Spherical mirror multi-working condition surface shape stability measurement method and system

Abstract

The application provides a spherical mirror multi-working-condition surface shape stability measurement method and system, relates to the optical testing technical field, and includes the following steps: constructing a double-layer state machine, monitoring a support contact state and an air path state in real time, generating corresponding state identifiers based on load sensing data and optical path monitoring data, determining topology annealing sequence parameters by using the state identifiers and performing topology annealing, and finally performing interference measurement in a stable state and generating stability evaluation index data. The application overcomes the problem of measurement uncertainty caused by the coupling of support topology micro-slippage and air path disturbance in the prior art, and improves the accuracy and reliability of surface shape stability evaluation of a large-aperture spherical mirror under multiple working conditions.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of optical testing, in particular to a spherical mirror multi-working-condition surface shape stability measurement method and system. BACKGROUND

[0002] In the practical application of a large-aperture spherical mirror, the surface shape stability of the mirror surface is usually evaluated under different attitudes, temperature and humidity conditions or after long-term standing, so as to ensure that the spherical mirror can stably meet the application requirements of a high-precision optical system. The current mainstream method is usually to predict the surface shape change under typical working conditions through finite element simulation, and then to measure several discrete working conditions by means of an interferometer, so as to determine the PV value or RMS value of the spherical mirror. However, this static and discrete evaluation method is only suitable for a single or a few specific working conditions, and does not consider the contact state uncertainty caused by the frictional viscosity, micro-slip and locking sequence difference in the support structure of the spherical mirror system with attitude adjustment, locking mechanism or long-term operation. This state uncertainty often leads to difficulty in reproducing or explaining the evaluation results, thereby seriously affecting the stability and reliability of the optical system in actual operation.

[0003] In addition, the current evaluation method also often ignores the influence of air disturbance existing in the long-path optical path during the interference measurement on the measurement results. Especially under the measurement conditions of a large-aperture mirror, the air disturbance path reconstruction problem caused by temperature gradient, air convection or attitude change makes the air disturbance component other than the mirror surface itself mixed into the measurement wavefront, further aggravating the uncertainty and non-repeatability of the measurement. The existing technology does not fully distinguish, analyze and eliminate the influence of such air disturbance on the measurement accuracy, and does not consider the possible cross-coupling effect between the support structure state and the air disturbance state from the mechanism. Therefore, how to evaluate the surface shape stability of the spherical mirror under multiple working conditions, overcome the measurement uncertainty problem caused by the support topology path dependence effect and air path disturbance coupling, and more accurately determine the stable measurement state to obtain accurate and reliable multi-working-condition surface shape stability evaluation indexes has become a key technical problem to be solved in the field. SUMMARY

[0004] In view of the deficiencies of the prior art, the present application provides a spherical mirror multi-working-condition surface shape stability measurement method and system.

[0005] In a first aspect, the present application provides a spherical mirror multi-working-condition surface shape stability measurement method, which comprises:

[0006] a double-layer state machine is constructed, the double-layer state machine comprising a first state machine for representing the support contact state of the spherical mirror and a second state machine for representing the air path state;

[0007] collecting load sensing data related to support contact states and optical path monitoring data related to air path states;

[0008] generating a first state identifier representing support contact states based on the load sensing data, generating a second state identifier representing air path states based on the optical path monitoring data, determining topology annealing sequence parameters based on the first state identifier and the second state identifier, and executing a topology annealing sequence;

[0009] updating the first state identifier and the second state identifier in real time during execution of the topology annealing sequence, and generating a starting state confirmation signal when both satisfy their respective preset stability conditions;

[0010] triggering formal interferometric measurement based on the starting state confirmation signal, obtaining spherical mirror surface data under current working conditions, and generating evaluation index data representing the spherical mirror multi-working condition surface stability based on the surface data and the corresponding first state identifier and second state identifier.

[0011] Optionally, the double-layer state machine includes:

[0012] determining contact micro-state characteristics of each support point based on output data of multiple load sensors of the multi-point support mechanism, performing clustering operation on the contact micro-state characteristics, generating state identifiers of each support point, and constituting a state mapping table of the first state machine;

[0013] determining disturbance characteristics of the air path based on output data of the optical path peripheral monitoring channel of the interferometric measurement assembly and the environmental monitoring sensor, performing classification operation on the disturbance characteristics, and generating a state mapping table of the second state machine;

[0014] establishing cross-state indexes between the first state machine and the second state machine, for identifying coupling mode description quantities between support contact states and air path states.

[0015] Optionally, the contact micro-state characteristics are feature vectors composed of load fluctuation quantities of the load sensors and support point displacement response quantities within a continuous time window;

[0016] The disturbance characteristics are modal coefficient fluctuation rates of the air path disturbance quantities projected to preset low-order spatial modal basis within a continuous time window;

[0017] The state mapping tables of the first state machine and the second state machine respectively record state switching sequences, and generate path history codes, and the cross-state indexes are established by associating the path history codes of the first state machine and the second state machine.

[0018] Optionally, the generating the first state identifier representing the support contact state based on the load sensing data comprises:

[0019] calculating a load average, a load fluctuation amplitude and a load change rate of each support point within a preset time window, and combining the three quantities as a load feature vector of the support point;

[0020] performing neighborhood normalization processing on the load feature vectors according to the spatial distribution of the support points to obtain a relative load feature vector eliminating the overall temperature drift effect, and calculating a deviation from a historical baseline in the relative load feature vector;

[0021] inputting the deviation into a micro-state classification model based on support point partition, and combining displacement response data or micro-vibration response data of the support point in the model to generate a state label of a bearing state, a touch state or an away state;

[0022] combining the state labels of all support points to form a first state identifier, and appending a time sequence code of state switching of each support point in the first state identifier to represent the path-dependent characteristics of the support contact topology.

[0023] Optionally, the generating the second state identifier representing the air path state based on the optical path monitoring data comprises:

[0024] performing attitude normalization processing on the air disturbance data collected by the optical path peripheral monitoring channel of the interferometric measurement assembly to obtain normalized air disturbance data;

[0025] determining a disturbance feature quantity based on the normalized air disturbance data, and inputting the disturbance feature quantity into an air path classification model to generate a state label of the air path;

[0026] combining the state label of the air path to form a second state identifier.

[0027] Optionally, the attitude normalization processing comprises performing optical path geometry remapping operation on the air disturbance data based on the optical path geometry offset caused by the change in pitch or azimuth of the spherical mirror to eliminate the geometric effect of attitude change on air disturbance measurement;

[0028] the determining the disturbance feature quantity comprises projecting the normalized air disturbance data to a preset low-order spatial modal basis, calculating a fluctuation rate and a phase consistency index of each modal coefficient within a preset time window, and combining the fluctuation rate and the phase consistency index as the disturbance feature quantity;

[0029] the second state identifier further comprises a time sequence code of air path state switching to represent the historical dependence characteristics of air path disturbance.

[0030] Optionally, the determining the topology annealing sequence parameters based on the first state identifier and the second state identifier, and performing the topology annealing sequence comprises:

[0031] reading a coupling mode entry corresponding to a current time window from the cross-state index, and generating a coupling mode descriptor representing a coupling relationship between a support contact state and an air path state;

[0032] selecting and parameterizing an annealing action primitive in a controller based on the coupling mode descriptor, to obtain a topology annealing control vector;

[0033] generating a topology annealing sequence parameter set according to the topology annealing control vector, and driving an annealing step sequence comprising unlocking, micro-amplitude attitude disturbance, standing still, and locking;

[0034] after the annealing step sequence is executed, updating the first state identifier and the second state identifier, and determining a preset annealing termination condition, and when the condition is not met, re-generating the topology annealing control vector according to the updated coupling mode descriptor and repeating the execution.

[0035] Optionally, the coupling mode descriptor comprises a coupling correlation strength and a coupling priority direction;

[0036] the annealing action primitive at least comprises an attitude disturbance direction, an attitude disturbance amplitude, a disturbance rhythm, a standing still time length, and a locking sequence, and a state mapping table of a first state machine is used to identify a coupling hotspot support point set to determine that the locking sequence is first non-hotspot support points and then hotspot support points;

[0037] the sampling trigger of the annealing step sequence is defined by the second state identifier as being turned on when the air path meets a preset stability criterion;

[0038] when the coupling correlation strength does not meet a threshold value, a preset attenuation adjustment is performed on the attitude disturbance amplitude and the disturbance rhythm along the coupling priority direction, and then the next annealing step sequence is entered.

[0039] Optionally, the triggering formal interferometric measurement based on the starting state confirmation signal comprises:

[0040] after the starting state confirmation signal is generated, the control channels of the attitude execution mechanism and the locking execution mechanism are frozen, and the first state identifier and the second state identifier are continuously read during sampling, and when any state identifier is detected to switch, the sampling is stopped and the starting state establishment process is rolled back;

[0041] Take the second state identifier as a sampling gate, and adopt a dual sampling strategy to obtain a first interferometric frame and a second interferometric frame in adjacent time slots, requiring that the second state identifier has a modal coefficient fluctuation rate lower than a preset threshold and its time sequence code is consistent, to form a candidate frame pair;

[0042] Perform a back mapping registration operation based on optical path geometry on the candidate frame pair, and unify the two frame interferometric data to a starting state reference coordinate system;

[0043] Based on the air disturbance modal coefficient recorded by the second state identifier, the interferometric data of the candidate frame pair is subjected to air component deduction processing to obtain face shape exclusive wavefront data, and the face shape exclusive wavefront data is written into the measurement data buffer as input for generating evaluation index data.

[0044] In a second aspect, the application provides a spherical mirror multi-working-condition face shape stability measurement system, comprising:

[0045] A construction module is configured to construct a double-layer state machine, which includes a first state machine for representing a spherical mirror support contact state and a second state machine for representing an air path state;

[0046] An acquisition module is configured to acquire load sensing data related to the support contact state and optical path monitoring data related to the air path state;

[0047] A processing module is configured to generate a first state identifier representing the support contact state based on the load sensing data, generate a second state identifier representing the air path state based on the optical path monitoring data, determine topology annealing sequence parameters based on the first state identifier and the second state identifier, and execute a topology annealing sequence;

[0048] An updating module is configured to update the first state identifier and the second state identifier in real time during the topology annealing sequence process, and generate a starting state confirmation signal when both of them meet their respective preset stability conditions;

[0049] A generation module is configured to trigger formal interferometric measurement based on the starting state confirmation signal, obtain spherical mirror face shape data under the current working condition, and generate evaluation index data representing the multi-working-condition face shape stability of the spherical mirror based on the face shape data and the corresponding first state identifier and second state identifier.

[0050] Compared with the prior art, the application introduces a double-layer state machine representing the support contact state and the air path state of the spherical mirror, and determines the support state identifier and the air path state identifier respectively, so that the surface stability evaluation process can fully consider the path dependence effect of the support structure state and the dynamic change of the air path disturbance. This method actively eliminates the state uncertainty caused by the support structure micro-slip, the locking sequence change or the air path change through a topology annealing sequence before actual measurement, and realizes accurate determination of the stable measurement state.

[0051] In addition, the method of the application also combines the double-layer state identifier updated in real time and the strict stability judgment condition, and only triggers the formal interference measurement when the state is truly stable, thereby fundamentally improving the reliability and repeatability of the measurement data. Further, the evaluation index data formed by the surface data obtained based on the stable measurement state and the state identifier comprehensively and accurately reflects the comprehensive influence of the cross-coupling between the support contact state and the air path disturbance on the surface stability of the spherical mirror, significantly improves the objectivity and accuracy of the multi-working-condition surface stability evaluation result, and has a significant advantage compared with the prior art. BRIEF DESCRIPTION OF DRAWINGS

[0052] Figure 1 A flowchart of a spherical mirror multi-working-condition surface stability measurement method provided by an embodiment of the application is shown in the figure.

[0053] Figure 2 A flowchart of a method for constructing a double-layer state machine provided by an embodiment of the application is shown in the figure.

[0054] Figure 3 A flowchart of a method for generating a first state identifier representing a support contact state provided by an embodiment of the application is shown in the figure.

[0055] Figure 4 A schematic diagram of a spherical mirror multi-working-condition surface stability measurement system provided by an embodiment of the application is shown in the figure. DETAILED DESCRIPTION

[0056] The technical solutions in the embodiments of the application will be clearly and completely described below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, not all the embodiments of the application.

[0057] Referring to Figure 1 A flowchart of a spherical mirror multi-working-condition surface stability measurement method provided by an embodiment of the application is shown in the figure, which includes steps S101-S105, wherein:

[0058] S101: Construct a double-layer state machine, which includes a first state machine for representing the support contact state of the spherical mirror and a second state machine for representing the air path state.

[0059] S102: Collecting load sensing data related to support contact state and optical path monitoring data related to air path state;

[0060] S103: Based on the load sensing data, generating a first state identifier representing the support contact state; based on the optical path monitoring data, generating a second state identifier representing the air path state; based on the first state identifier and the second state identifier, determining topology annealing sequence parameters and executing the topology annealing sequence;

[0061] S104: During the execution of the topology annealing sequence process, the first state identifier and the second state identifier are updated in real time, and when both meet their respective preset stability conditions, a starting state confirmation signal is generated;

[0062] S105: Based on the starting state confirmation signal, triggering formal interferometric measurement to obtain spherical mirror surface type data under current working conditions, and based on the surface type data and corresponding first state identifier and second state identifier, generating evaluation index data for representing the spherical mirror multi-condition surface type stability.

[0063] For the above S101:

[0064] In an embodiment, in order to effectively identify the key state affecting the surface type stability under the multi-condition of the spherical mirror, the present application first constructs a double-layer state machine. The double-layer state machine includes a first state machine and a second state machine, which are respectively used to represent the support contact state and the air path state of the spherical mirror. The first state machine takes the output of the multiple load sensors arranged on the back or side of the spherical mirror as input, and can classify the stress conditions of different support points into a limited state category, such as bearing state, light touch state or leaving state. The second state machine takes the monitoring channel data and environmental sensor data arranged on the periphery of the interferometer optical path as input, and is used to abstract the air disturbance situation into a limited state, such as a relatively stable laminar flow corridor state, a local convection disturbance state or a mixed band state.

[0065] Through this double-layer modeling method, different physical influence sources are converted into traceable state identifiers, thereby providing a data basis for subsequent annealing control and stability determination.

[0066] It should be emphasized that the "state machine" referred to in the present application is an abstract logical modeling method, which is used to discretize continuous physical quantities into a limited number of state sets, and define the transition conditions between these states. In this way, complex physical processes can be represented and tracked using a limited number of states and state switching, thereby facilitating the use in subsequent data processing and determination links.

[0067] For example, for the support structure of a spherical mirror, multiple load sensors continuously output values that exhibit continuous changes over time. By introducing a state machine, the force conditions of each support point can be discretized into limited states such as "bearing", "lightly touching", and "leaving", and the switching sequence between states can be recorded. Similarly, for the air disturbance path, although the refractive changes of the air flow field are continuous, the state machine can abstract them into limited states such as "laminar corridor", "convective plume", and "mixed zone". In this way, the complex support topology and air disturbance are both mapped into discrete states that are easy to calculate and compare. Subsequently, when performing topology annealing and stability determination, it is only necessary to determine whether the state identifier meets the preset conditions, without the need to solve all continuous physical quantities in real time.

[0068] In actual use, for example, when detecting a vertical installation spherical mirror with an aperture of about 3 meters, 48 back support points and several lateral support points simultaneously provide real-time load data. Through the first state machine modeling, it can be identified in a short time whether some support points have changed from the bearing state to the lightly touching state, which can easily cause local drift of the overall surface shape if not controlled. At the same time, in the same detection environment, the interferometer optical path is about 30 meters long, and the distribution of air disturbance changes with temperature gradient and pitch angle. The second state machine can give the discrimination result of "whether the current optical path is in a stable laminar state" within the sampling window.

[0069] For example, during the process of the ambient temperature rising from 20°C to 26°C, the fluctuation rate of the air disturbance mode exceeds the preset threshold, and the second state machine outputs "convective disturbance state".

[0070] Through the output of the above-mentioned double-layer state machine, the evaluation method can double the characterization of the support topology and air disturbance state in the early stage, and serve as a prerequisite for the subsequent stability evaluation process. Those skilled in the art can select appropriate sensor arrangements and state sets for modeling according to different apertures, support arrangements, or environmental conditions, and the specific state categories are not limited to the above examples.

[0071] For the above S102:

[0072] In an embodiment, in order to provide input data for the double-layer state machine, the method of the present application comprises collecting load sensing data related to the support contact state and optical path monitoring data related to the air path state. The load sensing data refers to the output signals of force sensors arranged at the positions of the multi-point support or lateral support of the back of the spherical mirror, which represent the magnitude and variation of the force borne by each support point. The optical path monitoring data refers to the data representing the air disturbance condition obtained by the peripheral channels or auxiliary sensors of the interferometer detection optical path, such as the amount of spot tilt, the amount of phase disturbance, the amount of light intensity fluctuation, and the temperature and humidity parameters, etc. Both types of data are accessed through a unified data acquisition system and are time-stamped at the time of acquisition to ensure the alignment of data from different sources on the time axis.

[0073] In a specific implementation, for example, for a vertical spherical mirror with an aperture of about 3 meters, strain gauge force sensors are installed at the positions of the 48-point back support and partial lateral support, with a range of 0-2000 N and a sampling frequency of 100 Hz, and the signals are transmitted to the data acquisition card through the CAN bus.

[0074] In the same detection scenario, the optical path monitoring part is arranged with 4 tilt detectors and 2 temperature and humidity sensors beside the interferometer optical path, with a sampling frequency of 20 Hz, which is transmitted to the host computer through the Ethernet interface. The raw data collected can also be subjected to zero-point calibration and sliding average filtering before entering the state machine, in order to remove high-frequency noise and reference drift. In one sampling period, the system can simultaneously obtain the force curves of dozens of support points and multiple air disturbance curves, and organize them into a unified multi-dimensional matrix input into the double-layer state machine, for generating subsequent state identification.

[0075] Those skilled in the art can freely select different types of force sensors, air monitoring devices, sampling frequencies and data interface methods according to the aperture size, support number and detection environment of the spherical mirror. The above examples are only one possible implementation and do not constitute a limitation on the method of the present application.

[0076] For S103 above:

[0077] In an embodiment, the generation of the first state identification is based on the continuous collection and synchronization of the aforementioned load sensing data. Specifically, the load sequence of each support point can be statistically analyzed within a fixed time window, for example, 0.5 s-2 s, to obtain a set of characteristic quantities for representing the steady state and disturbance degree of the support point, such as the time window average value, the fluctuation amplitude within the time window, and the change rate between adjacent samples; after zero removal, drift removal and simple smoothing, the system calls a preset set of state criteria to discretize the single-point force from the continuous value domain to a finite state set, such as bearing, light touch or leaving, and forms a mapping table of support point-state.

[0078] wherein the state criterion can be implemented by threshold band discrimination, clustering-based interval division or other equivalent supervised / semi-supervised discrimination procedures, and is not limited to a certain algorithm. Subsequently, the system combines the discrete states of all support points in a unified coordinate to obtain a "first state identifier" for the current moment.

[0079] For example, when the spherical mirror is switched at a small angle of pitch, the average load of a group of support points near the lower edge decreases and the fluctuation amplitude increases in a short time window, and the discrimination result tends to be a light touch state. This spatial distribution is included in the first state identifier, which is used to prompt the subsequent sequence to focus on the load recovery of this area.

[0080] In another embodiment, the generation of the second state identifier is based on the spatio-temporal stability discrimination of air disturbance based on optical path monitoring data. The observation quantity related to air refraction disturbance is obtained from the optical path peripheral monitoring channel, such as spot tilt, phase disturbance, light intensity fluctuation, and temperature and humidity change, etc. The fluctuation amplitude, stable duration and correlation or consistency index between multiple points are calculated in the same time window, and compared with the preset criterion. The air path is discretized into a finite state set, such as laminar corridor, convection disturbance or mixing zone, and output as the second state identifier.

[0081] In addition, the discrimination can introduce threshold hysteresis and minimum duration constraints to avoid frequent switching caused by transient disturbance. For example, during the process of slowly increasing the temperature of the factory building from 20 ℃ to 24 ℃, the system can judge that it is in a laminar corridor state when the slow drift of the spot tilt in multiple channels is in the same direction and the correlation is improved. When the air conditioner starts and stops and introduces local vortex, the system can judge that it is in a convection disturbance or mixing zone when the fluctuation amplitude of each channel increases and the consistency decreases in a short time, and update the second state identifier accordingly.

[0082] In yet another embodiment, based on the first state identifier and the second state identifier, the topology annealing sequence parameters are determined, and the topology annealing sequence is executed, which is a preprocessing procedure targeting at state convergence.

[0083] For ease of understanding, the topology annealing sequence parameters are understood by the present application as a parameterized description of a group of standardized actions, including but not limited to unlocking, micro attitude disturbance, static and locking, and can be executed in a predetermined order.

[0084] In a specific implementation, first, the current first state identifier and second state identifier are read, and the implementation direction of the micro attitude disturbance is selected according to the spatial distribution and stability indication, such as applying small amplitude disturbance in one degree of freedom of pitch or azimuth, disturbance amplitude and rhythm, such as small angle, small step, intermittent, and trigger sequence of static duration and locking operation; then the unlocking-micro attitude disturbance-static-locking step sequence is executed according to the parameters.

[0085] At the end of each sub-step, the system refreshes the two types of state indicators: if the first state indicator shows that the previous touch region has returned to bearing, and the second state indicator shows that the air path is in a stable state, proceed to the next sub-step; if either state is still unstable, the system maintains or fine-tunes the current parameters, and continues to disturb and alternate with small amplitude in the same degree of freedom to promote the state to converge. Through this "determine - execute - refresh" closed loop, the topology annealing sequence pushes the support contact state and the air path state into the stable domain in a gentle and traceable manner, thereby creating repeatable initial conditions for subsequent initial state confirmation and formal interference measurement.

[0086] For example, when the first state indicator prompts partial touch at the lower edge and the second state indicator prompts convective disturbance, the system can choose to disturb in the pitch degree of freedom with smaller amplitude and shorter rhythm, and prolong the static state to reduce the probability of triggering air disturbance switching again; when both types of indicators show stability, the topology annealing in this embodiment is completed.

[0087] It can be understood that the above parameter selection and action sequence can be configured according to different equipment, site and control resolution, and is not limited to a single implementation form.

[0088] For S104 described above:

[0089] In an embodiment, during the execution of the topology annealing sequence, i.e. the cycle process of unlocking - small amplitude attitude disturbance - static - locking, the first state indicator and the second state indicator are updated in real time, and a start state confirmation signal is generated when both meet their respective preset stable conditions.

[0090] For ease of understanding, the meanings and implementations of real-time updating, stable conditions and start state confirmation signals are described below.

[0091] For real-time updating:

[0092] At each sub-stage of the annealing sequence, for example after completing small amplitude attitude disturbance or static, the system reads new round of load sensing data and optical path monitoring data from the data acquisition system under the synchronization time stamp, after zero correction, de-drifting and smoothing, inputs the first state machine and the second state machine respectively, to obtain the first state indicator and the second state indicator at that time.

[0093] In order to suppress the influence of short-term burst on judgment, the system can maintain a sliding time window, for example 1-5 s, to perform consistency statistics on the state indicators of the last few frames: including state invariant count, state switching count, variance or change rate of characteristic quantity, etc. Thus, real-time updating not only includes refreshing the latest frame of state indicator, but also includes cumulative evaluation of state stability in the near period.

[0094] In this embodiment, the update of the first state identifier focuses on the spatial distribution of the discrete states (supported / carried / light touch / leave) of each support point; the update of the second state identifier focuses on the discrete state (laminar corridor / convection disturbance / mixing zone) of the air path and its accompanying disturbance intensity indication. The above-mentioned time window, sampling rhythm, and statistical caliber can be configured according to the device resolution and environmental disturbance characteristics, and are not uniquely limited.

[0095] For stable conditions:

[0096] To avoid misjudging transient pseudo-stable as a usable starting state, this embodiment adopts a double criterion with hysteresis and minimum duration constraint:

[0097] First, the first stable condition corresponding to the first state identifier: within the sliding time window, the number of support points marked as light touch or leave does not exceed the preset threshold, and the spatial position does not spread; for the support point set identified as a historical hot spot, its state needs to continuously maintain a trend of being carried or converging to being carried; at the same time, the load characteristic quantity (such as fluctuation amplitude, change rate) of each support point is overall lower than the preset upper limit, and does not appear reverse transition between two adjacent sub-stages.

[0098] Second, the second stable condition corresponding to the second state identifier: within the sliding time window, the air path state remains in the stable set, such as the laminar corridor, or its disturbance intensity indication continuously below the threshold; the consistency and correlation indicators between multiple monitoring channels are not lower than the lower limit; if it has just experienced a state switch, it needs to meet the minimum duration before being counted as stable. This embodiment can set a hysteresis interval for the second stable condition to reduce repeated switching near the critical value.

[0099] The above-mentioned threshold, minimum duration, hysteresis width, and correlation lower limit are configurable parameters, which can be adjusted according to different plant scales, optical path lengths, and sensor noise characteristics.

[0100] When the first stable condition and the second stable condition conflict, for example, the support side shows stability while the air side is still disturbed, the system does not output a starting state confirmation signal, but returns to the annealing sequence cycle: maintaining small amplitude, low energy level perturbation and alternating static, slightly extending the static duration or reducing the perturbation rhythm, so that the second state tends to converge; if the opposite (air stable while support not stable), then prefer to the first state of the hot spot area for more fine step amplitude posture perturbation and more strict locking sequence control. For the case of long time without simultaneous stability, a timeout branch can be triggered: record the un-converged reason and key statistics, stop the current annealing and use the information for parameter reconfiguration or manual review.

[0101] The conflict processing strategy of this embodiment is based on the principle of not expanding the disturbance source, avoiding new state switching caused by excessive intervention.

[0102] Further, when both the first stability condition and the second stability condition are satisfied within the same determination window, the system generates a start state confirmation signal. The signal can include, but is not limited to, a timestamp, a current attitude reading and a locking state snapshot, a digest of the first state identifier and the second state identifier, such as a hash or compressed description, key statistics for determining stability (such as an invariance count, an upper bound on the rate of change, a lower bound on the correlation), and a parameter version identifier. The start state confirmation signal is used to indicate that the current working condition has entered a stable measurement starting point that is repeatable and traceable; the subsequent process can be triggered by the signal to start the formal interferometric measurement, but the embodiment does not limit the measurement details.

[0103] For example, for a spherical mirror with an aperture of about 3 m and vertically installed, in one annealing cycle, the first state identifier converges from sporadic light touch at the lower edge to full-area bearing within 12 s, while the second state identifier transitions from the mixed band to the laminar flow corridor within 8 s, and the stable intervals of the two overlap for 4 s; accordingly, the system generates a start state confirmation signal at the 20th second.

[0104] If a temporary rebound of the air path occurs in the subsequent perturbation phase (the second state identifier leaves the stable set), the system cancels the incomplete determination, returns to the resting state and extends the duration until the double stability condition is met again. The above-mentioned time, threshold and rhythm parameters can be determined according to the equipment and environment, and do not constitute a limitation on the present application.

[0105] In this way, the execution of the annealing sequence, the refresh of the state identifier and the double criterion of stability form a closed loop, which can provide a starting reference with repeatability for subsequent surface measurement without relying on forced numerical compensation.

[0106] For the above S105:

[0107] After generating the start state confirmation signal, the system enters the formal interferometric measurement phase. In order to maintain the consistency of the measurement and the starting reference, the controller freezes the execution authority of the attitude and the locking, and only keeps the data acquisition and measurement trigger channel in working state.

[0108] It can be understood that the measurement can be triggered by the host computer issuing an instruction, or can be automatically triggered by the start state confirmation signal. The obtained interferometric data is processed by a reference to form surface data under an effective aperture mask, and the reference processing includes removing the piston term and the tilt term, and whether to retain the power term is determined by the configuration.

[0109] In order to facilitate tracing and comparison, the surface data is unified to the start state reference coordinate system, and is recorded together with the timestamp, attitude reading, locking state snapshot, environmental parameters and corresponding first state identifier and second state identifier of this measurement.

[0110] To control data quality, a short stabilization window after the start state confirmation can be used as the sampling interval. If a first state identifier or a second state identifier is detected to switch within this interval, the current frame is discarded and the start state establishment process is returned. Qualified frames within the same stabilization window are registered and synthesized to weaken the influence of random disturbances. Registration uses a back mapping method based on optical path geometry or an equivalent registration method to unify each frame to the same reference system; synthesis can take simple average, weighted average or robust estimation, and the weight can be determined by the stability indicator provided by the second state identifier. Through gating, registration and synthesis, the surface data consistent with the start state reference and meeting the stability criterion is obtained.

[0111] Based on the surface data and the one-to-one correspondence between the first state identifier and the second state identifier, the system generates evaluation index data for characterizing the surface stability of the spherical mirror under multiple working conditions. The evaluation index data includes single-condition indicators, cross-condition stability indicators, and path coupling related indicators.

[0112] Among them, the single-condition indicators include PV value and RMS value, band-limited RMS divided by spatial frequency band, low-order modal coefficient set obtained by modal decomposition, and slope statistics and maximum residual position. Cross-condition stability indicators are obtained by aggregating and comparing single-condition indicators of multiple conditions, including global stability index of RMS maximum deviation and its confidence in the condition set, modal stability spectrum composed of amplitude and phase changes of low-order modes between different conditions, sensitivity matrix of attitude and temperature on RMS or key low-order modes, and stability map showing available area and degradation area in attitude and temperature plane.

[0113] Path coupling related indicators are calculated using the first state identifier and the second state identifier recorded synchronously with the measurement, including path-dependent indicators of surface differences at the same set point through different return paths, correlation between state changes in support hot spot areas and low-order aberration changes, correlation between air path stability and low-frequency surface changes, and complementary stability window proportion within the time proportion that meets the double-state stability.

[0114] The specific caliber, order and threshold of the above indicators can be adjusted according to the equipment and environment, and are not the only limitation.

[0115] The evaluation index data is output in the form of a structured data packet, which includes the index entries of the current and previous measurements, the working condition labels, the corresponding first and second state identifier summaries, the statistical confidence and quality control marks, and can give a visualization reference of the stability map and the sensitivity matrix. Taking a spherical mirror with a diameter of about three meters and using a long-range collimation measurement as an example, the RMS obtained by the system under the working condition of 0° / 20 ℃ can be on the order of λ / 160, and the RMS obtained under the working condition of +1° / 26 ℃ can be on the order of λ / 120, and the global stability index across the working conditions shows that the differential amplitude is about 30% or more;

[0116] wherein λ herein refers to the nominal working wavelength used in the interferometric measurement, for example, 632.8 nm of He-Ne light, which is used to normalize the representation of wavefront or surface error, such as 0.02λ representing 2% of the working wavelength.

[0117] When re-measured according to different zero-return sequences, the path-dependent indicator shows distinguishable differences, and at the same time, the low-frequency RMS fluctuation is significantly lower than that during the convection disturbance period when the air path is in the laminar flow corridor. This result is used to illustrate the formation of the index and the typical phenomena that can be observed, and the actual value and threshold can be configured according to the field conditions.

[0118] Through the above process, the formal interferometric measurement and the initial state confirmation, the state identifier tracing, and the index generation form a closed-loop output. The obtained evaluation index data not only reflects the surface level under a single working condition, but also comprehensively depicts the influence of the support contact state and the air path state on the surface stability in the multi-working condition dimension, which can provide comparable, traceable, and decision-making data basis for the adjustment, operation, and maintenance of the spherical mirror.

[0119] Optionally, referring to Figure 2 A flowchart of a method for constructing a double-layer state machine provided by an embodiment of the present application, comprising steps S201-S203, wherein:

[0120] S201: determining the contact micro-state characteristics of each support point based on the output data of the multiple load sensors of the multi-point support mechanism, and performing clustering operation on the contact micro-state characteristics to generate the state identifier of each support point and constitute the state mapping table of the first state machine;

[0121] S202: determining the disturbance characteristics of the air path based on the output data of the optical path peripheral monitoring channel of the interferometric measurement assembly and the environmental monitoring sensor, and performing classification operation on the disturbance characteristics to generate the state mapping table of the second state machine;

[0122] S203: establishing the cross-state index between the first state machine and the second state machine, which is used to identify the coupling mode description quantity between the support contact state and the air path state.

[0123] In an embodiment, the construction of the dual-layer state machine starts with the processing of the load sensor outputs of the multi-point support mechanism. The system pre-processes the time-synchronized load sequences of each support point, including zero-point removal, drift removal, and smoothing, and then calculates the statistical and response features that can represent the steady state and the disturbance level within a sliding time window.

[0124] To discretize the continuous values into finite states, the system performs clustering operations or equivalent discriminant procedures in the feature space described above, and labels the force state of each support point at the current time as categories such as bearing, touching, or leaving, etc.

[0125] Subsequently, the support point number and its discrete state are combined and registered with the time stamp to form the state mapping table of the first state machine. This mapping table is continuously updated in time, reflecting not only the current spatial distribution, but also the state evolution trajectory in the near period, which is used to support subsequent stability judgment and process control.

[0126] For the air path state, the system extracts feature quantities that can characterize the disturbance level and stability of the refractive field based on the synchronous data of the light path peripheral monitoring channels of the interferometric assembly and the environmental monitoring sensors, including multi-point tilt, phase or light intensity swing amplitude indicators, cross-channel consistency and correlation indicators, temperature and humidity change rates, etc.

[0127] It should be noted that, in order to achieve discretized expression, classification operations are performed in this feature space, and the current air path is labeled as categories such as laminar corridor, convection disturbance, or mixing zone, etc. The category label and feature summary are written into the state mapping table of the second state machine with the time stamp as the index. This mapping table is also maintained in time series to reflect the stable maintenance and switching of the air path during the measurement phase.

[0128] After the establishment of the two state mapping tables, the system aligns them based on the unified time axis to form the cross-state index for identifying the relationship between the support contact state and the air path state. This index measures the covariation strength between the local state change on the support side and the disturbance change of the air path by statistically analyzing the co-occurrence and sequence of the two types of states within the same time window, and combines the sequence information and spatial distribution information of state switching to obtain a set of description quantities for quantifying the coupling relationship between the two, which are used as coupling mode description quantities for subsequent steps.

[0129] Exemplarily, when the spherical mirror enters the sampling stage at a pitch angle θ = + 1° and an ambient temperature T = 26℃, the state mapping table of the first state machine shows that the lower edge support points are switched from bearing to light touch, and the state mapping table of the second state machine shows that the air path is switched from laminar corridor to convective disturbance at the same time. The cross-state index gives a higher covariant indication and a clear time sequence accordingly. The result is recorded as the coupling mode description quantity of the current time window, which is used to drive the subsequent parameter setting and annealing process. In order to adapt to different apertures, support arrangements and environmental conditions, the above-mentioned feature selection, clustering and classification method, time window length and correlation aperture can be adjusted as needed, and are not the only limitation.

[0130] The optional implementation is aimed at the problem of measurement uncertainty caused by the difficulty in simultaneously identifying the support contact state path dependence and air path disturbance remapping in existing evaluation. By establishing a first state machine and a second state machine, forming two types of state mapping tables through clustering operation and classification operation, and constructing a cross-state index on a unified time axis to obtain a coupling mode description quantity, the quantitative description and real-time tracking of the dual-domain relationship of support-air are realized.

[0131] In this way, continuous noise type observation is converted into determinable discrete state, which significantly improves the identification robustness and reproducibility under unstable working conditions. The coupling mode description quantity provides a calculable basis for the selection of topological annealing parameters, reducing blind tuning and "pseudo-stable" release. The time sequence and spatial distribution record form a traceable data asset, which is convenient for cross-condition comparison and long-term trend evaluation, thereby improving the accuracy and operability of multi-condition surface stability evaluation as a whole.

[0132] Optionally, the contact micro-state feature is a feature vector composed of load fluctuation of the load sensor and support point displacement response in a continuous time window;

[0133] The disturbance feature is the modal coefficient fluctuation rate of the air path disturbance projected onto the preset low-order spatial modal basis in a continuous time window;

[0134] The state mapping tables of the first state machine and the second state machine respectively record the state switching sequence and generate a path history code, and the cross-state index is established by associating the path history codes of the first state machine and the second state machine.

[0135] In order to accurately describe the small state changes of support contact and the low-order disturbance patterns of air path under multiple working conditions, and to include the historical sequence of the two into the same index, so as to reveal the path dependence and coupling relationship.

[0136] The embodiment constructs a contact microstate feature vector for each support point within a continuous time window. The load fluctuation can take the standard deviation or the range within the window, or the residual energy after median filtering. The displacement response can be read from the micro-displacement sensor at the support point, or estimated from the actuator encoder and the stiffness model. Both constitute the feature vector of the support point in the window. The window length can be set to 0.5s to 2s according to the sampling frequency and the environmental disturbance characteristics, and the window sliding step can be an integer multiple of the sampling period.

[0137] To eliminate the influence of overall temperature drift or slow drift, the feature vector can be de-meaned and normalized before entering clustering or discrimination, and a neighborhood constraint can be introduced to prevent the amplification of adjacent support point anomalies. The system processes the feature vectors of all support points in parallel, outputs the discrete states such as bearing touch-off, and updates the state mapping table of the first state machine accordingly.

[0138] For air path disturbance, the system aggregates the tilt amount, phase disturbance amount, and light intensity fluctuation amount of the optical path peripheral monitoring channel within the same time window, first performs attitude and geometric mapping to make the data fall into a unified reference system, and then projects it onto a preset low-order spatial modal basis.

[0139] Among them, the modal basis can be a low-order Zernike term, or an orthogonal polynomial or sine-cosine basis, and the order and basis type are determined by the field conditions. The system calculates the fluctuation rate of each modal coefficient within the window, and can use the ratio of the coefficient standard deviation to the mean or the root mean square of the adjacent frame difference as the fluctuation rate measure. According to the combination of the fluctuation rates of each coefficient and the threshold rule, the window is classified as an air laminar flow corridor or a convection disturbance or a mixing zone, and the state mapping table of the second state machine is updated accordingly.

[0140] To avoid threshold boundary jitter, the category determination can set a minimum duration and a hysteresis band, and the specific values are determined by the device resolution and the plant air flow conditions.

[0141] The two state mapping tables not only record the current category, but also record the state switching sequence. The system generates a path history code for each time window. The first path history code is compressed from the category change sequence of each support point in the window queue, which can use the hash connection of support point index and state number, or use the order list according to spatial partitioning. The second path history code is compressed from the change sequence of the air path category over time, and is accompanied by the sign and amplitude interval identification of the dominant low-order modal term. The cross-state index aligns the two types of path history codes through a unified time axis to generate a co-occurrence table and a precedence relationship table, and further gives the covariant strength and time priority direction in the coupling mode description quantity.

[0142] For example, taking a 3 m vertical spherical mirror as an example, taking a 1 s window and a 0.1 s step, at the stage of θ = +1° and T = 26℃, the load fluctuation of the lower edge support points increases and the displacement response increases, which is judged to be converted from bearing to light touch, at the same time, the low order term fluctuation rate of the air mode increases and is converted from laminar corridor to convection disturbance, the first path history code and the second path history code of the time window are significantly co-occurring in the cross index and show that the support side precedes, the system accordingly marks the covariance intensity of the window as higher and records the time priority direction. This processing provides a calculable input for the subsequent parameter setting and annealing process, while retaining the complete sequential information, facilitating cross-condition playback and comparison.

[0143] Optionally, referring to Figure 3 A flowchart of a first state identification method for generating a state identification method for characterizing the support contact state is provided for the embodiments of the present application, comprising:

[0144] S301: Calculate the load average, load fluctuation amplitude and load change rate of each support point in a preset time window, and combine the above three quantities as the load feature vector of the support point;

[0145] S302: Perform neighborhood normalization processing on each load feature vector according to the spatial distribution of the support points to obtain a relative load feature vector that eliminates the influence of overall temperature drift, and calculate the offset from the historical baseline in the relative load feature vector;

[0146] S303: Input the offset into a micro-state classification model based on support point partition, and generate state labels of bearing state, light touch state or departure state in the model combined with displacement response data or micro-vibration response data of the support point;

[0147] S304: Combine the state labels of all support points to form a first state identification, and append the time sequence code of the state switching of each support point in the first state identification to represent the path-dependent characteristics of the support contact topology.

[0148] After collecting the load sensing data, if only the absolute value of single-point load or simple threshold is used for judgment, it is easy to be affected by overall temperature drift, slow offset and mechanical common mode disturbance, leading to misjudgment of bearing and light touch; at the same time, without introducing neighborhood constraints and historical reference, the occasional fluctuation of edge support points will be amplified, and the classification result will appear jitter and bounce, which is difficult to stabilize the subsequent process of support.

[0149] Therefore, the present embodiment performs difference and normalization processing in time and space dimensions, and combines mechanical characteristics with displacement or micro-vibration response for judgment, thereby improving the robustness and repeatability of support contact state recognition.

[0150] For the above S301:

[0151] When only single-point instantaneous load values are used for determination, false positives are easily introduced by environmental vibrations and sampling noise, and it is impossible to distinguish between steady-state loading and short-term disturbances.

[0152] Therefore, three quantities are extracted from the load sensing data of each support point within a preset time window and combined into a load feature vector in a fixed order for subsequent state discrimination and mapping. The time window can be 1.0 s, the sliding step can be 0.1 s, and the sampling frequency of the load signal can be 100 Hz.

[0153] For the load samples within the window, the load average is calculated using the arithmetic mean to reflect the load reference level of the support point in the current window; the load fluctuation amplitude is represented by the difference between the upper and lower bounds of the window, and the upper and lower bounds are obtained by performing median smoothing and mild outlier rejection on the window samples to suppress the influence of occasional spikes on amplitude estimation; the load change rate is obtained by averaging the absolute values of the differences between adjacent samples in the window and then time-normalizing, which is used to represent the dynamic activity of the support point in the current window.

[0154] The above three quantities are respectively referred to as the load average, load fluctuation amplitude, and load change rate of the support point, which are concatenated in this order to form a one-dimensional load feature vector, and are stored in association with the spatial index and timestamp of the support point. Taking a 3 m vertical spherical mirror as an example, under the working condition of θ = +1°, T = 26℃, the load average of a support point at the lower edge is about 940 N, the load fluctuation amplitude is about 32 N, and the load change rate is about 5.8 N per second within a 1.0 s window. Therefore, the load feature vector of the support point in the window is (940, 32, 5.8). The time window, step, sampling frequency, and measurement thresholds of fluctuation and rate can be adjusted according to the device resolution and field disturbance characteristics, and do not constitute a unique limitation.

[0155] For the above S302:

[0156] To reduce the impact of overall temperature drift and slow drift on the judgment, the system performs neighborhood normalization according to the spatial distribution of the support points after generating the load feature vector, and obtains the drift from the historical baseline on this basis. The determination of the neighborhood can be based on the mechanical layout to divide the support points into several annular zones or sectors, or a fixed size neighborhood set can be determined according to the number of neighbors of each support point. For each support point, the system first calculates three local statistics in its neighborhood, which correspond to the local center value and the local scale value of the load average, the load fluctuation amplitude and the load change rate respectively. The local center value is preferably a truncated mean or a median to improve the anti-outlier ability, and the local scale value is preferably a quartile range or a robust standard deviation. Then the three-dimensional load feature vector of the support point is subtracted by the local center value and divided by the local scale value according to the component, to obtain the relative load feature vector, so that it mainly reflects the difference relative to the nearest environment rather than the global common mode change. To avoid sharp discontinuity introduced by random noise in the neighborhood, the system can do a slight smoothing according to the spatial graph structure after normalization is completed, and the smoothing weight is set according to the adjacent distance and the structure connection strength.

[0157] To ensure the comparability of different sampling rounds, the system imposes a global consistency constraint on the relative vector, so that the relative average component of all support points in the same time window is close to zero. The specific method is to correct the relative component to zero mean according to the annular zone or sector, or to remove the residual linear trend in the whole field range. For the points of the edge or near-missing sensor, the system completes the missing component by neighborhood interpolation and adds a quality mark, and the low-quality samples do not participate in the subsequent threshold judgment, but are only used for interpolation display.

[0158] The historical baseline is fixed by one or more confirmed stable reference phases. The confirmation of the reference phase is based on the double-state stability condition, and the relative load feature vector is generated using the same time window length and preprocessing aperture as the current processing. Then the baseline center value and the baseline scale value are obtained according to the support point index, and the version number and the time stamp are recorded. To cope with the slow aging or long-term drift of the device, the system supports a rolling update strategy, which only updates the baseline with a small weight when it passes the stability judgment again, while retaining the old version for backtracking. The baseline and the current relative vector are aligned in the same coordinate and the same mask, ensuring the consistency of the comparison.

[0159] After obtaining the relative load feature vector, the system calculates the offset from the historical baseline. The offset includes three component differences and a dimensionless offset degree normalized by the baseline scale, and a weighted comprehensive strength can also be given for sorting. The component difference reflects the deviation of the point in the load level, the disturbance amplitude and the dynamic rate relative to the stable reference, and the dimensionless offset degree facilitates cross-point and cross-period comparison. To suppress occasional mutations, the offset can be short-window moving average in time and set an upper limit for clipping. Samples exceeding the upper limit are marked as abnormal and trigger review or rejection.

[0160] For example, the load eigenvector of a 3 m vertical spherical mirror at θ = +1°, T = 26℃ is 940, 32, 5.8 at a certain support point on the lower edge within a 1.0 s window. The local neighborhood center values are 950, 20, 3.0, and the local scale values are 20, 8, 1.0. The normalized relative load eigenvector of this point is about -0.5, 1.5, 2.8. The relative vector center of the historical baseline at the same index is -0.25, 1.0, 1.5. The component difference between the current and the baseline is about -0.25, 0.5, 1.3. The non-dimensional offset obtained by the baseline scale normalization falls in the medium level, and the corresponding quality mark is valid. This offset will be one of the inputs of the subsequent microstate classification model and will participate in the generation of the time sequence code for representing the evolution path of the contact topology. The time window, neighborhood division, robust statistical threshold, smoothing and clipping threshold can be adjusted according to the equipment resolution and the characteristics of the field disturbance, and do not constitute a unique limitation.

[0161] For S303 and S304 above:

[0162] In an embodiment, the system inputs the aforementioned offset from the historical baseline into a microstate classification model constructed by support point partitioning. The partitioning can be divided into annular zones or sectors according to the support geometry, so that the model learns or sets more appropriate thresholds and criteria in a local range. The input of the model at least includes the three-dimensional offset of the support point in the current time window, and displacement response data or micro-vibration response data are introduced as auxiliary quantities. The displacement response can be directly given by the displacement sensor of the support point, or it can be estimated by the encoder of the actuator combined with the support stiffness. The micro-vibration response can be given by the acceleration sensor close to the support point in a limited frequency band. The output of the model is a state mark of one of the three categories of bearing, touching and leaving, with a confidence value or distance measure attached for quality control. In order to reduce jitter, hysteresis and minimum duration constraints are set inside the model.

[0163] Taking a rule implementation as an example, a three-dimensional offset threshold can be set in each partition and jointly decided with an auxiliary quantity threshold, and different threshold bands are used when entering or exiting a certain state.

[0164] Taking a statistical learning implementation as an example, stable bearing and touching samples can be collected during the commissioning stage, and the cluster centers and covariances are fitted. The Mahalanobis distance and auxiliary quantity consistency are used for online judgment. Both implementations can deploy quality marks, and when the input signal is missing or noise is out of limit, the output maintains the last stable state and is marked as to be retested.

[0165] At the same timestamp, the system aggregates the state labels of all support points, and combines them into a first state identifier of the current moment according to the spatial index. The identifier reflects the discrete spatial distribution of the support contact state in the whole field, and is the direct input of the subsequent cross-indexing and annealing process. In order to represent path dependence, the system also generates a time sequence code to record the chronological events of state switching of each support point within the sliding window.

[0166] Among them, the time sequence code can be stored in the form of event sequence, and the events include timestamp, support point index, pre-switching state, post-switching state. The same type of switching of adjacent support points can also be merged and compressed at the spatial partition level to generate partition-level sequence entries.

[0167] In order to balance the volume and traceability, the system can de-duplicate and threshold cut the original event sequence, and generate sequence abstract and hash for quick comparison. The time sequence code and the first state identifier are archived synchronously on the time axis, so that the spatial distribution at the same moment and the recent evolution path can be retrieved and reconstructed at the same time.

[0168] For example, taking a 3 m vertical spherical mirror as an example, when θ = + 1° and T = 26℃, the lower edge support point of a certain ring appears load offset rise and displacement response increase within a 1.0 s window, and the micro-vibration energy rises synchronously. The partition model judges that the point is switched from bearing to light touch according to this, and the adjacent two points successively undergo the same type of switching at 0.2 s and 0.4 s. The system generates a first state identifier at this timestamp and writes three light touch labels, and records three switching events in time sequence in the time sequence code. The sequence abstract shows that the switching propagates along the edge direction. The subsequent annealing process can select the direction and rhythm of the micro attitude disturbance according to this, and take more strict locking sequence and longer static time in the hot spot area. The above partition method, the acquisition method and the calculation caliber of the auxiliary quantity, the specific implementation of the classification model, and the specific format of the sequence coding can be adjusted or replaced according to the device resolution and the field conditions, and should not be understood as a limitation on the claims of the present application.

[0169] Optionally, the second state identifier representing the state of the air path is generated based on the optical path monitoring data, comprising:

[0170] Perform attitude normalization processing on the air disturbance data collected by the interference measurement assembly optical path peripheral monitoring channel to obtain normalized air disturbance data;

[0171] Determine the disturbance feature quantity based on the normalized air disturbance data, and input the disturbance feature quantity into an air path classification model to generate a state label of the air path;

[0172] Combine the state label of the air path to form a second state identifier.

[0173] In an embodiment, the system takes the air disturbance data obtained by the interferometric assembly optical path monitoring channels as input, first performs attitude normalization processing, then determines the disturbance characteristic quantity according to this, and finally aggregates the state markers of each channel at the same timestamp to form the second state identifier.

[0174] Wherein, the purpose of attitude normalization is to eliminate the geometric remapping caused by pitch and azimuth fine adjustment, so that the monitoring quantity under different attitudes falls into the same reference coordinate system. The specific method is to calculate the relative displacement and incidence angle change of the current light beam in space according to the attitude encoder reading and the optical path geometric model, and to remap and resample the tilt, phase disturbance and light intensity fluctuation of each monitoring channel according to the reference path, while using cross-channel cross-correlation to estimate the time delay and make a slight phase alignment, and if necessary, a linear scale correction according to the reference amplitude, thereby obtaining the attitude-normalized air disturbance data sequence.

[0175] After attitude normalization is completed, the system calculates the characteristic quantity that can represent the disturbance strength and stability of each channel within a preset time window. The characteristic quantity at least includes the root mean square of the tilt, the root mean square of the phase disturbance, the coefficient of variation of the light intensity fluctuation, the consistency and correlation index of the cross-channel, and the temperature and humidity change rate.

[0176] In order to enhance the sensitivity to low-order stripe structure, the system projects the phase disturbance spatial distribution into low-order modal basis, records several dominant modal coefficients and their fluctuation rates. The above features are input into the air path classification model after mean removal and scale normalization. The model can use threshold rules with hysteresis and minimum duration constraints, or use offline calibrated statistical discriminators, output the air path state marker at the time level, and the state set can include laminar corridor, convection disturbance, mixing zone, and accompanying stability indicator for quality control.

[0177] At the same timestamp, the system aggregates the state markers and stability indicators of all monitoring channels for consistency, generating the second state identifier at that time.

[0178] Wherein, the aggregation strategy preferentially adopts the majority consistency principle and is weighted by stability, and if necessary, the conflicting channels are excluded or down-weighted, while recording the state retention duration and the last switching time for subsequent stability determination.

[0179] Exemplarily, in an actual scenario, for the working condition of a 3 m vertical spherical mirror with a positive small angle θ and a temperature in a high range, the root mean square of the tilt amount in the short time window after attitude normalization can be observed to rise from a low level to a significantly high level, the root mean square of the phase disturbance jumps from a level close to the baseline to a level significantly higher than the baseline, the cross-channel correlation gradually decreases from high consistency to medium consistency, the temperature and humidity change rates show an upward trend, and the fluctuation rate of the dominant term in the low-order mode continues to be higher than the preset threshold. Based on the above-mentioned joint features, the classification model determines that the air path is in a convection disturbance state, and gives a low stability indication; the second state identifier records this state accordingly and accumulates the duration of the state.

[0180] Exemplarily, the air path classification model adopts a two-level interpretable structure. The first level is feature standardization and weighted scoring. The tilt amount root mean square, phase disturbance root mean square, light intensity fluctuation coefficient of variation, cross-channel consistency and correlation, temperature change rate and humidity change rate, and low-order mode dominant term fluctuation rate are first normalized by median and absolute deviation, and then linearly combined to form a score S according to the weight vector obtained offline. The weight can be obtained by least squares or logistic regression with constraints on the labeled data, and can be fine-tuned in the field with a small number of working conditions. The second level is threshold determination with hysteresis and minimum duration. Set the upper threshold U and the lower threshold L. When S is greater than U, it is determined to be a convection disturbance, when S is less than L, it is determined to be a laminar corridor, and between the two, it is determined to be a mixed band. The hysteresis interval width and the minimum duration are matched with the field sampling frequency and the disturbance time scale to reduce boundary jitter. The model also introduces cross-channel consistency weighting and channel health mask. When some channels are lost or the quality is marked as low, the weight is automatically reduced and a robust summary is used. The final state is determined by the principle of weighted majority consensus. The window length and sliding step, the initial value of the weight vector, the initial recommended value of U and L, the hysteresis width, the minimum duration, and the selected order of the low-order mode are issued as configurable parameters in engineering deployment. A small amount of comparative measurement is completed for calibration in the field. The calibration process is recorded in the parameter version for auditing and backtracking.

[0181] The above normalization method, feature set, classification implementation and summary caliber can be adjusted with the device and the environment, and do not constitute a limitation on the present application.

[0182] Optionally, the attitude normalization process includes performing optical path geometry remapping operation on the air disturbance data based on the optical path geometry offset caused by the change in the elevation or azimuth of the spherical mirror, to eliminate the geometric influence of the attitude change on the air disturbance measurement.

[0183] The determination of the disturbance feature quantity includes projecting the normalized air disturbance data into a preset low-order spatial modal base, calculating the fluctuation rate and phase consistency index of each modal coefficient in a preset time window, and combining the fluctuation rate and the phase consistency index as the disturbance feature quantity.

[0184] The second state identifier further includes a time sequence code of air path state switching to represent the historical dependence feature of air path disturbance.

[0185] When the air path state is directly determined only by the original light path monitoring quantity, the small changes in pitch and azimuth will introduce geometric remapping errors, causing the same disturbance to be treated as different phenomena at different attitudes, and it is difficult to separate the low-order stripe structure and random disturbance in space. The present embodiment solves the problems of geometric consistency, feature interpretability and historical dependence expression by two-step processing of attitude normalization and low-order mode projection, and introduces a time sequence code in the second state identifier.

[0186] Attitude normalization is based on the geometric calibration parameters of the spherical mirror and the interference light path and the pitch and azimuth readings given by the attitude encoder, and calculates the transverse displacement and incident angle change of the interference beam on the monitoring plane at the current time. The tilt, phase disturbance and light intensity fluctuation collected by each monitoring channel are geometrically remapped and resampled according to the reference path, so as to align them to the common reference coordinate system. To reduce the phase offset caused by synchronization error, a small time delay can be estimated by cross-channel cross-correlation after remapping and sub-sampling alignment; to reduce the influence of inconsistent sensor gains, the amplitude can be linearly scaled according to the reference section. The normalized air disturbance data is comparable across attitudes and can be used as a unified input for subsequent feature calculation.

[0187] In determining the disturbance feature quantity, the system calculates two types of complementary indicators on the normalized data within a preset time window. The first type is the intensity and instability indicators, including the root mean square of tilt and phase disturbance, the coefficient of variation of light intensity fluctuation, and the consistency and correlation of multiple channels.

[0188] The second type is the spatial form and temporal phase indicators, which are obtained by projecting the disturbance field into a preset low-order spatial mode basis to obtain a number of dominant mode coefficients, and calculating the fluctuation rate and phase consistency of each coefficient within the window. The mode basis is preferably a low-order orthogonal basis that has good characterization ability for long air path low-frequency stripes, and the order and basis type can be configured according to the field conditions. The fluctuation rate is used to quantify the temporal fluctuation of the intensity of each mode, and the phase consistency is used to describe the degree of stable maintenance of the mode phase in time. After de-meaning and scale normalization, the two types of indicators are combined according to the weight to form the disturbance feature quantity, which is used as the input of the air path classification model. The classification model outputs state labels such as laminar flow corridor, convective disturbance or mixed zone, and stability indicators.

[0189] At the same timestamp, the system aggregates the status labels of all channels to generate a second state identifier, and appends a time sequence code of the air path state switching to represent the historical dependence. The time sequence code records the state entry, maintenance and exit events on a sliding time axis, including event time, previous and subsequent states and holding time, and filters critical jitter through hysteresis and minimum duration constraints. For ease of storage and rapid comparison, the sequence code can be summarized under the premise of ensuring traceability, while the original event queue is retained to support review.

[0190] In this way, the second state identifier is independent of the pose in space, has an interpretable low-order modal component in feature, and retains the switching sequence information in time, which can be cross-indexed with the first state identifier on a unified time axis to identify the coupling mode and drive the subsequent annealing and measurement process.

[0191] Optionally, based on the first state identifier and the second state identifier, the system determines a topological annealing sequence parameter and executes a topological annealing sequence, which includes:

[0192] Reading the coupling mode entry corresponding to the current time window from the cross-state index to generate a coupling mode description quantity for representing the coupling relationship between the support contact state and the air path state;

[0193] Selecting and parameterizing an annealing action primitive in the controller based on the coupling mode description quantity to obtain a topological annealing control vector;

[0194] Generating a set of topological annealing sequence parameters according to the topological annealing control vector, and driving an annealing step sequence including unlocking, micro-pose disturbance, standing and locking;

[0195] After the execution of the annealing step sequence ends, the first state identifier and the second state identifier are updated and a preset annealing termination condition is determined. If the condition is not met, the topological annealing control vector is regenerated based on the updated coupling mode description quantity and repeated execution.

[0196] When annealing is performed only with fixed parameters, it is easy to appear that the support side has converged while the air side is still disturbed or vice versa, resulting in a long waiting for the initial state or repeated oscillation. Therefore, in the present embodiment, the system does not directly use fixed parameters, but uses the coupling information generated by the first state identifier and the second state identifier together. The system adaptively determines the annealing action primitive and its parameters based on the coupling information, and implements parameter closed-loop update based on the latest state identifier after each round of annealing ends.

[0197] In a specific implementation, first, the coupling mode entry corresponding to the current time window is read from the cross-state index, and the coupling mode description quantity used to represent the coupling relationship between the support contact state and the air path state is calculated. The coupling mode description quantity includes the covariant intensity, the time priority direction, the hotspot distribution, and the stability summary. The covariant intensity is used to measure the joint significance of the local state change on the support side and the state change on the air side within the same window. The time priority direction is used to represent the sequence of the two types of changes. The hotspot distribution is given by the first state identifier to indicate the region set where the light touch or departure is concentrated and the spatial order. The stability summary is given by the second state identifier to indicate the holding time and consistency index of the current air state. The above quantities are calculated and archived on the same time axis, so that the coupling relationship within the same window becomes a directly callable input.

[0198] The controller selects and parameterizes the annealing action primitive in the action primitive library according to the coupling mode description quantity, and obtains the topology annealing control vector. The action primitive library at least includes the posture jitter direction, the posture jitter amplitude, the jitter rhythm, the static duration, and the locking sequence. The selection and parameterization follow two rules.

[0199] For example, the first rule is to prefer small-amplitude jitter along the coupling priority direction to reduce the remapping risk, and to use smaller amplitude and longer static when the covariant intensity is high, and to enable the hotspot-priority locking sequence.

[0200] For example, the second rule is to reduce the jitter rhythm and extend the static when the air side stability is low, and to use the non-hotspot-first and hotspot-second locking strategy and to extend the static in the hotspot area when the support side hotspots are concentrated.

[0201] For example, in the scenario of a 3 m vertical spherical mirror, θ is +1°, and the temperature is high, if the cross-state index shows that the lower edge is a hotspot and the air is in disturbance, the controller can select a negative pitch perturbation, the initial amplitude is 5″ to 10″, the rhythm is 0.3 s to 0.5 s once, the static is 3 s to 5 s, and the locking sequence is the upper edge and the two sides first, and then the lower edge. The above intervals can be adjusted according to the device resolution and the field disturbance.

[0202] The system generates a specific topology annealing sequence parameter set according to the topology annealing control vector, drives the step sequence of unlocking, small-amplitude posture disturbance, static, and locking, and refreshes the first state identifier and the second state identifier at the end of each sub-stage during the execution process.

[0203] If the first state identifier shows that the hotspot region is returned by light touch and no new spread occurs, and the second state identifier shows that the air path remains stable and meets the minimum duration, the current step is considered valid and the next sub-phase is entered; if either side is still unstable, the amplitude and rhythm are attenuated by a preset factor without changing direction, and the resting time is extended before repeating the sub-phase until the maximum number of attempts or the stability meets the conditions.

[0204] For example, in one cycle, if the air stability is still low and the covariance intensity does not decrease below the threshold, the amplitude is decreased from 10" to 7" and then to 5", the rhythm is increased from 0.3 s to 0.5 s, the resting time is increased from 3 s to 5 s, and the hotspot locking sequence strategy is maintained; if the air is stable and the support is still partially touched, the rhythm is maintained, the amplitude is increased to 12" and the resting time is shortened to 2 s to promote force chain redistribution, but the locking sequence still follows the non-hotspot first and then the hotspot.

[0205] After the sequence of annealing steps is executed, the system determines the preset annealing termination condition according to the updated first state identifier and second state identifier.

[0206] For example, the termination condition can include the covariance intensity being below the threshold, the two types of states remaining stable in the same determination window without new hotspots, and the convergence degree of the current parameters meeting the requirements. If the termination condition is met, a starting state confirmation signal is output and the formal measurement is entered; if not, the latest coupling mode description quantity is used as input to regenerate the topology annealing control vector and enter the next sequence.

[0207] Through the above processes of adaptive parameter selection according to the coupling mode, feedback fine-tuning according to the sub-phase, and closed-loop determination according to the termination condition, the annealing promotes the convergence of the support and the air on both sides with a small perturbation energy level, reducing the excessive intervention and pseudo-stable release risk caused by fixed parameters, and stably generating a reproducible starting state in typical field scenarios.

[0208] Optionally, the coupling mode description quantity includes coupling correlation strength and coupling priority direction.

[0209] The annealing action primitive includes at least attitude jitter direction, attitude jitter amplitude, jitter rhythm, resting time, and locking sequence, and the state mapping table of the first state machine identifies the coupling hotspot support point set to determine that the locking sequence is the non-hotspot support point first and then the hotspot support point.

[0210] The sampling trigger of the annealing step sequence is defined by the second state identifier to be started when the air path meets the preset stability criterion.

[0211] When the coupling correlation strength fails to satisfy the threshold, the posture jitter amplitude and the jitter rhythm are subjected to preset attenuation adjustment along the coupling priority direction before entering the next annealing step sequence.

[0212] In an embodiment, the coupling mode description quantity consists of the coupling correlation strength and the coupling priority direction. The coupling correlation strength is used to measure the joint significance of the support side state change and the air side state change within the same time window, which can be obtained by weighting the co-occurrence rate of the two state mapping tables, the cross-window consistency, and the low-order statistical correlation, with a value between zero and one, and the greater the value, the stronger the correlation.

[0213] The coupling priority direction is used to represent the precedence relationship between the two types of changes, which can be obtained by the time lag corresponding to the cross-correlation peak of the first state identifier and the second state identifier, and then the lag sign and size are mapped to the positive or negative direction of the posture control axis combined with the geometric mapping of the device.

[0214] For example, if the lag is support leading and the amplitude falls within a short time range, the priority direction is set to a small step amplitude action with a negative pitch. The above two quantities are calculated and archived in each determination window as direct input for parameter selection.

[0215] In a specific implementation, the action primitive library includes at least posture jitter direction, posture jitter amplitude, jitter rhythm, static duration, and locking sequence. The posture jitter direction is consistent with or opposite to the coupling priority direction, and by default it is consistent to reduce the risk of remapping. The posture jitter amplitude and the jitter rhythm are jointly determined by the coupling correlation strength and the air side stability, and when the correlation is strong, a smaller amplitude and a slower rhythm are taken, and when the correlation is weak, a slightly larger amplitude is allowed to speed up the force chain redistribution. The static duration is set to be longer or shorter according to the air side stability, and when the stability is low, a longer duration is taken to wait for the convergence of the air field. The locking sequence identifies the coupling hotspot support point set according to the state mapping table of the first state machine, and the hotspot set is defined as the region where the touch or leave events appear in space. The sequence strategy is to lock the non-hotspots first and then the hotspots to reduce the probability of secondary switching.

[0216] Taking a 3 m vertical spherical mirror as an example, when the lower edge is identified as a hotspot and the air stability is low, the sequence usually selects a small step amplitude jitter with a negative pitch, the amplitude falls within a low end of the order of seconds, the rhythm slows down to sub-second level, the static duration takes several seconds, and the upper edge and the two sides are locked first, and the lower edge is locked later.

[0217] The sampling trigger of the annealing step sequence is defined by the second state identifier as opening when the air path satisfies a preset stability criterion. The stability criterion requires that the air path state is in a stable set and remains for a minimum duration, while the cross-channel consistency and correlation are not lower than a lower limit. The sequence performs one or more samplings within the trigger window, and if the second state identifier switches during this period, the current sample is discarded and the system reverts to the annealing cycle. This gating strategy ensures that the data used for decision-making and quality control is consistent with the stable period of the air field, reducing false positives caused by the flow plume from the source.

[0218] When the coupling correlation strength does not meet the threshold, the amplitude and rhythm of the attitude jitter are adjusted according to the preset attenuation strategy in the coupling priority direction, and then enter the next annealing step sequence. The attenuation strategy requires that the amplitude monotonically decreases until it reaches the allowed lower limit, and the rhythm monotonically slows down or remains unchanged, and the resting time is extended as needed.

[0219] Through the above selection of parameters based on coupling, sampling gated by the second state identifier, and iteration mechanism constrained by the attenuation law, the annealing process can promote the convergence of the two states with small disturbance energy levels under typical field conditions, and provide repeatable preconditions for the initial state confirmation.

[0220] Optionally, after the initial state confirmation signal is generated, the controller closes the write channels of the attitude actuator and the locking actuator, leaving only the read-only monitoring and measurement trigger channels in working condition, while continuously reading the first state identifier and the second state identifier on a unified time axis. If either state identifier switches during sampling, the current sampling process is immediately terminated, the cached measurement frames for this round are discarded, and the system reverts to the initial state establishment process and records the trigger event and timestamp for subsequent parameter tuning and review. This freezing and rollback mechanism ensures the consistency of formal measurement and initial steady state, avoiding the generation of incomparable data during state drift.

[0221] The formal measurement adopts a dual sampling strategy gated by the second state identifier. The system only triggers sampling when the second state identifier indicates that the air path satisfies the preset stability criterion, and obtains the first and second interference measurement frames within adjacent time slots. The span of adjacent time slots and the sampling rhythm are configured according to the interferometer readout speed and the time scale of the air field, forming a tightly coupled frame pair.

[0222] It should be noted that to form a candidate frame pair, the second state identifiers corresponding to the two frames must have a modal coefficient fluctuation rate below the threshold and consistent time sequence codes. If either condition is not met, the system discards the pair and waits for the next stable time slot. This gating and dual sampling design is used to obtain data under similar conditions for two frames during periods of slow air field changes rather than sudden changes, providing stable input for subsequent registration and subtraction.

[0223] The system performs a back-mapping registration based on optical path geometry for the candidate frame pair passing the gating. The registration targets the initial state reference coordinate system, and according to the attitude encoder readings and optical path calibration parameters, the two frame interference phase fields are back-mapped to the reference geometry to eliminate the spatial misplacement caused by the small pitch or azimuth errors. Then, the cross-channel cross-correlation is used to estimate the small time delay and perform the sub-sampling level phase alignment, and if necessary, the linear scale correction is performed according to the reference amplitude, so that the two frames are aligned in space and amplitude. The registration residual and the alignment quality are recorded as quality control quantities, and are written into the metadata for subsequent screening and tracing.

[0224] After the registration is completed, the system performs air component subtraction for the candidate frame pair according to the air disturbance modal coefficients recorded in the second state identifier. Specifically, the two phase fields are projected onto the preset low-order spatial modal basis, the air disturbance dominant modal coefficients and the stability indicators of the corresponding time slots in the second state identifier are read, the air modal components of the two frames are jointly estimated according to the robust weights and are subtracted from the respective phase fields, and the air-subtracted surface-specific wavefront data is obtained. In order to suppress overdrawing or underdrawing, the weights and threshold values can be set to have hysteresis and upper limit clipping, and whether to accept the result is determined jointly by the registration residual and the stability indicator. If the quality control quantity does not meet the standard, the candidate frame pair is discarded and waits for the next stable time slot.

[0225] The surface-specific wavefront data is written into the measurement data buffer together with the timestamp of this measurement, the attitude reading, the locking state snapshot, the environmental parameters, and the corresponding first state identifier and second state identifier summary, and is used as direct input for generating evaluation index data. In order to ensure comparability, the system preferably accumulates several qualified surface-specific wavefront data within a single stable window, performs spatial registration consistency review and robust synthesis, and the synthesis method can be simple average, stability weighted average or robust estimation targeting outlier suppression, and the weights and rejection criteria used are coordinated with the stability indicators of the second state identifier.

[0226] In this way, the synthesis result is used as the surface-specific wavefront representative data under this working condition, and the quality label and confidence indicator are output, providing reliable drafts for subsequent cross-condition stability calculation and report generation.

[0227] The above-mentioned freezing and rollback strategies, gating and dual sampling apertures, back-mapping registration method, air component subtraction weight setting, and quality control threshold values can be adjusted according to the device resolution and the disturbance characteristics on site, and do not constitute a limitation on the present application.

[0228] Based on the same inventive concept, the spherical mirror multi-condition surface stability measurement system corresponding to the spherical mirror multi-condition surface stability measurement method in the embodiments of the present application is also provided. Since the principle of the system in the embodiments of the present application solves the problem similarly to the spherical mirror multi-condition surface stability measurement method in the embodiments of the present application, the implementation of the system can be referred to the implementation of the method, and the repeated parts will not be described again.

[0229] Referring to Figure 4 Fig. 1 is a schematic diagram of a spherical mirror multi-working condition surface shape stability measurement system provided by an embodiment of the present application, which comprises:

[0230] a construction module 10 configured to construct a double-layer state machine, wherein the double-layer state machine comprises a first state machine configured to represent a support contact state of the spherical mirror and a second state machine configured to represent an air path state;

[0231] a collection module 20 configured to collect load sensing data related to the support contact state and optical path monitoring data related to the air path state;

[0232] a processing module 30 configured to generate a first state identifier representing the support contact state based on the load sensing data, generate a second state identifier representing the air path state based on the optical path monitoring data, determine a topology annealing sequence parameter based on the first state identifier and the second state identifier, and execute the topology annealing sequence;

[0233] an updating module 40 configured to update the first state identifier and the second state identifier in real time during the topology annealing sequence process, and generate a starting state confirmation signal when both of them meet their respective preset stability conditions;

[0234] a generation module 50 configured to trigger formal interferometric measurement based on the starting state confirmation signal, obtain spherical mirror surface shape data under a current working condition, and generate evaluation index data representing the spherical mirror multi-working condition surface shape stability based on the surface shape data and the corresponding first state identifier and second state identifier.

[0235] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, but not to limit them; although the present application has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions recorded in the above embodiments, or make equivalent replacement to some or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for measuring the surface stability of a spherical mirror in multiple working conditions, characterized in that, The method comprises the following steps: constructing a double-layer state machine, which comprises a first state machine for representing the contact state of the spherical mirror support and a second state machine for representing the air path state; collecting load sensing data related to the support contact state and optical path monitoring data related to the air path state; generating a first state identifier representing the support contact state based on the load sensing data; generating a second state identifier representing the air path state based on the optical path monitoring data; determining the topology annealing sequence parameters based on the first state identifier and the second state identifier, and executing the topology annealing sequence; updating the first state identifier and the second state identifier in real time during the execution of the topology annealing sequence, and generating a starting state confirmation signal when both of them meet the respective preset stability conditions; triggering a formal interference measurement based on the starting state confirmation signal to obtain the spherical mirror surface data under the current working condition, and generating evaluation index data for representing the spherical mirror multi-working-condition surface stability based on the surface data and the corresponding first state identifier and second state identifier; the generation of the first state identifier representing the support contact state comprises: calculating the load average value, load fluctuation amplitude and load change rate of each support point within a preset time window, and combining the three quantities as the load feature vector of the support point; performing neighborhood normalization processing on each load feature vector according to the spatial distribution of the support points to obtain the relative load feature vector after eliminating the overall temperature drift effect, and calculating the offset from the historical baseline in the relative load feature vector; inputting the offset into a micro-state classification model based on the support point partition, and generating the state mark of the bearing state, the light touch state or the leaving state in the model combined with the displacement response data or the micro-vibration response data of the support point; combining the state marks of all support points to form the first state identifier, and appending the time sequence code of the state switching of each support point in the first state identifier to represent the path dependence characteristics of the support contact topology; the generation of the second state identifier representing the air path state comprises: performing attitude normalization processing on the air disturbance data collected by the optical path peripheral monitoring channel of the interference measurement assembly to obtain normalized air disturbance data; determining the disturbance feature quantity based on the normalized air disturbance data, inputting the disturbance feature quantity into the air path classification model to generate the state mark of the air path, and combining the state mark of the air path to form the second state identifier.

2. The method of claim 1, wherein, The construction of the double-layer state machine comprises: determining the contact micro-state characteristics of each support point based on the output data of the multiple load sensors of the multi-point support mechanism, and performing clustering operation on the contact micro-state characteristics to generate the state identifier of each support point and constitute the state mapping table of the first state machine; determining the disturbance characteristics of the air path based on the output data of the optical path peripheral monitoring channel and the environmental monitoring sensor of the interference measurement assembly, and performing classification operation on the disturbance characteristics to generate the state mapping table of the second state machine; The cross-state index between the first state machine and the second state machine is established to identify the coupling mode description between the support contact state and the air path state.

3. The method of claim 2, wherein, The contact micro-state feature is a feature vector composed of a load fluctuation of the load sensor and a support point displacement response in a continuous time window; The disturbance feature is a modal coefficient fluctuation rate after projecting an air path disturbance in a continuous time window to a preset low-order spatial modal basis; The state mapping table of the first state machine and the second state machine respectively records the state switching sequence, and generates a path history code, and the cross-state index is established by associating the path history codes of the first state machine and the second state machine.

4. The spherically powered multi-condition surface figure stability measurement method of claim 1, wherein, The attitude normalization processing includes performing optical path geometry remapping operation on the air disturbance data based on the optical path geometry offset caused by the change of the spherical mirror pitch or azimuth to eliminate the geometric influence of the attitude change on the air disturbance measurement; The determination of the disturbance feature quantity includes projecting the normalized air disturbance data to a preset low-order spatial modal basis, calculating the fluctuation rate and phase consistency index of each modal coefficient in a preset time window, and combining the fluctuation rate and the phase consistency index as the disturbance feature quantity; The second state identifier further includes a time sequence code of the air path state switching to represent the historical dependence feature of the air path disturbance.

5. The spherically powered multi-condition surface figure stability measurement method of claim 2, wherein, Based on the first state identifier and the second state identifier, the topology annealing sequence parameters are determined, and the topology annealing sequence is executed, which includes: reading the coupling mode entry corresponding to the current time window from the cross-state index to generate a coupling mode description quantity for representing the coupling relationship between the support contact state and the air path state; selecting and parameterizing the annealing action primitive in the controller based on the coupling mode description quantity to obtain a topology annealing control vector; generating a set of topology annealing sequence parameters according to the topology annealing control vector, and driving the annealing step sequence including unlocking, micro-attitude disturbance, static and locking; After the execution of the annealing step sequence ends, the first state identifier and the second state identifier are updated, and a preset annealing termination condition is determined. If it is not met, the topology annealing control vector is regenerated according to the updated coupling mode description quantity and repeated execution.

6. The spherically powered multi-condition surface figure stability measurement method of claim 5, wherein, The coupling mode description quantity includes coupling correlation strength and coupling priority direction; The annealing action primitive at least includes attitude jitter direction, attitude jitter amplitude, jitter rhythm, static time length and locking sequence, and the locking sequence is determined as first non-hot spot support point and then hot spot support point by identifying the coupling hot spot support point set through the state mapping table of the first state machine; The sampling trigger of the annealing step sequence is limited to be started when the air path meets the preset stability criterion according to the second state identifier; When the coupling correlation strength does not meet the threshold value, the attitude jitter amplitude and the jitter rhythm are adjusted after a preset attenuation adjustment along the coupling priority direction, and then the next annealing step sequence is entered.

7. The spherically powered multi-condition surface figure stability measurement method of claim 1, wherein, The formal interference measurement triggered based on the starting state confirmation signal includes: Freezing the control channels of the posture actuator and the locking actuator after the initial state confirmation signal is generated, and continuously reading the first state identifier and the second state identifier during sampling, and stopping sampling and returning to the initial state establishment process when any state identifier is detected to switch; Using the second state identifier as a sampling gate, acquiring a first interferometric measurement frame and a second interferometric measurement frame in adjacent time slots using a dual sampling strategy, requiring that the modal coefficient fluctuation rate of the second state identifier be lower than a preset threshold and that its time sequence code be consistent, to form a candidate frame pair; Performing a back mapping registration operation based on optical path geometry on the candidate frame pair to unify the two frames of interferometric measurement data to the initial state reference coordinate system; Based on the air disturbance modal coefficient recorded by the second state identifier, performing air component subtraction processing on the interferometric measurement data of the candidate frame pair to obtain face-specific wavefront data, and writing the face-specific wavefront data as input to generate evaluation index data into the measurement data buffer.

8. A system for measuring the stability of the surface of a spherical mirror in multiple working conditions, characterized in that it comprises: Comprise: A construction module for constructing a double-layer state machine, the double-layer state machine comprising a first state machine for representing a spherical mirror support contact state and a second state machine for representing an air path state; An acquisition module for acquiring load sensing data related to the support contact state and optical path monitoring data related to the air path state; A processing module for generating a first state identifier representing the support contact state based on the load sensing data; Generating a second state identifier representing the air path state based on the optical path monitoring data; Determining topology annealing sequence parameters based on the first state identifier and the second state identifier, and performing a topology annealing sequence; The first state identifier representing the support contact state comprises: calculating the load average, load fluctuation amplitude and load change rate of the load sensing data of each support point within a preset time window, and combining the three quantities as the load feature vector of the support point; Normalizing each load feature vector according to the spatial distribution of the support points to obtain a relative load feature vector that eliminates the overall temperature drift effect, and calculating the offset from the historical baseline in the relative load feature vector; Inputting the offset into a micro-state classification model based on support point partitioning, and generating a state marker of a bearing state, a touch state or an away state in the model combined with displacement response data or micro-vibration response data of the support point; Combining the state markers of all support points to form the first state identifier, and appending the time sequence code of each support point state switch in the first state identifier to represent the path-dependent characteristics of the support contact topology; The second state identifier representing the air path state comprises: performing posture normalization processing on the air disturbance data collected by the optical path peripheral monitoring channel of the interferometric measurement assembly to obtain normalized air disturbance data; Based on the normalized air disturbance data, determine the disturbance feature quantity, and input the disturbance feature quantity into the air path classification model to generate a state marker of the air path; combine the state marker of the air path to form the second state identifier; an updating module, configured to update the first state identifier and the second state identifier in real time during the topological annealing sequence process, and generate a start state confirmation signal when the two identifiers both satisfy respective preset stability conditions; a generating module, configured to trigger formal interferometric measurement based on the start state confirmation signal, to obtain spherical mirror surface data under a current working condition, and to generate evaluation index data for characterizing the spherical mirror multi-working-condition surface stability based on the surface data and the corresponding first state identifier and second state identifier.

Images