DRAMs, VRAMs, flash memory devices, magnetic or optical cards, nanosystems (including molecular memory ICs), or any type of medium or device suitable for storing instructions and / or data.

[0111] According to another aspect of an embodiment of the present application, an electronic device for implementing the above-mentioned reference model switching method based on model estimation error trend is also provided. The electronic device may be a server, a terminal, or a combination thereof.

[0112] Figure 7 is a schematic diagram of the structure of an optional electronic device according to an embodiment of the present application, such as Figure 7As shown, it includes a processor 702, a communication interface 704, a memory 706 and a communication bus 708, wherein the processor 702, the communication interface 704, and the memory 706 communicate with each other through the communication bus 708, wherein:

[0113] Memory 706, used to store computer programs;

[0114] The processor 702 is used to implement the following steps when executing the computer program stored in the memory 706:

[0115] S1, constructing a segmented model of the controlled object system, and collecting system state information of the controlled object system through sensors, wherein the segmented model includes an initial zero dynamic model, and the system state information includes the current working condition of the controlled object system;

[0116] S2, extracting adjacent first models and second models from the model set according to the system state information, and respectively estimating first system state values ​​and second system state values ​​of the first model and the second model;

[0117] S3, calculating the errors between the system state value of the controlled object system acquired by the sensor and the first system state value and the second system state value, respectively obtaining a first deviation value and a second deviation value, calculating a first error trend set based on the first deviation value, and calculating a second error trend set based on the second deviation value;

[0118] S4, respectively calculating the first sub-control law and the second sub-control law according to the first model and the second model, combining the first deviation value and the second deviation value, the first error trend set and the second error trend set, and outputting the target control law.

[0119] Optionally, the communication bus may be a PCI (Peripheral Component Interconnect) bus, or an EISA (Extended Industry Standard Architecture) bus, etc. The communication bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 7 The communication interface is used for communication between the electronic device and other devices.

[0120] The memory may include RAM, or may include non-volatile memory, such as at least one disk memory. Optionally, the memory may also be at least one storage device located away from the aforementioned processor.

[0121] As an example, the memory 706 may include, but is not limited to, the acquisition unit 602, the estimation unit 604, the first calculation unit 606, and the first output unit 608 in the reference model switching device based on the model estimation error trend. In addition, other module units in the reference model switching device based on the model estimation error trend may also be included but are not limited to, which will not be described in detail in this example.

[0122] The above-mentioned processor can be a general-purpose processor, which can include but not be limited to: CPU (Central Processing Unit), NP (Network Processor), etc.; it can also be DSP (Digital Signal Processing), ASIC (Application Specific Integrated Circuit), FPGA (Field-Programmable Gate Array) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0123] Optionally, the specific examples in this embodiment may refer to the examples described in the above embodiments, and this embodiment will not be described in detail here.

[0124] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0125] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0126] In the several embodiments provided in the present application, it should be understood that the disclosed devices can be implemented in other ways. For example, the device embodiments described above are only schematic, such as the division of the units, which is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some service interfaces, and the indirect coupling or communication connection of devices or units can be electrical or other forms.

[0127] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0128] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0129] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a memory, including a number of instructions to enable a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned memory includes: U disk, read-only memory (ROM), random access memory (RAM), mobile hard disk, disk or optical disk and other media that can store program code.

[0130] A person of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable memory, which can include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0131] The above is only an exemplary embodiment of the present disclosure, and the scope of the present disclosure cannot be limited thereto. That is, any equivalent changes and modifications made according to the teachings of the present disclosure are still within the scope of the present disclosure. After considering the specification and practicing the disclosure here, those skilled in the art will easily think of the implementation scheme of the present disclosure. This application is intended to cover any modification, use or adaptation of the present disclosure, which follows the general principles of the present disclosure and includes common knowledge or customary technical means in the technical field not recorded in the present disclosure. The description and examples are regarded as exemplary only, and the scope and spirit of the present disclosure are defined by the claims.

[0132] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0133] It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A reference model switching method based on model estimation error trend, characterized in that: include: Constructing a segmented model of a controlled object system, and collecting system state information of the controlled object system through a sensor, wherein the segmented model includes an initial zero dynamic model, and the system state information includes a current working condition of the controlled object system; Extracting adjacent first and second models from a model set according to the system state information, and respectively estimating first system state values ​​and second system state values ​​of the first and second models; Calculate the errors between the system state value of the controlled object system acquired by the sensor and the first system state value and the second system state value to obtain a first deviation value and a second deviation value respectively, calculate a first error trend set based on the first deviation value, and calculate a second error trend set based on the second deviation value; A first sub-control law and a second sub-control law are calculated according to the first model and the second model respectively, and a target control law is outputted by combining the first deviation value and the second deviation value and the first error trend set and the second error trend set.

2. The reference model switching method based on model estimation error trend according to claim 1, characterized in that: After constructing the segmented model of the controlled object system and collecting the system state information of the controlled object system through sensors, the method further includes: The supervision rule selects adjacent first model and second model from the model set including multiple fixed models at each calculation step according to the system operation parameters in the system state information; The first model serves as a current initial state model of the adaptive reference model, and the second model serves as a candidate state model of the adaptive reference model.

3. The reference model switching method based on model estimation error trend according to claim 1, characterized in that: After respectively obtaining a first deviation value and a second deviation value by calculating the error between the system state value of the controlled object system acquired by the calculation sensor and the first system state value and the second system state value, the method further includes: Constructing a first initial error set based on the first deviation value, wherein the first initial error set rollingly records the errors of ten consecutive calculation steps of the first model when the algorithm starts; A second initial error set is constructed based on the second deviation value, wherein the second initial error set rollingly records the errors of ten consecutive calculation steps of the second model when the algorithm starts.

4. The reference model switching method based on model estimation error trend according to claim 3, characterized in that: After constructing a second initial error set based on the second deviation value, the method further includes: Calculate the first error trend set according to the first initial error set, and count the first negative value number, the first positive value sum, and the first negative value sum in the first error trend set; The second error trend set is calculated according to the second initial error set, and the second number of negative values, the second sum of positive values, and the second sum of negative values ​​in the second error trend set are counted.

5. The reference model switching method based on model estimation error trend according to claim 4, characterized in that: After calculating the second error trend set according to the second initial error set and counting the second number of negative values, the second sum of positive values, and the second sum of negative values ​​in the second error trend set, the method further includes: When the first negative value is greater than the second negative value, the output target control law is the first sub-control law; When the first negative value is smaller than the second negative value, the output target control law is the second sub-control law; When the first negative number is equal to the second negative number, the target control law is output.

6. The reference model switching method based on model estimation error trend according to claim 5, characterized in that: When the first negative value is equal to the second negative value, outputting the target control law comprises: Determine a first weight and a second weight based on the first negative value sum and the second negative value sum, and calculate an output weight by using the first weight and the second weight; The target control law is determined by utilizing the output weight and combining the first sub-control law with the second sub-control law.

7. The reference model switching method based on the model estimation error trend according to any one of claims 1 to 6, characterized in that: After determining the target control law by using the output weight and combining the first sub-control law with the second sub-control law, the method further includes: The output change rate of the target control law is limited based on the first calibration value and the second calibration value, wherein the output change rate is the difference between two adjacent outputs of the target control law.

8. A reference model switching device based on model estimation error trend, characterized in that: include: A collection unit, used to construct a segmented model of a controlled object system, and collect system state information of the controlled object system through a sensor, wherein the segmented model includes an initial zero dynamic model, and the system state information includes a current working condition of the controlled object system; an estimating unit, configured to extract adjacent first and second models from a model set according to the system state information, and estimate a first system state value and a second system state value of the first model and the second model respectively; a first calculation unit, configured to calculate errors between a system state value of the controlled object system acquired by a sensor and the first system state value and the second system state value, respectively obtaining a first deviation value and a second deviation value, calculating a first error trend set based on the first deviation value, and calculating a second error trend set based on the second deviation value; The first output unit is used to calculate the first sub-control law and the second sub-control law according to the first model and the second model respectively, and output the target control law in combination with the first deviation value and the second deviation value, the first error trend set and the second error trend set.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein the program executes the method according to any one of claims 1 to 7 when executed.

10. An electronic device comprising a memory and a processor, characterized in that: A computer program is stored in the memory, and the processor is configured to execute the method according to any one of claims 1 to 7 through the computer program.

Images