A method and system for synergistic optimization of reactive power compensation and harmonic suppression of a refining furnace transformer

Abstract

The present application belongs to the technical field of power quality control, and particularly relates to a refining furnace transformer reactive power compensation and harmonic suppression collaborative optimization method and system, comprising the following steps: S1, real-time signal acquisition and feature extraction: real-time acquisition of voltage signals and current signals in a refining furnace transformer power supply system, decomposition of the current signals, extraction of the root mean square value of the fundamental reactive current and characteristic harmonics, and real-time calculation of the instantaneous arc impedance; S2, robust evaluation of transient arc dispersion: based on the pre-constructed reference impedance characteristic sequence and the real-time acquired arc impedance transient sequence. The present application solves the problem of excessive sensitivity of the index to noise in the traditional method, guarantees the reliability and stability of the dynamic weight distribution, and completely avoids the compensation system oscillation and misjudgment caused by noise.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of power quality control, and particularly relates to a method and system for synergistic optimization of reactive power compensation and harmonic suppression of a refining furnace transformer. BACKGROUND

[0002] As a key equipment in modern steel short process steelmaking, the refining furnace mainly performs refining, heating and desulfurization on molten steel through arc heating. However, the arc combustion process of the refining furnace is a typical nonlinear and highly random load, and its operating characteristics will cause huge reactive power fluctuations and serious characteristic harmonic pollution on the power grid side, which will have a serious impact on the power quality and system safety of the power grid. Therefore, in the power supply system of the refining furnace, high-performance reactive power compensation and harmonic suppression devices must be configured to ensure the efficient and stable operation of the system.

[0003] In the prior art, in order to cope with the rapid and severe changes of the refining furnace load, some advanced reactive power compensation control strategies attempt to use a dynamic optimization method based on the least square method. These methods usually introduce real-time indicators such as transient dispersion of arc impedance, and dynamically adjust the weight factors of reactive power compensation and / or harmonic suppression based on this. The purpose of this dynamic weight distribution is to enable the response speed of the compensation command to follow the severe changes of the arc in real time and accurately, so as to reduce the lag effect of compensation.

[0004] However, in actual smelting conditions, the nonlinear transient fluctuations of the arc itself are extremely complex, and the current and voltage signals collected in real time will inevitably be mixed with a large amount of industrial high-frequency noise, transient spike pulses and data acquisition errors. When these original, contaminated signals are directly used to calculate the arc impedance dispersion, the evaluation result shows extreme sensitivity to noise. This sensitivity will cause the dispersion index evaluated to jump irregularly and unrealistically, and thus distort the subsequent reactive power compensation weight and harmonic suppression weight used for dynamic control. Ultimately, this will cause oscillation of the compensation current command or misjudgment of the wrong signal, seriously affecting the stable operation of the compensation system and making it difficult to ensure the accuracy and robustness of power quality control. Therefore, there is an urgent need in the industry for a mechanism that is robust to noise and anti-interference to accurately evaluate the true fluctuation degree of the transient arc and fundamentally solve the problems of compensation system oscillation and misjudgment caused by evaluation deviation. SUMMARY

[0005] The present application provides a method and system for synergistic optimization of reactive power compensation and harmonic suppression of a refining furnace transformer to solve the technical problems of excessive sensitivity of transient arc evaluation results to noise, distortion of dynamic weight factors, and compensation current command oscillation or misjudgment caused thereby in the prior art.

[0006] In a first aspect, the present application provides a method for synergistic optimization of reactive power compensation and harmonic suppression of a refinery furnace transformer, comprising the following steps:

[0007] S1, real-time signal acquisition and feature extraction: real-time acquisition of voltage signals and current signals in the power supply system of the refinery furnace transformer, decomposition of the current signals, extraction of the root mean square value of the fundamental reactive current and characteristic harmonics, and real-time calculation of the instantaneous arc impedance;

[0008] S2, robust evaluation of transient arc dispersion: based on the pre-constructed reference impedance feature sequence and the real-time acquired arc impedance transient sequence, the dynamic time warping distance is used to evaluate the similarity of arc fluctuation, and combined with the energy entropy of the real-time arc impedance transient sequence and the preset threshold, the energy entropy correction factor is generated, the dynamic time warping distance is multiplied by the energy entropy correction factor, and the robust arc dispersion index is constructed;

[0009] S3, dynamic weight allocation and synergistic constraint: based on the robust arc dispersion index, the reactive power compensation weight and the harmonic suppression weight are dynamically generated, and based on the reactive power compensation weight, the harmonic suppression weight and the preset dynamic response ratio, the synergistic constraint factor is generated;

[0010] S4, dynamic least squares optimization and compensation instruction output: the reactive power compensation weight, the harmonic suppression weight and the synergistic constraint factor are embedded into the objective function of the least squares method, real-time optimization is solved, and the compensation current instruction is generated to realize the synergistic optimization of reactive power compensation and harmonic suppression.

[0011] Further, in S2, the energy entropy correction factor is calculated by the following formula:

[0012]

[0013] wherein, is the energy entropy of the real-time arc impedance transient sequence, is the preset energy entropy reference threshold, is the exponential correction coefficient, is the natural exponential function.

[0014] Further, the construction formula of the robust arc dispersion index is:

[0015]

[0016] wherein, is the dynamic time warping distance, is the energy entropy correction factor.

[0017] Further, in S3, the calculation formula of the synergistic constraint factor is:

[0018]

[0019] wherein, is a reactive compensation weight, is a harmonic suppression weight, is a preset dynamic response ratio, is an allowed weight difference tolerance criterion, is an exponential correction coefficient, is a natural exponential function.

[0020] Further, in S3, the dynamic weight generation formula is:

[0021]

[0022] wherein, is a reactive compensation weight, is a harmonic suppression weight, and are base multipliers of the reactive compensation weight and the harmonic suppression weight, respectively, is a robust arc dispersion index.

[0023] Further, in S4, the target function is constructed according to the formula:

[0024]

[0025] wherein, is a compensation current instruction, is a cooperative constraint factor, is a reactive compensation weight, is a fundamental reactive current, is a reactive component of the compensation current instruction, is a harmonic suppression weight, is a root mean square value of a characteristic harmonic, is a harmonic component of the compensation current instruction.

[0026] Further, in S1, the decomposition of the collected current signal includes a fast Fourier transform or a decomposition based on a digital phase-locked loop on the current signal .

[0027] Further, in S2, the construction of the reference impedance characteristic sequence includes: when the refining furnace is in a stable working condition, a preset short time window is selected, a plurality of groups of time series data of arc impedance are collected, and the reference impedance characteristic sequence is constructed through averaging and filtering.

[0028] Further, in S4, the optimal compensation current is obtained by real-time minimization of the objective function, and the compensation current is output to the compensation device to perform compensation.

[0029] In a second aspect, the present application provides a refining furnace transformer reactive power compensation and harmonic suppression collaborative optimization system, comprising a memory and a processor, and the memory stores computer program instructions, which realize the above-mentioned refining furnace transformer reactive power compensation and harmonic suppression collaborative optimization method when executed by the processor.

[0030] The beneficial effect is: the present application innovatively proposes an evaluation mechanism of the robust arc dispersion index, which successfully isolates the arc fluctuation evaluation process from the interference of high-frequency noise and sharp pulses by combining dynamic time warping similarity evaluation and energy entropy correction factor. This significantly enhances the anti-interference ability of transient arc evaluation, solves the problem of excessive sensitivity of the index to noise in traditional methods, ensures the reliability and smoothness of dynamic weight distribution, and completely avoids the oscillation and misjudgment of the compensation system caused by noise.

[0031] At the same time, by introducing a collaborative constraint factor, the present application realizes real-time constraint of the dynamic response ratio of the reactive power compensation weight and the harmonic suppression weight, and forces the two control loops to maintain a collaborative relationship when the arc fluctuates violently. This eliminates the mutual interference of the reactive power and harmonic control loops in transient state, so that the entire power quality control system achieves a higher level of collaborative optimization, and significantly reduces the transient control error. BRIEF DESCRIPTION OF DRAWINGS

[0032] Figure 1 The flowchart of the present application.

[0033] Figure 2 The schematic diagram of the real-time arc impedance transient sequence of the refining furnace of the present application.

[0034] Figure 3 The schematic diagram of the robustness comparison of the arc fluctuation dispersion evaluation of the present application.

[0035] Figure 4 The schematic diagram of the absolute error comparison of the compensation instruction under transient noise of the present application. DETAILED DESCRIPTION

[0036] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0037] The embodiment of the refining furnace transformer reactive power compensation and harmonic suppression collaborative optimization method provided by the application comprises the following steps:

[0038] As shown in the refining furnace transformer reactive power compensation and harmonic suppression collaborative optimization method, the method comprises the following steps: Figure 1

[0039] S1, real-time signal acquisition and feature extraction.

[0040] This step is mainly responsible for acquiring necessary power quality parameters through a high-speed acquisition system to provide original data input for subsequent optimization control.

[0041] Through a high-speed data acquisition system, three-phase voltage signals and current signals in the refining furnace transformer power supply system are synchronously acquired in real time. For example, a high-speed acquisition card with a sampling rate of kHz can be used to synchronously acquire three-phase voltage and current to ensure that transient changes and high-frequency harmonics of the electric arc can be captured. The acquired current signals are decomposed to extract the root mean square values of the fundamental reactive current and the 3rd, 5th, 7th, etc. characteristic harmonics . Current decomposition can be realized by fast Fourier transform or by a digital phase-locked loop method for real-time fundamental and harmonic separation. For example, the digital phase-locked loop technology is used to track the grid fundamental frequency in real time, and the current signal is decomposed into fundamental current and non-fundamental current (i.e. harmonic) components. Among them, represents the main source of reactive impulse, represents the harmonic pollution amount.

[0042] On this basis, the instantaneous arc impedance is calculated in real time. The instantaneous arc impedance can be approximately obtained by the ratio of the real-time voltage signal and the current signal , which serves as the original input of the arc load characteristics. For example, within a ms window, the time sequence data of the instantaneous arc impedance is calculated, and the sequence reflects the original characteristics of the current arc load transient change.

[0043] By real-time acquisition and decomposition of the current signal, the characteristic values of the reactive impulse and the harmonic pollution amount can be accurately extracted, and the instantaneous arc impedance is calculated, which provides necessary power quality parameters for subsequent robust evaluation and optimization control.

[0044] S2, robust evaluation of transient arc dispersion.

[0045] ​This step aims to address the problem of excessive sensitivity to noise in traditional arc dispersion indices, to robustly assess the true degree of arc fluctuation, and to construct a robust arc dispersion index that is highly immune to noise. .

[0046] First, a reference impedance characteristic sequence is constructed. When the refining furnace is under standard and stable operating conditions, for example, during the middle stage of refining when the arc is relatively stable, a preset short time window is selected. ,For example ms, collect multiple sets of arc impedance The time-series data is used to construct a reference impedance characteristic sequence through averaging and filtering. This reference impedance characteristic sequence represents the typical fluctuation pattern of the electric arc under normal, acceptable operating conditions. In real-time smelting processes, every control cycle, for example... ms, the current time window is extracted from the instantaneous arc impedance. Arc impedance transient sequence .

[0047] Next, short-time series similarity assessment is performed, using dynamic time-normalized distance. This is used to quantify the similarity between current arc fluctuations and stable operating conditions. The dynamic time warping distance can tolerate minor misalignments on the time axis, accurately capturing the morphological similarity of sequences. The higher the value, the lower the similarity and the more the arc fluctuation deviates from the normal pattern.

[0048] Then, energy entropy correction is performed to identify whether high-frequency noise and spike pulses exist in the transient sequence of arc impedance. Energy analysis is performed on the transient sequence of arc impedance to calculate its energy entropy. High-frequency noise and pulses can cause signal energy to disperse in the frequency domain, thereby increasing the energy entropy value.

[0049] Constructing an energy entropy correction factor Used to punish those caused by noise rather than actual arc fluctuations. Increased assessment results, energy entropy correction factor The calculation formula is as follows:

[0050]

[0051] in, The energy entropy of the real-time transient sequence of arc impedance; For example, a preset energy entropy benchmark threshold. , representing the average energy entropy under standard operating conditions; For example, the exponential correction factor. , It is a natural exponential function.

[0052] Compute example: Assume preset energy entropy baseline threshold , ;

[0053] Case 1, signal pure: Real-time energy entropy ;

[0054] Then .

[0055] Case 2, heavily contaminated by noise: Real-time energy entropy ;

[0056] Then .

[0057] When the signal is pure value is high, i.e. , while when heavily contaminated by noise value is suppressed, i.e. .

[0058] Finally, construct the robust arc dispersion index : .

[0059] Compute example:

[0060] Scenario A, true drastic fluctuation: , , ;

[0061] Then .

[0062] Scenario B, false fluctuation caused by noise: , , ;

[0063] Then .

[0064] In scenario B, although the original evaluation shows drastic fluctuation, due to the penalty of , the final value is significantly suppressed. This ensures that only when value is high (i.e. arc fluctuation is drastic) and value is high (i.e. signal is pure), the can get a high value, thus avoiding the system from generating excessive compensation instructions due to noise.

[0065] By introducing dynamic time warping and energy entropy correction, a robust arc dispersion index with strong immunity to noise is successfully constructed, which can accurately evaluate the real transient fluctuations of arc and effectively enhance the anti-interference ability of arc transient evaluation.

[0066] S3, dynamic weight allocation and collaborative constraint.

[0067] This step dynamically generates reactive compensation weight and harmonic suppression weight based on the robust arc dispersion index obtained in step S2, and introduces a collaborative constraint factor to ensure the dynamic response collaboration of the two.

[0068] Dynamic weight generation:

[0069] Based on the robust arc dispersion index , the reactive compensation weight and the harmonic suppression weight are dynamically generated:

[0070]

[0071] wherein and are the basic multipliers of reactive and harmonic, used to set the reference ratio between the two. For example, it can be set as , to ensure that the reactive compensation is higher than the harmonic suppression in response priority. Since has robustness, the generated and can smoothly and reliably reflect the real arc fluctuation degree.

[0072] Generation of collaborative constraint factor:

[0073] To ensure the collaboration of reactive and harmonic in dynamic response, a preset dynamic response ratio , for example is introduced as the reference of collaborative constraint. The collaborative constraint factor is constructed to intervene in punishment when the dynamic response ratio of is too large, and to constrain the dynamic response speed of the two. The calculation formula of is as follows:

[0074]

[0075] wherein is the preset dynamic response ratio; is the allowed weight difference tolerance reference, exemplarily as ; is the exponential correction coefficient, for example , is the natural exponential function.

[0076] Calculation example: suppose , , ;

[0077] Case 1, response synergy: , At this time ;

[0078] Then .

[0079] Case 2, response unsynergy: , At this time ;

[0080] Then .

[0081] When the response is unsynergy, The value is lower than that when the response is synergy. This indicates that when The dynamic response ratio of The greater , The closer to , thus imposing a stronger penalty on the subsequent least squares objective function.

[0082] Through the robust arc dispersion index, the reactive power compensation weight and the harmonic suppression weight are dynamically generated, and the dynamic response ratio of the two is constrained in real time by the synergy constraint factor, forcing the two to maintain a synergistic relationship when the arc fluctuates violently, effectively achieving dynamic synergistic regulation.

[0083] S4, dynamic least squares optimization and compensation instruction output.

[0084] This step embeds the reactive power compensation weight and the harmonic suppression weight and the synergy constraint factor into the objective function of the least squares method, and optimizes and solves in real time to generate the final compensation current instruction .

[0085] The objective function is constructed:

[0086] The goal is to find the optimal compensation current , which minimizes the objective function :

[0087]

[0088] Where, and These are the total compensation currents. The reactive component and harmonic component; This is the fundamental reactive current; is the root mean square value of the characteristic harmonic; and These are dynamic weighting factors; As a collaborative constraint factor, in this embodiment, the objective function is a weighted sum of squared errors model.

[0089] When the dynamic response coordination of reactive power and harmonics is good, , At this point, the first weight of the objective function... Higher, second term weight The system will primarily optimize reactive power compensation due to the lower reactive power level. The error.

[0090] When dynamic response coordination is poor , At this point, the first weight of the objective function... Lower, second term weight If the noise level is high, the system will prioritize harmonic suppression in its optimization. The error is minimized, thus forcing both to maintain a coordinated ratio in the dynamic response. This mechanism ensures that when the risk of coordination increases, the system can self-regulate and adjust by optimizing the weight allocation of the objective function.

[0091] Real-time optimization solution: The optimal compensation current is obtained by minimizing the objective function in real time. Due to the extremely short control cycle of the refining furnace, this optimization should be implemented in a digital signal processor or field-programmable gate array using a fast iterative algorithm to meet real-time requirements.

[0092] Compensation command output: The optimal compensation current is output to the compensation device to perform compensation and complete the dynamic response. Due to the robustness and smoothness of the robust arc dispersion index, the compensation command can smoothly and in real time follow the arc fluctuations of the refining furnace, effectively avoiding oscillations and misjudgments, and ensuring the efficient and stable operation of the entire system.

[0093] The effectiveness of this solution can be clearly demonstrated through three data charts, please refer to the attached images for details:

[0094] Appendix Figure 2 The refining furnace was demonstrated. The arc impedance value within the s-time series, the curve contains several real stage fluctuations and several transient noise interference areas marked by shaded areas.

[0095] Appendix Figure 3The robustness comparison result of arc fluctuation dispersion evaluation is shown, and two normalized curves are compared: the traditional dispersion index and the robust dispersion index, in the transient noise interference area, the traditional dispersion index curve will appear unreasonable and sharp spikes, and the robust dispersion index curve of the present application can remain extremely smooth, only a low amplitude response to the true phase fluctuation, which proves that the energy entropy correction factor successfully suppresses the interference of noise.

[0096] The present application is characterized in that Figure 4 The absolute error comparison result of compensation instruction under transient noise is shown, and two curves are compared: the compensation error of the traditional method and the compensation error of the present application, in all non-noise areas, the fluctuation amplitude and baseline level of the compensation error curve of the present application are always significantly lower than those of the traditional method, which proves the effectiveness of the present application. In the transient noise interference area, the compensation error of the traditional method will produce a huge peak due to the false response to noise, while the compensation error of the present application can still maintain a small fluctuation near the low baseline, and even does not produce a peak at all, which directly proves that the robust evaluation is transformed into the improvement of actual control effect, and successfully solves the technical problem.

[0097] The embodiment of the refining furnace transformer reactive power compensation and harmonic suppression collaborative optimization system provided by the present application comprises a processor and a memory, and the memory stores computer program instructions, when the computer program instructions are executed by the processor, the above-mentioned refining furnace transformer reactive power compensation and harmonic suppression collaborative optimization method is realized.

[0098] The refining furnace transformer reactive power compensation and harmonic suppression collaborative optimization system further comprises a communication interface and other components familiar to those skilled in the art, the settings and functions of which are known in the art, and therefore will not be described here.

[0099] The refining furnace transformer reactive power compensation and harmonic suppression collaborative optimization system further comprises a communication interface and other components familiar to those skilled in the art, the settings and functions of which are known in the art, and therefore will not be described here.

[0100] In this disclosure, a "storage medium" or "computer readable medium" can be any available medium that can be accessed by a general purpose or special purpose computer system. By way of example, and not limitation, such computer readable media can comprise RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to carry or store desired program code means in the form of computer readable instructions or data structures and that can be accessed by a general purpose or special purpose computer system, or a combination thereof. When information is transferred or provided over a network or another communications connection (either hardwired, wireless, or combination thereof) to a computer, the computer properly views the connection as a computer readable medium. Thus, any such connection is properly termed a computer readable medium. Combinations of the above should also be included within the scope of the computer readable media.

[0101] The above are preferred embodiments of the present application, not to limit the protection scope of the present application, therefore: any equivalent changes made according to the structure, shape, principle of the present application should be covered within the protection scope of the present application.

Claims

1. A method for synergistic optimization of reactive power compensation and harmonic suppression in a refining furnace transformer, characterized in that, Includes the following steps: S1, Real-time signal acquisition and feature extraction: Real-time acquisition of voltage and current signals in the power supply system of the refining furnace transformer, decomposition of current signals, extraction of root mean square values ​​of fundamental reactive current and characteristic harmonics, and real-time calculation of instantaneous arc impedance. S2, Robust evaluation of transient arc dispersion: Based on the pre-constructed reference impedance characteristic sequence and the real-time acquired transient arc impedance sequence, the similarity of arc fluctuation is evaluated by dynamic time warping distance. Combined with the energy entropy of the real-time transient arc impedance sequence and the preset threshold, an energy entropy correction factor is generated. The dynamic time warping distance is multiplied by the energy entropy correction factor to construct a robust arc dispersion index. S3, Dynamic Weight Allocation and Cooperative Constraints: Based on the robust arc dispersion index, reactive power compensation weights and harmonic suppression weights are dynamically generated, and cooperative constraint factors are generated based on the reactive power compensation weights, harmonic suppression weights, and preset dynamic response ratios. S4, Dynamic Least Squares Optimization and Compensation Command Output: The reactive power compensation weight, harmonic suppression weight, and collaborative constraint factor are embedded into the objective function of the least squares method for real-time optimization and solution, generating compensation current commands to achieve collaborative optimization of reactive power compensation and harmonic suppression.

2. The method for coordinated optimization of reactive power compensation and harmonic suppression in refining furnace transformers according to claim 1, characterized in that, In S2, the energy entropy correction factor The calculation formula is: in, The energy entropy of the real-time transient sequence of arc impedance. The preset energy entropy benchmark threshold, This is the exponential correction factor. It is a natural exponential function.

3. The method for coordinated optimization of reactive power compensation and harmonic suppression in refining furnace transformers according to claim 1, characterized in that, Robust arc dispersion index The construction formula is: in, For dynamic time-normalized distance, This is the energy entropy correction factor.

4. The method for coordinated optimization of reactive power compensation and harmonic suppression in refining furnace transformers according to claim 1, characterized in that, In S3, the collaborative constraint factor The calculation formula is: in, For reactive power compensation weight, For harmonic suppression weights, The preset dynamic response ratio, As a baseline for allowable weight difference tolerance, This is the exponential correction factor. It is a natural exponential function.

5. The method for coordinated optimization of reactive power compensation and harmonic suppression in refining furnace transformers according to claim 1, characterized in that, In S3, the dynamic weight generation formula is: in, For reactive power compensation weight, For harmonic suppression weights, and These are the base multipliers for reactive power compensation weight and harmonic suppression weight, respectively. It is a robust arc dispersion index.

6. The method for coordinated optimization of reactive power compensation and harmonic suppression in refining furnace transformers according to claim 1, characterized in that, In S4, the objective function The construction formula is: in, To compensate for current commands, As a collaborative constraint factor, For reactive power compensation weight, For fundamental reactive current, To compensate for the reactive component of the current command, For harmonic suppression weights, The root mean square value of the characteristic harmonic. To compensate for the harmonic components of the current command.

7. The method for coordinated optimization of reactive power compensation and harmonic suppression in refining furnace transformers according to claim 1, characterized in that, In S1, the acquired current signal Decomposition includes the analysis of current signals. Perform Fast Fourier Transform or decomposition based on digital phase-locked loop.

8. The method for coordinated optimization of reactive power compensation and harmonic suppression in refining furnace transformers according to claim 1, characterized in that, In S2, the construction of the reference impedance characteristic sequence includes: when the refining furnace is in a stable operating condition, selecting a preset short time window, collecting multiple sets of time-series data of arc impedance, and constructing the reference impedance characteristic sequence through averaging and filtering.

9. The method for coordinated optimization of reactive power compensation and harmonic suppression in refining furnace transformers according to claim 1, characterized in that, In S4, the optimal compensation current is obtained by minimizing the objective function in real time, and the compensation current is output to the compensation device to perform compensation.

10. A collaborative optimization system for reactive power compensation and harmonic suppression in a refining furnace transformer, characterized in that, It includes a memory and a processor, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the refining furnace transformer reactive power compensation and harmonic suppression synergistic optimization method according to any one of claims 1-9 is implemented.

Images