A method for determining TOPCon film parameters based on spectral response
By using module-level spectral data packages and albedo spectral templates, a back-side local film layer parameter model was established, which solved the problem of the difficulty in uniquely determining the back-side local film layer parameters of TOPCon modules under high albedo bearing surfaces, and achieved accurate film layer parameter differentiation and process backtracking.
Patent Information
- Application Number
- CN202611275523.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-21
- Publication Date
- 2026-11-17
AI Technical Summary
Under high albedo bearing surface conditions, existing technologies struggle to accurately distinguish the local film layer parameters on the back side of bifacial TOPCon modules, affecting the uniqueness of the film layer parameter determination results and the basis for process backtracking.
By acquiring component-level spectral data packages and albedo spectral templates, a backside local film parameter model is established to distinguish film parameters in metallized and non-metallized regions. The parameters are then matched and determined based on spectral response characteristics, outputting a unique combination of film parameters or identifying the source of conflict.
It enables accurate differentiation of local film parameters on the back side of bifacial TOPCon modules in high albedo scenarios, reduces the mixed influence of equivalent film parameters on local film differences, and improves the uniqueness of film parameter determination and the reliability of process backtracking.
Smart Images

Figure CN123124760A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of photovoltaic module testing and film parameter inversion technology, and in particular to a method for determining TOPCon film parameters based on spectral response. Background Technology
[0002] TOPCon cells form a passivated contact structure through a tunneling oxide layer and a doped poly-Si passivated contact layer, and combine this with an antireflective coating and metallized electrodes to achieve light absorption and current collection. When bifacial TOPCon modules are applied to surfaces with high albedo, such as roofs, concrete floors, snowfields, and light-colored surfaces, the back side of the module receives back-incident light reflected from the surface. The spectral distribution and incident angle of this back-incident light vary with the type of surface, surface condition, and installation conditions, causing the back poly-Si passivated contact layer, the back antireflective coating, and the back metallized region to exhibit band-specific differences in their influence on the spectral response.
[0003] Existing methods for determining TOPCon film parameters typically employ reflectance spectral fitting, ellipsometry measurement, film color comparison, standard sheet comparison, test sheet sampling, or process formulation retrospection to assess poly-Si layer thickness, antireflection film thickness, film uniformity, or deposition process deviations. These methods often output parameters of the entire film layer or equivalent film layer, making them suitable for determining whether film deposition deviates from the process range.
[0004] In back-side localized or patterned poly-Si structures, the metallized and non-metallized regions exhibit different poly-Si retention states, local thicknesses, parasitic absorptions, and contact interactions. If the film parameters are determined based on a uniform film layer across the entire surface, the local differences in the back-side reflectance spectrum and back-side spectral response data will be factored into the same equivalent parameters, resulting in a mixture of differences between the poly-Si contact retention states in the metallized regions, the poly-Si thinning states in the non-metallized regions, and the back-side antireflection film states.
[0005] Meanwhile, the back-side incident spectrum formed by a high-albedo bearing surface is not a fixed standard spectrum. Different bearing surfaces will change the response weights of relevant bands on the back side, causing the same film layer state to exhibit different back-side response differences under different bearing surface conditions. When using a single standard spectrum or a single equivalent film layer result for judgment, changes in local back-side film layer parameters, back-side antireflective film deviation, front-side response deviation, and albedo spectrum changes are prone to overlap in the matching results, thus affecting the uniqueness of the film layer parameter determination results and the basis for process backtracking.
[0006] Therefore, it is urgent to solve the technical problem of how to distinguish and determine the local film parameters on the back side of bifacial TOPCon modules based on the front and back spectral responses, back reflection information, and the albedo spectral state of the bearing surface under high albedo bearing surface conditions. Summary of the Invention
[0007] This invention provides a method for determining TOPCon film parameters based on spectral response, solving the problem that it is difficult to uniquely determine the parameters of the local back-side film of TOPCon in high-albedo scenarios.
[0008] This invention provides a method for determining TOPCon film parameters based on spectral response, comprising:
[0009] Obtain the component-level spectral data package of the double-sided TOPCon component under test and the albedo spectral template corresponding to the high albedo bearing surface condition and with scene tags. The component-level spectral data package includes front spectral response data, back spectral response data and back reflectance spectral data. Generate back weighted spectral response features from the back spectral response data, back reflectance spectral data and albedo spectral template.
[0010] A parameter model for the back local film layer is established. The model divides the effective light-receiving area on the back into a metallized region and a non-metallized region, and associates the poly-Si equivalent parameters of the metallized region and the non-metallized region, the equivalent parameters of the back antireflection film, and the region ratio parameters.
[0011] Based on the model, candidate film layer parameter combinations, candidate front spectral response features, and candidate back spectral response features are generated. The candidate front spectral response features are matched with the front spectral response data by band, and the candidate back spectral response features are weighted by an albedo spectral template and matched with the back weighted spectral response features by band to obtain the matching results.
[0012] The matching results generate a matching level, candidate matching gap, scenario consistency level, and attribution identifier, which are used to determine uniqueness. When the uniqueness determination condition is met, the target membrane parameter combination and parameter credibility identifier are output; when the condition is not met, the parameter non-unique determination identifier and conflict source identifier are output.
[0013] In some embodiments, the component-level spectral data package further includes frontal reflectance spectral data, and the component-level spectral data package is generated through homogenization processing, which includes:
[0014] Map the front spectral response data, back spectral response data, front reflectance data, and back reflectance data to the same band identification system;
[0015] Establish data association based on the component identifier, test direction identifier, and test batch identifier of the double-sided TOPCon component under test;
[0016] For each band identifier, generate a frontal response sequence, a backal response sequence, a frontal reflection sequence, and a backal reflection sequence;
[0017] Set a missing measurement marker for the missing band and use the missing measurement marker as a shielding condition for band matching.
[0018] In some embodiments, obtaining the albedo spectral template includes:
[0019] Obtain bearing surface category information, which includes at least one of roof surface category, cement surface category, light-colored ground category, and snow surface category;
[0020] The data sources for the albedo spectral template include test data of the bearing surface material, component installation scene records, back irradiation test data, and preset bearing surface category data;
[0021] The bearing surface category is marked according to the material type, surface color, and surface condition, while the incident angle category is marked according to the component installation tilt angle, component height from the bearing surface, bearing surface category, and back test angle.
[0022] In some embodiments, obtaining the albedo spectral template further includes:
[0023] Read the band albedo sequence from the albedo spectral template library according to the aforementioned bearing surface category information;
[0024] The incident angle category is labeled and normalized on the albedo sequence of the band to generate normalized albedo spectral values and normalized reference intensity, forming an albedo spectral template with scene tags; the normalization mode identifier is used to record the intensity reference used for the normalized albedo spectral values, and the scene tag is generated by combining the bearing surface category, incident angle category and normalization mode identifier.
[0025] When the same double-sided TOPCon component under test corresponds to multiple scene labels, the back-side weighted spectral response features under each scene label are generated respectively.
[0026] In some embodiments, the backside local membrane parameter model includes a region parameter layer, a membrane parameter layer, and a response mapping layer; the region parameter layer records metallized region identifiers, non-metallized region identifiers, and region ratio parameters;
[0027] The film parameter layer records the equivalent thickness parameters of poly-Si in the metallized region, the contact retention parameters of poly-Si in the metallized region, the equivalent thickness parameters of poly-Si in the non-metallized region, the absorption category of poly-Si in the non-metallized region, and the equivalent parameters of the antireflective film on the back side.
[0028] The response mapping layer maps the parameter combinations in the region parameter layer and the film parameter layer into front response contribution, back response contribution, and front-back differential response contribution.
[0029] In some embodiments, the metallized region is the region on the back side that forms an electrical connection with a metal electrode, and the non-metallized region is the region on the back side effective light-receiving region other than the metallized region;
[0030] The region ratio parameter is determined by the back metallization pattern, the area of the effective light-receiving area on the back, and the area of the non-metallized area.
[0031] The poly-Si contact retention parameters of the metallized region are obtained by associating the equivalent thickness parameters of the poly-Si metallized region, the contact resistance category, and the annealing batch identifier.
[0032] The absorption category of the nonmetallized poly-Si region is obtained by correlating the equivalent thickness parameter of the nonmetallized poly-Si region with the absorption intensity of the corresponding band.
[0033] In some embodiments, generating multiple candidate membrane parameter combinations includes:
[0034] Read process constraint information, which includes at least one of the following: back poly-Si deposition process identifier, back poly-Si thinning process identifier, back localization process identifier, back anti-reflection film deposition process identifier, and metallization contact process identifier;
[0035] The range of candidate parameter values is determined based on the process constraint information;
[0036] Candidate membrane parameter combinations are generated by combining candidate parameters within their range; candidate membrane parameter combinations that conflict with process constraint information are eliminated to form a set of candidate membrane parameters to be matched.
[0037] In some embodiments, eliminating candidate film parameter combinations that conflict with process constraint information includes:
[0038] Eliminate combinations where the equivalent thickness parameter of the non-metallized region poly-Si does not match the poly-Si thinning process markings on the back side.
[0039] Eliminate combinations where the regional proportion parameters do not match the localized processing technology markings on the back;
[0040] Eliminate combinations where the poly-Si contact retention parameters in the metallized region do not match the metallized contact process specifications; eliminate combinations where the equivalent parameters of the back antireflection film do not match the back antireflection film deposition process specifications.
[0041] In some embodiments, band-specific matching includes:
[0042] According to the preset band division rules, the results of front spectral response data, back weighted spectral response features, candidate front spectral response features, and candidate back spectral response features after being weighted by the albedo spectral template are divided into multiple band response segments.
[0043] For each band response segment, generate a band deviation data item between the measured response and the candidate response;
[0044] The scene labels and missing data markers of the front band deviation data items, back band deviation data items, and albedo spectral templates are combined into the matching results;
[0045] The matching results retain the band source identifier and scene label.
[0046] In some embodiments, after outputting the target film parameter combination and parameter credibility identifier, the target film parameter combination, parameter credibility identifier, scene label of the albedo spectral template, matching result and attribution identifier are written into the parameter record of the double-sided TOPCon component under test.
[0047] Process backtracking labels are generated based on the target film layer parameter combination. The process backtracking labels include at least one of the following: non-metallized region poly-Si thinning bias label, metallized region poly-Si retention bias label, back antireflection film bias label, and albedo spectral scene adaptation label.
[0048] When the output parameter cannot be uniquely identified, the conflict source identifier is written into the parameter record, and the parameter record is set to the verification state.
[0049] Through the above technical solution, the present invention can achieve at least the following beneficial effects:
[0050] This invention generates back-side weighted spectral response features using component-level spectral data packages and albedo spectral templates with scene tags. It also distinguishes film parameters in metallized and non-metallized regions using a back-side local film parameter model, enabling candidate film parameter combinations to participate in uniqueness determination under frontal spectral response, back-side weighted spectral response, and scene consistency level. Candidate matching gaps are used to differentiate the matching differences between optimal and suboptimal candidate film parameter combinations. Back-side local poly-Si parameter differences, back-side antireflective film parameter differences, and carrier surface albedo spectral differences are respectively included in the matching results and attribution identifiers. The mixing of local film differences by the overall equivalent film parameters is reduced, and the identifier for target film parameter combinations or parameters that cannot be uniquely determined can be output.
[0051] By homogenizing the component-level spectral data packets, the front spectral response data, back spectral response data, front reflectance data, and back reflectance data are mapped to the same band identification system. Missing bands are marked with missing indicators and used as masking conditions for sub-band matching, so that the input data for sub-band matching maintains the same band basis.
[0052] Scene labels are generated by identifying the bearing surface type, incident angle type, and normalization mode. Albedo spectral templates under different high albedo bearing surface conditions are distinguished and recorded. Backside weighted spectral response features under each scene label are generated separately, so that the source of albedo spectral changes is preserved in the matching results.
[0053] A back-side local film parameter model is established by using a region parameter layer, a film parameter layer, and a response mapping layer. The poly-Si contact retention parameters in the metallized region, the poly-Si absorption category in the non-metallized region, the equivalent parameters of the back-side antireflection film, and the region ratio parameters are jointly mapped to the front response contribution, the back response contribution, and the front-back differential response contribution, thus establishing a correspondence between the candidate film parameter combinations and the front-back response differences.
[0054] By eliminating candidate film parameter combinations that conflict with process identifiers through process constraint information, the candidate film parameter combinations that enter the band matching process correspond to the back poly-Si deposition, back poly-Si thinning, back localization treatment, back antireflection film deposition and metallization contact process records, and the candidate results that reduce process state conflicts are entered into the uniqueness determination.
[0055] By writing the target membrane parameter combination, parameter credibility identifier, matching result, attribution identifier, and process backtracking label into the parameter record, the membrane parameter determination result, albedo spectral scene, and source of response difference are stored together; when the output parameter cannot be uniquely identified, the conflict source identifier is written into the parameter record and a verification status is formed. Attached Figure Description
[0056] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope of the present invention.
[0057] Figure 1 This is a flowchart of the TOPCon film parameter determination method based on spectral response in the embodiments. Detailed Implementation
[0058] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.
[0059] All terms used in this invention (including technical and scientific terms) have the meanings commonly understood by those skilled in the art, unless otherwise defined.
[0060] To facilitate understanding, the relevant terms and concepts involved in the embodiments of this application will be introduced below.
[0061] The high-reflectivity carrier surface condition refers to the application condition formed by the band albedo characteristics, incident angle category, and scene label of the carrier surface on the back of the bifacial TOPCon module under test. Spectral response data refers to the band-specific electrical response data formed by the bifacial TOPCon module under test under preset band markings. The front and back spectral response data are obtained from the front and back test directions of the module, respectively. Reflectance spectrum data refers to the band-specific reflection intensity data formed by the bifacial TOPCon module under test under preset band markings. The front and back reflection spectrum data are obtained from the front and back test directions of the module, respectively. The effective back-side light-receiving area refers to the area on the back of the bifacial TOPCon module under test that participates in the modeling of the back-side local film layer parameters. It includes the metal electrode projection area and the non-metallized light-receiving opening area, but excludes the module frame shading area. The metal electrode projection area is classified as the metallized area, and the non-metallized light-receiving opening area is classified as the non-metallized area. Its boundary is jointly determined by the back metallization pattern, the effective test boundary of the module, and the shading area record.
[0062] A component-level spectral data package refers to a data set associated with the component identifier of the same bifacial TOPCon component under test. It includes at least front spectral response data, back spectral response data, back reflectance spectrum data, test direction identifier, test batch identifier, and band identifier. When generating a component-level spectral data package through homogenization processing, the component-level spectral data package also includes front reflectance spectrum data. An albedo spectral template is a data template used to characterize the back incident spectral distribution formed after the carrier surface reflects incident light in a band. It includes carrier surface category, band albedo sequence, incident angle category, normalization mode identifier, and scene label.
[0063] Specifically, each data item in the component-level spectral data package uses component identifier, test batch identifier, test direction identifier, and band identifier as alignment fields. The test batch identifier is used to record the batch affiliation of the same spectral response test and reflectance spectrum test. The test direction identifier is used to distinguish between the front test direction and the back test direction. The band identifier is used to map the band data output by different test equipment to the same band position. If the spectral response data and reflectance spectrum data under the same band identifier come from different acquisition times, the consistency of the test batch identifier and the consistency of the band identifier are used as the association condition, and the associated data is written into the same component-level spectral data package.
[0064] Backside weighted spectral response characteristics refer to the band-specific response characteristics formed by backside spectral response data under the joint constraints of the albedo spectral template and backside reflectance spectral data. Backside localized film parameter model refers to a data processing model that distinguishes the effective light-receiving area on the backside as metallized and non-metallized regions, and establishes a mapping relationship between regional parameters, film parameters, and response contributions.
[0065] Example 1:
[0066] like Figure 1 As shown, this embodiment employs a method for determining TOPCon film parameters based on spectral response, including:
[0067] Step S1: Obtain the component-level spectral data package of the double-sided TOPCon component under test and the albedo spectral template corresponding to the high albedo bearing surface condition and with scene tags. The component-level spectral data package includes front spectral response data, back spectral response data and back reflectance spectrum data.
[0068] Step S2: Generate back-side weighted spectral response features based on the back-side spectral response data, the back-side reflectance spectral data, and the albedo spectral template;
[0069] Step S3: Establish a back-side local film parameter model. The back-side local film parameter model divides the effective light-receiving area on the back side into a metallized region and a non-metallized region, and associates the equivalent parameters of poly-Si in the metallized region, the equivalent parameters of poly-Si in the non-metallized region, the equivalent parameters of the back-side antireflection film, and the region ratio parameters.
[0070] Step S4: Generate multiple candidate film parameter combinations based on the back local film parameter model, and generate candidate front spectral response features and candidate back spectral response features corresponding to each candidate film parameter combination.
[0071] Step S5: Perform band-wise matching between the candidate frontal spectral response features and the frontal spectral response data, and perform band-wise matching between the candidate back spectral response features weighted by the albedo spectral template and the back weighted spectral response features to obtain the matching result;
[0072] Step S6: Generate a matching level, candidate matching gap, scene consistency level, and attribution identifier based on the matching results, and make a determination based on the uniqueness determination criteria jointly determined by the matching level, candidate matching gap, scene consistency level, and attribution identifier; wherein, the candidate matching gap is used to characterize the matching difference between the first-ranked candidate membrane parameter combination and the second-ranked candidate membrane parameter combination, and the scene consistency level is used to characterize the consistency status of the matching results of the same candidate membrane parameter combination under different scene labels;
[0073] When the uniqueness determination condition is met, the target film layer parameter combination and parameter reliability identifier are output; when the uniqueness determination condition is not met, the parameter non-unique determination identifier and conflict source identifier are output.
[0074] In one implementation, the process of generating back-side weighted spectral response features includes: reading the back-side response value from the back-side spectral response data, the back-side reflectance value from the back-side reflectance spectral data, and the band albedo value from the albedo spectral template according to the band identifier; obtaining the albedo spectral band weight by normalizing the band albedo value; obtaining the back-side effective response correction amount by mapping the back-side reflectance value to the reflectance correction category in the calibration sample library; and associating the back-side response value, the albedo spectral band weight, and the back-side effective response correction amount as a back-side weighted response data item. The reflectance correction category is the reflectance intensity category to which the back-side reflectance value belongs under the corresponding band identifier, used to map the back-side reflectance value to the back-side effective response correction amount. The back-side weighted response data item includes the band identifier, scene label, back-side response value, back-side reflectance value, band albedo value, and back-side weighted response value; when any band lacks a back-side response value, back-side reflectance value, or band albedo value, the band is written with a missing measurement identifier and treated as a masked band in the sub-band matching.
[0075] Based on the above implementation, the candidate matching gap is jointly determined by the front band deviation data item, the back band deviation data item, and the albedo scene deviation data item of the candidate film parameter combination. The attribution identifier is determined by the dominant deviation data item among the front band deviation data item, the back band deviation data item, and the albedo scene deviation data item, and includes the back local poly-Si attribution identifier, the back antireflection film attribution identifier, the front response attribution identifier, and the albedo spectral scene attribution identifier. Uniqueness determination is performed in the order of matching level, candidate matching gap, scene consistency level, and attribution identifier; when the candidate film parameter combination ranked first meets the matching level, and the candidate matching gap between it and the candidate film parameter combination ranked second reaches the distinguishing boundary given by the calibration sample library, the candidate matching gap determination item is recorded as satisfied. When only one candidate membrane parameter combination simultaneously satisfies the corresponding conditions of matching level, candidate matching gap, scenario consistency level, and attribution identifier, the candidate membrane parameter combination is determined as the target membrane parameter combination, and a parameter trust identifier is generated; when there are two or more candidate membrane parameter combinations simultaneously satisfying the above conditions, or when at least one of the above conditions is not generated, a parameter non-unique identifier and a conflict source identifier are generated.
[0076] In one embodiment, the component-level spectral data package is generated through homogenization processing, which includes: mapping the front spectral response data, back spectral response data, front reflectance spectrum data, and back reflectance spectrum data to the same band identification system; establishing data association according to the component identification, test direction identification, and test batch identification of the bifacial TOPCon component under test; generating a front response sequence, a back response sequence, a front reflectance sequence, and a back reflectance sequence for each band identification; setting a missing detection identification for missing bands, and using the missing detection identification as a shielding condition for band-specific matching.
[0077] In one embodiment, the backside local membrane parameter model includes a region parameter layer, a membrane parameter layer, and a response mapping layer; the region parameter layer records metallized region identifiers, non-metallized region identifiers, and region ratio parameters.
[0078] The film parameter layer records the equivalent thickness parameters of poly-Si in the metallized region, the contact retention parameters of poly-Si in the metallized region, the equivalent thickness parameters of poly-Si in the non-metallized region, the absorption category of poly-Si in the non-metallized region, and the equivalent parameters of the antireflective film on the back side.
[0079] The response mapping layer maps the parameter combinations in the region parameter layer and the film parameter layer into front response contribution, back response contribution, and front-back differential response contribution.
[0080] The back-side local film parameter model was established using a calibration sample library. The data sources for this library included mass-produced component spectral test data, standard sample film test data, back-side layout data, and process batch records. Labeling methods included film thickness category labeling, contact retention category labeling, absorption intensity category labeling, anti-reflective film state category labeling, and region proportion category labeling. Training target categories included front-side response contribution categories, back-side response contribution categories, and front-back differential response contribution categories. The response mapping layer established a mapping relationship between candidate film parameter combinations and band-specific response contributions according to the parameter category boundaries in the calibration sample library. For each candidate film parameter combination, the band-specific contact response contribution corresponding to the poly-Si equivalent parameters of the metallized region, the band-specific absorption response contribution corresponding to the poly-Si equivalent parameters of the non-metallized region, and the band-specific reflection response contribution corresponding to the back-side anti-reflective film equivalent parameters were determined, and these were synthesized into candidate back-side spectral response features according to the region proportion parameters. Candidate front-side spectral response features were jointly determined by the front-side response contribution and the front-back differential response contribution. During deployment, the back-side local membrane parameter model adopts the same band identification system and parameter category system as the calibration sample library. It receives component-level spectral data packets, albedo spectral templates, and candidate membrane parameter combinations, and outputs candidate front-side spectral response features, candidate back-side spectral response features, and corresponding response contribution categories. If there are missing bands in the input data, they are masked with missing measurement identifiers, and candidate front-side spectral response features and candidate back-side spectral response features are generated with aligned bands.
[0081] In one embodiment, the metallized region is the region on the back side that forms an electrical connection with the metal electrode, and the non-metallized region is the region on the back side effective light-receiving region excluding the metallized region; the region ratio parameter is determined by the back side metallization pattern, the area of the back side effective light-receiving region, and the area of the non-metallized region; the poly-Si contact retention parameter of the metallized region is obtained by associating the poly-Si equivalent thickness parameter of the metallized region, the contact resistance category, and the annealing batch identifier; the poly-Si absorption category of the non-metallized region is obtained by associating the poly-Si equivalent thickness parameter of the non-metallized region and the absorption intensity of the corresponding band.
[0082] Specifically, the back metallization pattern refers to the data record that documents the location of the back metal electrodes, the boundary of the metallized area, and the boundary of the effective light-receiving area on the back. The contact resistance category refers to the category identifier corresponding to the electrical connection state of the metallized area, which is mapped from metallization contact test data or process batch records. The annealing batch identifier refers to the heat treatment batch record corresponding to the formation process of the poly-Si contact on the back of the bifacial TOPCon module under test. The corresponding band absorption intensity refers to the absorption intensity category of the non-metallized poly-Si area under the specified band, which is obtained from the absorption category labeling in the standard sample film layer test data, mass-produced module spectral test data, or calibration sample library.
[0083] The equivalent parameters of the poly-Si metallized region include the equivalent thickness parameter and the contact retention parameter. The equivalent parameters of the poly-Si non-metallized region include the equivalent thickness parameter and the absorption category. The equivalent parameters of the back antireflection coating include the equivalent thickness category, the equivalent refractive category, and the band reflection category. The region proportion parameters include the proportion of metallized region, the proportion of non-metallized region, and the localization coverage category. These parameters are stored in association with the same component identifier.
[0084] In one embodiment, generating multiple candidate film layer parameter combinations includes: reading process constraint information, the process constraint information including at least one of back poly-Si deposition process identifier, back poly-Si thinning process identifier, back localization process identifier, back anti-reflection film deposition process identifier, and metallization contact process identifier; determining the value range of candidate parameters based on the process constraint information; generating candidate film layer parameter combinations within the candidate parameter value range; and eliminating candidate film layer parameter combinations that conflict with the process constraint information to form a set of candidate film layer parameters to be matched.
[0085] The set of candidate film parameters to be matched includes multiple combinations of candidate film parameters. Each candidate film parameter combination includes a combination identifier, equivalent poly-Si parameters for the metallized region, equivalent poly-Si parameters for the non-metallized region, equivalent parameters for the backside antireflection film, region ratio parameters, and a process constraint matching identifier. The process constraint matching identifier is used to record the correspondence between the candidate film parameter combination and the process constraint information, including a matching state, a mismatch state, and a pending verification state. Candidate film parameter combinations in a mismatch state are removed from the set of candidate film parameters to be matched, while candidate film parameter combinations in a pending verification state are retained and their pending verification identifier is written into the matching results.
[0086] In one implementation, eliminating candidate film parameter combinations that conflict with process constraint information includes: eliminating combinations where the equivalent thickness parameter of the non-metallized region poly-Si does not match the back poly-Si thinning process identifier; eliminating combinations where the region ratio parameter does not match the back localization process identifier; eliminating combinations where the poly-Si contact retention parameter of the metallized region does not match the metallized contact process identifier; and eliminating combinations where the equivalent parameter of the back antireflection film does not match the back antireflection film deposition process identifier.
[0087] In one embodiment, band matching includes: dividing the front spectral response data, back-side weighted spectral response features, candidate front spectral response features, and candidate back-side spectral response features into multiple band response segments according to a preset band division rule; generating band deviation data items between the measured response and the candidate response for each band response segment; combining the front band deviation data items, back band deviation data items, scene labels of the albedo spectral template, and missing measurement identifiers into a matching result; the matching result retains the band source identifier and scene label.
[0088] The matching results include a combination identifier, band identifier, band source identifier, scene label, front band deviation data item, back band deviation data item, albedo scene deviation data item, and missing measurement identifier. The band deviation data item records the difference between the measured response and the candidate response within the same band response segment. The band source identifier is used to distinguish front spectral response data, back weighted spectral response features, candidate front spectral response features, and candidate back spectral response features.
[0089] Matching levels include satisfying matching level, not satisfying matching level, and restricted matching level; the candidate matching gap is generated based on the difference in matching level between the first-ranked candidate film parameter combination and the second-ranked candidate film parameter combination, as well as the difference in band deviation data items; when all deviation data items involved in the matching meet the matching boundary given by the calibration sample library, a satisfying matching level is generated; when at least one deviation data item involved in the matching exceeds the matching boundary, a not satisfying matching level is generated; the restricted matching level is used to record the status of missing measurement identifiers, albedo spectral templates marked as restricted, low-weight restricted scene labels, or version conflict source identifiers affecting the complete generation of the matching level.
[0090] In one implementation, after outputting the target film parameter combination and parameter credibility identifier, the target film parameter combination, parameter credibility identifier, scene label of the albedo spectral template, matching result, and attribution identifier are written into the parameter record of the double-sided TOPCon component under test; a process backtracking label is generated based on the target film parameter combination, the process backtracking label including at least one of the following: non-metallized region poly-Si thinning bias label, metallized region poly-Si retention bias label, back antireflection film bias label, and albedo spectral scene adaptation label; when the output parameter cannot uniquely determine the identifier, the conflict source identifier is written into the parameter record, and the parameter record is set to a verification state.
[0091] The parameter record includes component identifier, test batch identifier, target membrane parameter combination, parameter reliability identifier, parameter non-uniquely determinable identifier, conflict source identifier, scene label, matching result, attribution identifier, process backtracking label, and review status. Conflict source identifiers include positive response conflict identifiers, negative response conflict identifiers, albedo spectral scene conflict identifiers, and process constraint conflict identifiers. Version conflict source identifiers, low-weight constraint identifiers, differential response constraint identifiers, scene weight interpretation constraint identifiers, and multi-scene consistency constraint identifiers are written as subfields into the parameter record. The process backtracking label is generated jointly by the target membrane parameter combination and the attribution identifier, and is stored in association with the corresponding process batch record.
[0092] For example, the parameter trust identifier is used to record the uniqueness determination status of the target membrane parameter combination. Its field values include trust status and restricted trust status. The trust status indicates that the target membrane parameter combination is determined by candidate membrane parameter combinations that meet the matching level, candidate matching gap, scene consistency level, and attribution identifier. The restricted trust status indicates that the target membrane parameter combination has been determined, but there are missing measurement identifiers, albedo spectral templates marked as restricted, low-weight restricted scene labels, or differential response restricted identifiers. When the output parameter cannot be uniquely determined identifier is present, the verification status record in the parameter record is recorded as pending verification and stored in association with the parameter cannot be uniquely determined identifier and conflict source identifier.
[0093] In a preferred embodiment of Example 1, when determining the scene consistency level of the matching results of the same candidate film layer parameter combination under multiple scene labels, the back band deviation data item, front and back differential response contribution and attribution identifier formed by the same candidate film layer parameter combination under each scene label are statistically analyzed in the same group.
[0094] The same-group statistics take the candidate film parameter combination as a fixed object, and the scene label is only used as the source of change of the albedo spectral weight and the effective band set. The equivalent parameters of poly-Si in the metallized region, the equivalent parameters of poly-Si in the non-metallized region, the equivalent parameters of the back antireflection film, and the regional ratio parameters in the candidate film parameter combination remain fixed in the same-group statistics process.
[0095] Before generating the intrinsic backside deviation, albedo spectral band weights are generated based on the normalized albedo spectral values and bandwidths:
[0096] ,
[0097] in, For the first Under the first scene tag The weight of the albedo spectral band corresponding to each band identifier; For the first Under the first scene tag The normalized albedo spectral values corresponding to each band identifier; For the first Each band identifier corresponds to a band width; This refers to the band summation sequence number in the calculation of the weights of the albedo spectrum bands. For the first Under the first scene tag The normalized albedo spectral values corresponding to each band identifier; For the first Each band identifier corresponds to a band width; For the first The set of valid bands under each scene label that are not blocked by the missing measurement marker; To prevent zero bias in the weighted normalization of the albedo spectral bands; and Following the previous definition, for band identifiers that do not fall into the valid band set, the corresponding albedo spectral band weights are reset to 0.
[0098] In determining the scene consistency level, a lower limit for the weight of the albedo spectral band is set. , The minimum albedo spectral band weight allowed to participate in cross-scene statistics during scene consistency level determination is determined by the upper quantile value of the albedo spectral band weight corresponding to the noise floor in the calibration component samples. When At that time, the first Under the first scene tag Each band identifier is retained only in the albedo spectral template and is not involved in the calculation of cross-scene intrinsic backface deviation reference value, cross-scene front and backface differential response contribution reference value, multi-scene response consistency deviation, and scene weight interpretation deviation. The corresponding valid band identifier value is then used. Set to 0. If the same scene label satisfies If the number of bands is lower than the minimum band requirement, the scene label will be marked as a low-weight restricted scene label in the scene consistency level determination.
[0099] To distinguish between the band weight changes caused by scene labels and the response differences caused by film parameters, the back-side band deviation data item is first converted into intrinsic back-side deviation:
[0100] ,
[0101] in, For the first The candidate film layer parameter combinations in the first Under the first scene tag The intrinsic backside deviation corresponding to each band identifier; The sequence number of the candidate film layer parameter combination; For scene label sequence numbers; For band identification serial number; For the first The candidate film layer parameter combinations in the first Under the first scene tag Each band identifier corresponds to a back band deviation data item, which is obtained by the difference between the back weighted spectral response features and the candidate back spectral response features after weighting by an albedo spectral template. For the first Under the first scene tag The weight of the albedo spectral band corresponding to each band identifier; To prevent zero bias in band weighting.
[0102] Then, under the same candidate film layer parameter combination, cross-scene reference values are generated for the intrinsic back face deviation and front-back differential response contributions in multiple scene labels:
[0103] ; ;
[0104] in, For the first The candidate film layer parameter combinations in the first Each band identifier corresponds to the cross-scenario intrinsic backside deviation reference value; For the first A set of scene labels for determining the scene consistency level using combinations of candidate membrane parameters; For the first Under the first scene tag The scene consistency level determination for each band identifier is a valid band identifier value, provided that the band is not blocked by the missing measurement identifier. Greater than or equal to And the version binding field corresponding to this scene label The version binding field is set to 1 when it is consistent with the version binding field of other scene labels under the same candidate membrane layer parameter combination; This field is used to record the rule version adopted when determining the scene consistency level of the same candidate film layer parameter combination under multiple scene labels. It is masked by the missing measurement identifier in this band. Less than Or the version binding field corresponding to the scene tag. The version binding field is set to 0 when it is inconsistent with the version binding field of other scene labels under the same candidate membrane layer parameter combination; For the first The candidate film layer parameter combinations in the first Reference values for cross-scene front and back differential response contributions corresponding to each band identifier; For the first The candidate film layer parameter combinations in the first Under the first scene tag The front and back differential response contributions corresponding to each band identifier To prevent zero bias in the number of valid scenes, this is used to avoid the denominator being 0 when the number of valid scene labels involved in the calculation is 0. , , and The aforementioned definition will be used.
[0105] Cross-scenario reference values use valid bands in the calculation; bands with missing measurements, low-weight restricted bands, and bands with version inconsistencies are masked and do not participate in the scenario consistency level determination. For the... Before generating cross-scenario reference values for each band identifier, statistics are performed. ;when At that time, no identifier corresponding to that band will be generated. and This band identifier will be masked from the calculation of subsequent multi-scenario response consistency deviation and scenario weight interpretation deviation. The minimum number of valid scene labels required to calculate cross-scene reference values is determined by the minimum scene coverage requirement for scene consistency level determination in the calibration component sample.
[0106] When generating the consistency deviation of multi-scenario response, the weights of the albedo spectral bands in the actual participating band set are renormalized so that the sum of the weights of the albedo spectral bands in the actual participating band set is 1; the weights of the albedo spectral bands in the consistency deviation of multi-scenario response adopt the renormalized albedo spectral band weights.
[0107] Based on cross-scenario reference values, generate the consistency deviation of the multi-scenario response for the same candidate membrane layer parameter combination:
[0108] ,
[0109] in, For the first Consistency deviation of multi-scenario response of candidate membrane layer parameter combinations; For the first The candidate film layer parameter combinations in the first The set of actual participating bands under each scene label for determining scene consistency level, wherein the set of actual participating bands consists of... The missing test label was not blocked. Greater than or equal to Composed of band identifiers that meet the minimum number of valid scene tags and whose version binding fields are consistent; The deviation weights contributed to the front and back differential responses; , , , , , , , , and Following the aforementioned definition, the smaller the deviation in consistency of responses across multiple scenarios, the more stable the differences in the back-side local response and the front-back differential response under different albedo scenarios for the same candidate film layer parameter combination.
[0110] Deviation weights of front and back differential response contributions The deviation is determined by the proportion of the differential response contribution between the front and back sides in the calibration component samples relative to the intrinsic back side deviation. The value is greater than or equal to 0 and less than or equal to 1, and the upper limit of the deviation is consistent with the multi-scene response, using the same calibration version. When the front and back side differential response contribution is missing under a certain scene label, or when the number of valid bands corresponding to the front and back side differential response contribution is less than the minimum band requirement, no front and back side differential response contribution deviation term is generated for that scene label, and it is not included in the calculation. Time is not included in the calculation under this scene tag The item is written with a differential response limited identifier; when all participating scene tags are written with a differential response limited identifier, the front and back differential response contributions are not used as the distinguishing criterion in the uniqueness determination.
[0111] When the number of scene tags participating in the scene consistency level determination reaches a preset minimum, consistency statistics are performed on the attribution identifiers, and the attribution identifiers that remain stable under multiple scene tags for the same candidate film layer parameter combination are used as candidate dominant attribution identifiers to generate the attribution consistency ratio:
[0112] ,
[0113] in, For the first The proportion of attribution consistency for each candidate membrane parameter combination; This is an indicator function; it takes the value 1 when the condition inside the parentheses is true and 0 when the condition inside the parentheses is false. For the first The candidate film layer parameter combinations in the first Attribution tags under each scene label; The candidate dominant attribution identifier is one of the following: backside local poly-Si attribution identifier, backside antireflection film attribution identifier, frontside response attribution identifier, and albedo spectral scene attribution identifier. For the first The number of scene labels involved in the scene consistency level determination for each candidate membrane layer parameter combination; , and Following the previous definition, a higher attribution consistency rate indicates that the difference in response of the candidate membrane parameter combination under different scene labels is dominated by the same attribution source.
[0114] To determine whether scene label switching primarily alters the band weight order, the deviation in scene weight interpretation is based on the common actual participating bands among different scene labels:
[0115] ,
[0116] in, For the first The scene weight interpretation deviation of each candidate membrane layer parameter combination; A scene label pair consisting of the sequence numbers of the two scene label pairs that are participating in the scene switching comparison; For the first A set of scene label pairs that participate in the scene consistency level determination of candidate membrane layer parameter combinations; For the first The candidate film layer parameter combinations in the first The first scene tag and the first The common actual participation band set among scene labels, by and The intersection is obtained; For the first The candidate film layer parameter combinations in the first Under the first scene tag Each band identifier corresponds to the back band deviation data item; For the first The candidate film layer parameter combinations in the first Under the first scene tag Each band identifier corresponds to the back band deviation data item; For the first Under the first scene tag The weight of the albedo spectral band corresponding to each band identifier; For the first Under the first scene tag The weight of the albedo spectral band corresponding to each band identifier; To jointly participate in the band set An operator generates a normalized sort value after sorting the values from smallest to largest. The normalized sort value is greater than or equal to 0 and less than or equal to 1, with 0 corresponding to the smallest sort position and 1 corresponding to the largest sort position. When sorting in pairs, the average of the parallel positions is used. When the number of bands in the common actual participating band set is less than the minimum number of bands required for sorting, no normalized sort value is generated for that scene label pair, and the scene label pair is removed from the list. Remove from the middle; The number of bands in the band set that are actually participated in; To explain the deviation of scene weights and prevent zero bias; and Continuing with the previous definition, the smaller the scene weight interpretation deviation, the closer the sorting change of the backband deviation data items after scene label switching is to the sorting change of the albedo spectral band weights. If... If it is an empty set, then no calculation is performed. The candidate membrane parameter combination is marked as having limited scene weight interpretation. When the scene weight interpretation is limited, the candidate membrane parameter combination does not meet the scene consistency level under multiple scene labels.
[0117] When the same candidate membrane parameter combination satisfies the following formula, it is determined that the matching result of the candidate membrane parameter combination under multiple scene labels reaches the scene consistency level:
[0118] ,
[0119] in, For the first The scenario consistency level achievement indicator for each candidate membrane layer parameter combination is set to 1, indicating that the scenario consistency level has been achieved, and 0 indicates that the scenario consistency level has not been achieved. To ensure consistent response across multiple scenarios with an upper limit on deviation; This represents the lower limit of the attribution consistency ratio. Explain the upper limit of deviation for scene weights; , , and Using the previous definition. If , or If any key quantity in the process is not generated, then... Set it to 0, and write the restricted identifier corresponding to the ungenerated key quantity to the conflict source identifier.
[0120] The upper limit of multi-scenario response consistency deviation, the lower limit of attribution consistency ratio, and the upper limit of scenario weight interpretation deviation are obtained from the calibration component samples. The calibration component samples include bifacial TOPCon components with known backside local poly-Si retention state, known backside antireflective film state, and known carrier surface type. The upper limit of multi-scenario response consistency deviation is set according to the upper quantile value of the cross-scenario deviation of the same film layer state in the calibration samples; the lower limit of attribution consistency ratio is set according to the lowest consistency ratio when the backside local poly-Si attribution remains stable in the calibration samples; the upper limit of scenario weight interpretation deviation is set according to the maximum allowable deviation when the scene label switching only causes a change in the band weight order.
[0121] When multiple scene labels of the same dual-sided TOPCon component under test participate in the scene consistency level determination, the albedo spectral template library version, angle scoring weight version, category boundary version, normalized baseline version, band division version, and scene consistency level threshold version are written into the version binding field. The above version uses offline calibration batches as the refresh window; when any of the following versions is updated: albedo spectral template library version, angle scoring weight version, category boundary version, normalized benchmark version, band division version, or scene consistency level threshold version, the updated version is only used for subsequent test batches, and test batches that have already generated matching results continue to use the original version. If multiple scene labels corresponding to the same candidate film layer parameter combination exist... If there is a discrepancy, the scenario consistency level of the candidate membrane layer parameter combination is marked. Set it to 0 and output the version conflict source identifier.
[0122] When the number of scene tags participating in the scene consistency level determination is less than the preset minimum number, or the common actual participating band set among multiple scene tags is lower than the minimum band coverage requirement, a multi-scene consistency restriction identifier is written to the candidate membrane parameter combination, and the scene consistency level is recorded as a restricted state; when the matching level, candidate matching gap, and attribution identifier meet the uniqueness determination conditions, the target membrane parameter combination and the restricted credibility state are output; when at least one of the matching level, candidate matching gap, or attribution identifier does not meet the uniqueness determination conditions, the parameter non-unique determination identifier and conflict source identifier are output.
[0123] Example 2:
[0124] Based on Example 1, this example provides a specific method for incident angle category labeling and normalization in the TOPCon film parameter determination method;
[0125] Obtaining the albedo spectral template includes:
[0126] Obtain bearing surface category information, which includes at least one of roof surface category, cement surface category, light-colored ground category, and snow surface category;
[0127] The data sources for the albedo spectral template include test data of the bearing surface material, component installation scene records, back irradiation test data, and preset bearing surface category data. The bearing surface category is labeled according to material type, surface color, and surface condition, and the incident angle category is labeled according to component installation tilt angle, component height from bearing surface, bearing surface category, and back test angle.
[0128] In one embodiment, obtaining the albedo spectral template further includes: reading the band albedo sequence in the albedo spectral template library according to the bearing surface category information; performing incident angle category labeling and normalization processing on the band albedo sequence to generate normalized albedo spectral values and normalized reference intensity, forming an albedo spectral template with scene tags; the normalization mode identifier is used to record the intensity reference used for the normalized albedo spectral values, and the scene tag is generated by combining the bearing surface category, incident angle category, and normalization mode identifier; when the same double-sided TOPCon component under test corresponds to multiple scene tags, the back-side weighted spectral response features under each scene tag are generated respectively.
[0129] In a preferred embodiment of Example 2, when labeling the incident angle category of the band albedo sequence, the component installation tilt angle, component height from the bearing surface, bearing surface category, and back test angle of the double-sided TOPCon component under the corresponding scene label are used as the same set of angle label inputs, and the angle scattering coefficient corresponding to the bearing surface category is read from the albedo spectral template library.
[0130] The angle annotation input is fixed according to the same test batch. When the component installation tilt angle, component height from the bearing surface, bearing surface type or back test angle change, the incident angle category is regenerated for the scene label; if the above inputs do not change within the same test batch, the already generated incident angle category is used.
[0131] Before calculating the angle category score, the calibration ranges for component installation tilt angle, component height from the bearing surface, and rear test angle are truncated. Component installation tilt angles less than 0 are counted as 0 in the calculation; those greater than 0 are counted as 0. Press at time Participate in calculations; the height of the component from the bearing surface is less than Press at time Participating in the calculation, greater than Press at time Involved in the calculation; when the back-side test angle is less than 0, it is counted as 0 in the calculation; when it is greater than 0, it is counted as 0 in the calculation. Press at time Participate in the calculation. The aforementioned and All are positive values. Greater than , Take a positive value. When any angle annotation input triggers truncation, write the angle annotation boundary truncation identifier; when multiple angle annotation inputs under the same scene label trigger truncation simultaneously, use that scene label as a restricted scene label in subsequent band matching.
[0132] The angle category score corresponding to the incident angle category is obtained according to the following formula:
[0133] ,
[0134] in, For the first Angle category rating corresponding to each scene label; For scene label sequence numbers; This is a truncation operator that restricts the value within parentheses to the range of 0 to 1; Assign angle scoring weights to components based on their tilt angle. For the first The component installation tilt angle corresponding to each scene label after the calibrated range has been truncated; The upper limit of the calibrated tilt angle for component installation; The angle score weight corresponding to the height of the component above the bearing surface; This is the calibrated upper limit of the height of the component from the bearing surface; For the first The height of the component from the bearing surface after the calibrated range is truncated for each scene label; This is the lower limit of the calibration for the height of the component from the bearing surface; To ensure highly normalized zero bias; The angle score weights are assigned to the corresponding categories of bearing surfaces. For the first The category of the carrying surface in each scene label The corresponding angle scattering coefficient, after being normalized by the same albedo spectral template library version, takes a value greater than or equal to 0 and less than or equal to 1. For the first The carrying surface category corresponding to each scene label; The angle scoring weights are assigned to the backside test angles. For the first The back test angle corresponding to each scene label after the calibrated range has been truncated; This represents the upper limit of the calibration for the backside test angle. The weights for each angle score are calibration parameters, with values greater than or equal to 0 and less than or equal to 1. The sum of the weights for each angle score is 1. The smaller the height of the module from the supporting surface, the stronger the influence of the angle of reflected light received from the supporting surface on the backside. Therefore, the height item uses a reverse normalization method. As the module installation tilt angle, backside test angle, and angle scattering coefficient increase, the angle category score also increases.
[0135] After the angle category score is generated, it is mapped to the incident angle category according to the pre-defined category boundaries:
[0136] ,
[0137] in, For the first The incident angle category corresponding to each scene label; For the first One preset incident angle category; Preset incident angle category number; The maximum index of the preset incident angle category; Preset the incident angle category for the final stage; For the first The lower bound of the scoring for each preset incident angle category; For the first The upper bound of the scoring for each preset incident angle category; Set the lower boundary for the scoring of the incident angle category in the final stage; Set the upper boundary for the scoring of the incident angle category in the final stage; Following the aforementioned definitions, the boundaries of each category are obtained from calibration samples with known installation tilt angles, known component heights from the bearing surface, known bearing surface categories, and known rear-side test angles. The scoring boundaries are set in ascending order of influence from low angles to high angles, and satisfy the following conditions: , Adjacent incident angle categories share the same boundary value. The final preset incident angle category includes its upper boundary score, so that the score of the same angle category falls into only one incident angle category.
[0138] When normalizing the band albedo sequence, missing bands are first masked from the normalization reference, and the albedo of the valid bands is limited to the allowable albedo range of 0 to 1. The limited band albedo is then used for subsequent normalization calculations. The normalization reference includes the integral intensity reference and the specified reference band intensity reference. When the valid bands cover the preset normalized band set, the integral intensity reference is used; when the preset normalized band set has missing measurements but the specified reference band set meets the validity conditions, the specified reference band intensity reference is used.
[0139] The normalized albedo spectral values are obtained according to the following formula:
[0140] ,
[0141] in, For the first Under the first scene tag The normalized albedo spectral values corresponding to each band identifier; For band identification serial number; For the first Under the first scene tag The albedo of each band identifier after the albedo allowable range is limited; This refers to the band summation sequence number in the normalization calculation; For the first The set of valid bands under each scene label that are not blocked by the missing measurement marker; For the first Under the first scene tag The albedo of each band identifier after the albedo allowable range is limited; For the first Each band identifier corresponds to a band width; This is the zero-bias quantity for albedo normalization; For the first The normalized pattern identifier corresponding to each scene label This indicates that an integral strength benchmark is used. This indicates that a specified reference band intensity reference is used; For a specified set of reference bands; This is the zero-offset value for the reference band width. The integrated intensity reference is used to form a relative band distribution based on the integrated intensity of the effective bands, while the specified reference band intensity reference is used to form a relative band distribution based on a specified set of reference bands.
[0142] The normalized reference intensity is the data item used to determine the intensity reference before normalization. Under the integrated intensity reference, it is determined by the band albedo and bandwidth of the effective bands. Under the specified reference band intensity reference, it is determined by the band albedo and bandwidth of the specified reference band set. The normalized reference intensity is written into the albedo spectral template and participates in the processing together with the normalized albedo spectral values when generating the backside weighted response values.
[0143] After normalization, the surface type, incident angle type, normalization mode identifier, normalized albedo spectral value, and normalized reference intensity are written into the albedo spectral template to form an albedo spectral template with scene tags:
[0144] ,
[0145] in, For the first Albedo spectral templates corresponding to each scene label; , , , , and The aforementioned definitions are used for all cases.
[0146] If the component installation tilt angle, component height from the bearing surface, or backside test angle is missing, a temporary angle category score is generated using the calibrated median value corresponding to the non-missing angle annotation input and the missing input. The incident angle category on the low albedo side of the boundary where this temporary angle category score is located is determined as the incident angle category. If the missing input prevents the generation of a temporary angle category score, the incident angle category with the lowest albedo intensity among the preset incident angle categories is determined as the incident angle category. If the specified reference band set does not meet the validity conditions, the integral intensity reference is switched to. If the valid band set cannot meet the minimum band coverage requirements of the integral intensity reference, the normalized albedo spectral value and albedo spectral band weight under this scene label are not generated. Only the scene label and missing measurement mark are retained, and the albedo spectral template is marked as restricted.
[0147] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.
[0148] Furthermore, those skilled in the art will understand that although some embodiments herein include certain features included in other embodiments but not others, combinations of features from different embodiments are meant to be within the scope of the invention and form different embodiments. For example, all the embodiments above can be used in any combination. The information disclosed in the Background section is intended only to enhance the understanding of the general background of the invention and should not be construed as an admission or in any way implying that such information constitutes prior art known to those skilled in the art.
Claims
1. A method for determining TOPCon film parameters based on spectral response, characterized in that, include: Obtain the component-level spectral data package of the double-sided TOPCon component under test and the albedo spectral template corresponding to the high albedo bearing surface condition and with scene tags. The component-level spectral data package includes front spectral response data, back spectral response data and back reflectance spectrum data. Back-side weighted spectral response features are generated from back-side spectral response data, back-side reflectance spectral data, and albedo spectral templates. A back-side local film parameter model is established, which divides the effective light-receiving area on the back side into a metallized region and a non-metallized region, and associates the poly-Si equivalent parameters of the metallized region and the non-metallized region, the back-side antireflection film equivalent parameters, and the region ratio parameters. Based on the back-side local membrane parameter model, candidate membrane parameter combinations, candidate front-side spectral response characteristics, and candidate back-side spectral response characteristics are generated. The candidate frontal spectral response features are matched with the frontal spectral response data by band, and the candidate back spectral response features are weighted by the albedo spectral template and matched with the back weighted spectral response features by band to obtain the matching results. The matching results generate a matching level, candidate matching gap, scenario consistency level, and attribution identifier, which are used to determine uniqueness. When the uniqueness determination condition is met, the target membrane parameter combination and parameter credibility identifier are output; when the condition is not met, the parameter non-unique determination identifier and conflict source identifier are output.
2. The method for determining TOPCon film parameters based on spectral response according to claim 1, characterized in that, The component-level spectral data package also includes frontal reflectance spectral data, and the component-level spectral data package is generated through homogenization processing, which includes: Map the front spectral response data, back spectral response data, front reflectance data, and back reflectance data to the same band identification system; Establish data association based on the component identifier, test direction identifier, and test batch identifier of the double-sided TOPCon component under test; For each band identifier, generate a frontal response sequence, a backal response sequence, a frontal reflection sequence, and a backal reflection sequence; Set a missing measurement marker for the missing band and use the missing measurement marker as a shielding condition for band matching.
3. The method for determining TOPCon film parameters based on spectral response according to claim 1, characterized in that, Obtaining the albedo spectral template includes: Obtain bearing surface category information, which includes at least one of roof surface category, cement surface category, light-colored ground category, and snow surface category; The data sources for the albedo spectral template include test data of the bearing surface material, component installation scene records, back irradiation test data, and preset bearing surface category data; The bearing surface category is marked according to the material type, surface color, and surface condition, while the incident angle category is marked according to the component installation tilt angle, component height from the bearing surface, bearing surface category, and back test angle.
4. The method for determining TOPCon film parameters based on spectral response according to claim 3, characterized in that, Obtaining the albedo spectral template further includes: Read the band albedo sequence from the albedo spectral template library according to the aforementioned bearing surface category information; The incident angle category is labeled and normalized on the albedo sequence of the band to generate normalized albedo spectral values and normalized reference intensity, forming an albedo spectral template with scene tags; the normalization mode identifier is used to record the intensity reference used for the normalized albedo spectral values, and the scene tag is generated by combining the bearing surface category, incident angle category and normalization mode identifier. When the same double-sided TOPCon component under test corresponds to multiple scene labels, the back-side weighted spectral response features under each scene label are generated respectively.
5. The method for determining TOPCon film parameters based on spectral response according to claim 1, characterized in that, The back-side local membrane parameter model includes a region parameter layer, a membrane parameter layer, and a response mapping layer; the region parameter layer records the metallized region identifier, the non-metallized region identifier, and the region ratio parameter; The film parameter layer records the equivalent thickness parameters of poly-Si in the metallized region, the contact retention parameters of poly-Si in the metallized region, the equivalent thickness parameters of poly-Si in the non-metallized region, the absorption category of poly-Si in the non-metallized region, and the equivalent parameters of the antireflective film on the back side. The response mapping layer maps the parameter combinations in the region parameter layer and the film parameter layer into front response contribution, back response contribution, and front-back differential response contribution.
6. The method for determining TOPCon film parameters based on spectral response according to claim 5, characterized in that, The metallized region is the region on the back side that forms an electrical connection with the metal electrode, and the non-metallized region is the region on the back side effective light-receiving region other than the metallized region. The region ratio parameter is determined by the back metallization pattern, the area of the effective light-receiving area on the back, and the area of the non-metallized area. The poly-Si contact retention parameters of the metallized region are obtained by associating the equivalent thickness parameters of the poly-Si metallized region, the contact resistance category, and the annealing batch identifier. The absorption category of the nonmetallized poly-Si region is obtained by correlating the equivalent thickness parameter of the nonmetallized poly-Si region with the absorption intensity of the corresponding band.
7. The method for determining TOPCon film parameters based on spectral response according to claim 1, characterized in that, Generating multiple candidate film layer parameter combinations includes: Read process constraint information, which includes at least one of the following: back poly-Si deposition process identifier, back poly-Si thinning process identifier, back localization process identifier, back anti-reflection film deposition process identifier, and metallization contact process identifier; The range of candidate parameter values is determined based on the process constraint information; Candidate membrane parameter combinations are generated by combining candidate parameters within their range; candidate membrane parameter combinations that conflict with process constraint information are eliminated to form a set of candidate membrane parameters to be matched.
8. The method for determining TOPCon film parameters based on spectral response according to claim 7, characterized in that, Candidate film parameter combinations that conflict with process constraints include: Eliminate combinations where the equivalent thickness parameter of the non-metallized region poly-Si does not match the poly-Si thinning process markings on the back side. Eliminate combinations where the regional proportion parameters do not match the localized processing technology markings on the back; Eliminate combinations where the poly-Si contact retention parameters in the metallized region do not match the metallized contact process specifications; eliminate combinations where the equivalent parameters of the back antireflection film do not match the back antireflection film deposition process specifications.
9. The method for determining TOPCon film parameters based on spectral response according to claim 1, characterized in that, Band-by-band matching includes: According to the preset band division rules, the results of front spectral response data, back weighted spectral response features, candidate front spectral response features, and candidate back spectral response features after being weighted by the albedo spectral template are divided into multiple band response segments. For each band response segment, generate a band deviation data item between the measured response and the candidate response; The scene labels and missing data markers of the front band deviation data items, back band deviation data items, and albedo spectral templates are combined into the matching results; The matching results retain the band source identifier and scene label.
10. The method for determining TOPCon film parameters based on spectral response according to claim 1, characterized in that, After outputting the target film parameter combination and parameter credibility identifier, the target film parameter combination, parameter credibility identifier, scene label of the albedo spectral template, matching result and attribution identifier are written into the parameter record of the double-sided TOPCon component under test; Process backtracking labels are generated based on the target film layer parameter combination. The process backtracking labels include at least one of the following: non-metallized region poly-Si thinning bias label, metallized region poly-Si retention bias label, back antireflection film bias label, and albedo spectral scene adaptation label. When the output parameter cannot be uniquely identified, the conflict source identifier is written into the parameter record, and the parameter record is set to the verification state.
