Steel rail specification adjustment model construction method and device and steel rail specification adjustment method

Abstract

The invention relates to the field of steel rolling, and discloses a steel rail specification adjustment model construction method and device and a steel rail specification adjustment method. The model construction method comprises the steps of obtaining steel rail rolled piece specification data; calculating a first difference value vector, a second difference value vector and a third difference value vector, wherein the first difference value vector is equal to the standard specification array minus the specification array of the rolled piece before adjustment; the second difference vector is obtained by subtracting the specification array of the rolled piece before adjustment from the specification array of the rolled piece after adjustment; the third difference value vector is equal to the standard specification array minus the specification array of the adjusted rolled piece; on the basis of the first difference vector, the second difference vector and the third difference vector, determining whether each group of adjustment data is valid or not, so as to screen out multiple groups of valid adjustment data from the multiple groups of adjustment data; and constructing a steel rail specification adjustment model based on the multiple groups of effective adjustment data. According to the method, the artificial experience can be refined and reconstructed, and it is guaranteed that the effective steel rail specification adjustment model is constructed.

Description

Technical Field

[0001] The present invention relates to the technical field of steel rolling, and in particular to a method and device for constructing a rail specification adjustment model, and a rail specification adjustment method. Background Art

[0002] The research on the final specification accuracy control of high-precision rails has become a frontier topic actively researched by major high-precision rail manufacturers around the world. For the final specification accuracy control of high-precision rails, the traditional method is to adjust them manually, which has accumulated a lot of adjustment experience. However, manual adjustment is accompanied by instability. Such adjustments vary from person to person. There are quality differences and adjustment ideas between different production shifts of the same product. In addition, the working conditions of the adjustment personnel will also cause quality problems, which will greatly affect the improvement of the accuracy of the rail size. Therefore, it is necessary to study a rail specification adjustment model based on the existing many adjustment experiences to realize automatic adjustment and improve the accuracy and stability of the adjustment.

[0003] When constructing the rail specification adjustment model, due to various issues such as many related factors, large cross-influence between parameters, and parameter accuracy, the analysis results of equipment parameters and specification parameters are inconsistent with the actual deformation law and actual adjustment method. How to refine and reconstruct manual experience is the key to building an automated rail high-precision control model based on manual experience. Summary of the invention

[0004] The main purpose of the present invention is to provide a rail specification adjustment model construction method and device, and a rail specification adjustment method to solve the problem of how to refine and reconstruct artificial experience to construct an effective rail specification adjustment model.

[0005] According to one aspect of the present invention, a method for constructing a rail specification adjustment model is provided, comprising:

[0006] Acquire the specification data of the rail rolled piece, the specification data of the rail rolled piece includes a standard specification array and a plurality of corresponding groups of adjustment data, each group of adjustment data includes a specification array of the rolled piece before adjustment and a specification array of the rolled piece after adjustment;

[0007] Calculate the first difference vector, the second difference vector and the third difference vector, the first difference vector is equal to the standard specification array minus the specification array of the rolled piece before adjustment; the second difference vector is equal to the specification array of the rolled piece after adjustment minus the specification array of the rolled piece before adjustment; the third difference vector is equal to the standard specification array minus the specification array of the rolled piece after adjustment;

[0008] Based on the first difference vector, the second difference vector and the third difference vector, determining whether each set of adjustment data is valid, thereby screening out multiple sets of valid adjustment data from the multiple sets of adjustment data;

[0009] A rail specification adjustment model is constructed based on multiple groups of effective adjustment data, wherein the model is configured to determine corresponding adjustment schemes and affected specifications and affected values ​​corresponding to the adjustment schemes based on specific specifications that need to be adjusted and their adjustment values.

[0010] According to one embodiment of the present invention, the specification array includes data of multiple specifications of the rolled product, and the multiple specifications include multiple of the following: rail height, waist thickness, rail crown fullness, upper head width, bottom width, upper leg tip thickness, lower leg tip thickness, upper abdominal cavity height, lower abdominal cavity height, upper symmetry, lower symmetry, head deviation, upper profile, and lower profile.

[0011] According to an embodiment of the present invention, determining whether each set of adjustment data is valid based on the first difference vector, the second difference vector, and the third difference vector includes:

[0012] For each set of adjustment data, let SUM[ΔSam1]-SUM[ΔXY]=X,

[0013] If lim X→0 f([ΔSam2])=0, then the set of adjustment data is valid;

[0014] Among them, [ΔSam1] is the first difference vector; [ΔXY] is the second difference vector; [ΔSam2] is the third difference vector.

[0015] According to one embodiment of the present invention, a rail specification adjustment model is constructed based on multiple groups of effective adjustment data, including: using a decision tree algorithm and a K-nearest neighbor algorithm to perform data classification to determine the affected specifications and their impact values ​​corresponding to different adjustment schemes.

[0016] According to one embodiment of the present invention, the construction method further includes: judging whether the rolling stand is in a normal state based on the rolling force and the pressing amount of each rolling stand during rolling; when the rolling stand is judged to be in a normal state, obtaining the specification data of the rail rolled piece;

[0017] Among them, when the following formula is satisfied, the rolling stand is judged to be in normal state:

[0018] (Z1-Z2) / (S1-S2)≤Td

[0019] Among them, Z1 is the rolling force of the first rolled piece without considering the fluctuation value; Z2 is the rolling force of the second rolled piece without considering the fluctuation value; S1 is the pressing amount of the first rolled piece; S2 is the pressing amount of the second rolled piece; Td is the proportion of the pressing amount controlled by the rolling force.

[0020] According to another aspect of the present invention, a rail specification adjustment model construction device is provided, comprising:

[0021] An acquisition module is configured to acquire specification data of a rail rolled piece, wherein the specification data of the rail rolled piece includes a standard specification array and a corresponding plurality of groups of adjustment data, wherein each group of adjustment data includes a specification array of the rolled piece before adjustment and a specification array of the rolled piece after adjustment;

[0022] A calculation module is configured to calculate a first difference vector, a second difference vector and a third difference vector, wherein the first difference vector is equal to the standard specification array minus the specification array of the rolled piece before adjustment; the second difference vector is equal to the specification array of the rolled piece after adjustment minus the specification array of the rolled piece before adjustment; and the third difference vector is equal to the standard specification array minus the specification array of the rolled piece after adjustment;

[0023] a screening module configured to determine whether each set of adjustment data is valid based on the first difference vector, the second difference vector and the third difference vector, thereby screening out multiple sets of valid adjustment data from the multiple sets of adjustment data;

[0024] A construction module is configured to construct a rail specification adjustment model based on multiple sets of valid adjustment data, wherein the model is configured to determine a corresponding adjustment scheme and an affected specification and its affected value corresponding to the adjustment scheme based on a specific specification that needs to be adjusted and its adjustment value.

[0025] According to another aspect of the present invention, a method for adjusting rail specifications is provided, comprising the following steps:

[0026] S21, determining multiple specifications that need to be adjusted and their adjustment values ​​based on the specification array of the current rolled piece and the corresponding standard specification array, the upper limit exceeding specification array, and the lower limit exceeding specification array;

[0027] S22, determining a priority adjustment specification from multiple specifications that need to be adjusted;

[0028] S23, substituting the priority adjustment specifications and their adjustment values ​​into the rail specification adjustment model to obtain the corresponding adjustment scheme and the affected specifications and their impact values; wherein the rail specification adjustment model is constructed according to any one of the methods of claims 1-5;

[0029] S24, based on the adjustment value of the priority adjustment specification and the impact value of the affected specification, updating the specification array of the current rolled piece, and completing the first adjustment;

[0030] S25, judging whether the updated specification array of the current rolled piece is qualified or whether a predetermined number of adjustments have been made; when the judgment is yes, the adjustment is ended; when the judgment is no, repeating steps S21 to S24.

[0031] According to one embodiment of the present invention, step S21 includes:

[0032] Subtract the specification array of the current rolled piece from the standard specification array to obtain the adjustment vector;

[0033] The adjustment vector is compared with the upper limit exceeding specification array and the lower limit exceeding specification array. When a certain specification data in the adjustment vector does not meet the upper limit exceeding specification array and the lower limit exceeding specification array, it is determined that the specification needs to be adjusted.

[0034] According to one embodiment of the present invention, step S21 includes:

[0035] If the specification to be adjusted is closer to the upper limit, the difference between the current value of the specification to be adjusted and 50% to 100% of the upper limit is calculated as the adjustment value of the specification to be adjusted;

[0036] If the specification to be adjusted is closer to the lower limit, the difference between the current value of the specification to be adjusted and 50% to 100% of the lower limit is calculated as the adjustment value of the specification to be adjusted.

[0037] According to one embodiment of the present invention, step S22 includes: sorting multiple specifications that need to be adjusted according to the adjustment difference percentage, and determining the specification with the largest adjustment difference percentage as the priority adjustment specification; wherein, if the specification that needs to be adjusted is closer to the upper limit, the adjustment difference percentage = adjustment value / upper limit excess data*100%; if the specification that needs to be adjusted is closer to the lower limit, the adjustment difference percentage = adjustment value / lower limit excess data*100%.

[0038] In the technical solution of the present invention, the first difference vector, the second difference vector, and the third difference vector are calculated based on the standard specification array, the specification array of the rolled piece before adjustment, and the specification array of the rolled piece after adjustment, and the validity of each set of adjustment data is determined based on the first difference vector, the second difference vector, and the third difference vector, so as to screen out multiple sets of effective adjustment data from multiple sets of adjustment data for constructing a rail specification adjustment model. Through the solution of the present invention, artificial experience can be refined and reconstructed to ensure the construction of an effective rail specification adjustment model, based on which automated rail specification adjustment can be realized, ensuring stable rolling while improving rolling accuracy, and improving adjustment effectiveness and standardization of rolling adjustment. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0040] Figure 1 A flow chart showing a method for constructing a rail specification adjustment model according to an embodiment of the present invention;

[0041] Figure 2A schematic diagram showing a rail specification adjustment model building device according to an embodiment of the present invention;

[0042] Figure 3 A flow chart showing a method for adjusting rail specifications according to an embodiment of the present invention;

[0043] Figure 4 Another flow chart of the method for adjusting rail specifications according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0044] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the embodiments of the present invention are further described in detail below in combination with specific embodiments and with reference to the accompanying drawings.

[0045] It should be noted that all expressions using "first" and "second" in the embodiments of the present invention are for distinguishing two non-identical entities with the same name or non-identical parameters. It can be seen that "first" and "second" are only for the convenience of expression and should not be understood as limitations on the embodiments of the present invention. The subsequent embodiments will not explain this one by one.

[0046] refer to Figure 1 The present invention proposes a method for constructing a rail specification adjustment model, comprising:

[0047] S11, obtaining specification data of rail rolled products, the specification data of rail rolled products including a standard specification array and corresponding multiple groups of adjustment data, each group of adjustment data including a specification array of the rolled products before adjustment and a specification array of the rolled products after adjustment;

[0048] S12, calculating a first difference vector, a second difference vector and a third difference vector, wherein the first difference vector is equal to the standard specification array minus the specification array of the rolled piece before adjustment; the second difference vector is equal to the specification array of the rolled piece after adjustment minus the specification array of the rolled piece before adjustment; and the third difference vector is equal to the standard specification array minus the specification array of the rolled piece after adjustment;

[0049] S13, determining whether each set of adjustment data is valid based on the first difference vector, the second difference vector, and the third difference vector, thereby selecting multiple sets of valid adjustment data from the multiple sets of adjustment data;

[0050] S14, constructing a rail specification adjustment model based on multiple groups of valid adjustment data, wherein the model is configured to determine a corresponding adjustment scheme and an affected specification and an affected value corresponding to the adjustment scheme based on a specific specification that needs to be adjusted and an adjustment value thereof.

[0051] In an embodiment of the present invention, the first difference vector, the second difference vector, and the third difference vector are calculated based on the standard specification array, the specification array of the rolled piece before adjustment, and the specification array of the rolled piece after adjustment, and the validity of each set of adjustment data is determined based on the first difference vector, the second difference vector, and the third difference vector, so that multiple sets of effective adjustment data are screened out from multiple sets of adjustment data for constructing a rail specification adjustment model. Through the scheme of the present invention, manual experience can be refined and reconstructed to ensure the construction of an effective rail specification adjustment model, and automatic rail specification adjustment can be achieved based on the model to ensure stable rolling while improving rolling accuracy, improving adjustment effectiveness and standardization of steel rolling adjustment. It can not only solve the problem of low calculation efficiency of manual adjustment, but also avoid errors caused by manual adjustment, and effectively reduce product specification control fluctuations between shifts.

[0052] In some embodiments, the specification array includes data of multiple specifications of the rolled piece, and the multiple specifications include multiple of the following: rail height, waist thickness, rail crown fullness, upper head width, bottom width, upper leg tip thickness, lower leg tip thickness, upper abdominal cavity height, lower abdominal cavity height, upper symmetry, lower symmetry, head deviation, upper profile, and lower profile. Among them, "upper symmetry" refers to the symmetry of the upper leg of the rolling line; "lower symmetry" refers to the symmetry of the upper and lower legs of the rolling line; "head deviation" refers to the degree of deviation of the center line of the head on the rolling line relative to the center line of the waist, and the center line of the head is above the center line of the waist, which is defined as head deviation upper, and vice versa; "upper profile" refers to the degree of deviation of the upper profile of the rail relative to the ideal upper profile; "lower profile" refers to the degree of deviation of the lower profile of the rail relative to the ideal lower profile.

[0053] The rail rolling product specification data can be obtained by collecting the on-site rolling conditions through the data acquisition equipment. The standard specification array can include standard values ​​of multiple specifications. For example, [Sav] represents the standard specification array, [Sav] = [Hg Sav ,Wd Sav ,Tew Sav ....], Hg Sav ,Wd Sav ,Tew Sav.... represents the standard values ​​of multiple specifications. The specification array of the rolled piece before adjustment may include multiple specifications of pre-adjustment data. For example, [QR1] represents the specification array of the rolled piece before adjustment, [QR1] = [Hg1, Wd1, Tew1....], where Hg1, Wd1, Tew1.... represent multiple specifications of pre-adjustment data. The specification array of the rolled piece after adjustment may include multiple specifications of post-adjustment data. For example, [QR2] represents the specification array of the rolled piece after adjustment, [QR2] = [Hg2, Wd2, Tew2....], where Hg2, Wd2, Tew2.... represent multiple specifications of post-adjustment data. In some embodiments of the present invention, an upper limit exceeding specification array and a lower limit exceeding specification array are also involved. The upper limit exceeding specification array may include multiple specifications of upper limit exceeding data (the upper limit exceeding data indicates the upper limit exceeding the standard value allowed, and the value is positive). The upper limit exceeding specification array is represented by [Url], [Url] = [H Url ,W Url ,T Url ....]. The lower limit exceeding specification array may include multiple specification lower limit exceeding data (lower limit exceeding data indicates that the lower limit exceeding the standard value is allowed, and its value is negative), and [Ll] represents the upper limit exceeding specification array, [Ll] = [H Ll ,W Ll ,T Ll ....].

[0054] [ΔSam1] represents the first difference vector, [ΔSam1] = [Sav] - [QR1] = [Hg Sav -Hg1,Wd Sav -Wd1,Tew Sav -Tew1....]. The second difference vector is represented by [ΔXY], [ΔXY] = [QR2] - [QR1] = [Hg2 - Hg1, Wd2 - Wd1, Tew2 - Tew1....]. The third difference vector is represented by [ΔSam2], [ΔSam2] = [Sav] - [QR2] = [Hg Sav -Hg2,Wd Sav -Wd2,Tew Sav -Tew2....].

[0055] The first difference vector [ΔSam1] can indicate the specification data and adjustment direction that need to be adjusted. When [Url]>[ΔSam1]>[Ll], the rolled piece has normal specifications and does not need to be adjusted. When a certain specification data in [ΔSam1] exceeds [Url] or [Ll], it is necessary to adjust the specification, and the adjustment direction is also clear. The second difference vector [ΔXY] can indicate the actual adjusted specification data and adjustment direction.

[0056] In some embodiments, determining whether each set of adjustment data is valid based on the first difference vector, the second difference vector, and the third difference vector includes:

[0057] For each set of adjustment data, let SUM[ΔSam1]-SUM[ΔXY]=X,

[0058] If lim X→0 f([ΔSam2])=0 (i.e., when X approaches 0, [ΔSam2] also approaches 0), then the set of adjustment data is valid;

[0059] Among them, [ΔSam1] is the first difference vector; [ΔXY] is the second difference vector; [ΔSam2] is the third difference vector.

[0060] Data classification can be performed based on multiple groups of valid adjustment data, and a corresponding table can be established, which can indicate the corresponding relationship between different specifications and their adjustment values ​​and different adjustment schemes, and the corresponding relationship between different adjustment schemes and different affected specifications and their impact values. Based on the corresponding table, the specifications to be adjusted and their adjustment values ​​can be used as inputs, and the corresponding adjustment schemes and the affected specifications and their impact values ​​can be obtained as outputs.

[0061] In some embodiments, a rail specification adjustment model is constructed based on multiple groups of effective adjustment data, including: using a decision tree algorithm and a K nearest neighbor algorithm to classify data to determine the affected specifications and their impact values ​​corresponding to different adjustment schemes. The adjustment amount can be split for big data classification, and the decision tree algorithm and the K nearest neighbor algorithm are used to complete the classification of the data parameters affected by the frame pressure: for example, the X frame Y roller pressure a value affects the final specifications of the rolling mill, and the affected rail height is [b1, b2, b3...], and the crown fullness is [c1, c2, c3...]. The average value is finally taken: the affected rail height is Δb, and the crown fullness is Δc. The priority adjustment direction of specification adjustment: the rail height needs to adjust the z value, and the combination adjustment is performed using the A frame B roller and the C frame D roller.

[0062] In some embodiments, the construction method of the present invention further includes: judging whether the rolling stand is in a normal state based on the rolling force and the downward pressure of each rolling stand during rolling; when judging that the rolling stand is in a normal state, obtaining the specification data of the rail rolled piece, ensuring that the obtained data is obtained under a normal rolling state, and ensuring the validity of the data. The rolling roll gap state can be judged according to the rolling force fluctuation change formula.

[0063] Among them, when the following formula is satisfied, the rolling stand is judged to be in normal state:

[0064] (Z1-Z2) / (S1-S2)≤Td

[0065] Among them, Z1 is the rolling force of the first rolled piece without considering the fluctuation value; Z2 is the rolling force of the second rolled piece without considering the fluctuation value; S1 is the pressing amount of the first rolled piece; S2 is the pressing amount of the second rolled piece; Td is the proportion of the pressing amount controlled by the rolling force, which is a known value.

[0066] The first and second pieces are rolled by the same stand, and the rolling force is the rolling force of the same roller of the same stand, for example, the rolling force of the Y roller of the X stand, which is normal in the range of 2200 to 3000. The rolling force of the first piece is Z1+R1, and the rolling force of the second piece is Z2+R2. R1 and R2 are the fluctuation values ​​in the rolling force signal. Except for special circumstances, R1-R2 is infinitely close to 0.

[0067] refer to Figure 2 The present invention further provides a rail specification adjustment model building device 100, comprising:

[0068] The acquisition module 10 is configured to acquire the specification data of the rail rolled product, the specification data of the rail rolled product includes a standard specification array and a plurality of corresponding adjustment data groups, each adjustment data group includes a specification array of the rolled product before adjustment and a specification array of the rolled product after adjustment;

[0069] The calculation module 20 is configured to calculate a first difference vector, a second difference vector and a third difference vector, wherein the first difference vector is equal to the standard specification array minus the specification array of the rolled piece before adjustment; the second difference vector is equal to the specification array of the rolled piece after adjustment minus the specification array of the rolled piece before adjustment; and the third difference vector is equal to the standard specification array minus the specification array of the rolled piece after adjustment;

[0070] A screening module 30 is configured to determine whether each set of adjustment data is valid based on the first difference vector, the second difference vector and the third difference vector, so as to screen out multiple sets of valid adjustment data from the multiple sets of adjustment data;

[0071] The construction module 40 is configured to construct a rail specification adjustment model based on multiple sets of valid adjustment data, wherein the model is configured to determine a corresponding adjustment scheme and an affected specification and its affected value corresponding to the adjustment scheme based on a specific specification to be adjusted and its adjustment value.

[0072] The rail specification adjustment model of the present invention can intervene in the whole process of rail rolling, monitor the state of the roll gap rolling force in real time, control the various pressing amounts of the racks, keep them in a stable working state, and ensure the effectiveness and stability of the model adjustment. The model participates in the adjustment in real time monitoring and real time adjustment to ensure that the final specifications of the rails meet the specification requirements of high-precision rails.

[0073] refer to Figure 3 The present invention also provides a method for adjusting rail specifications, comprising the following steps:

[0074] S21, determining multiple specifications that need to be adjusted and their adjustment values ​​based on the specification array of the current rolled piece and the corresponding standard specification array, the upper limit exceeding specification array, and the lower limit exceeding specification array;

[0075] S22, determining a priority adjustment specification from multiple specifications that need to be adjusted;

[0076] S23, substituting the priority adjustment specifications and their adjustment values ​​into the rail specification adjustment model to obtain the corresponding adjustment scheme and the affected specifications and their impact values; wherein the rail specification adjustment model is constructed according to any one of the methods of claims 1-5;

[0077] S24, based on the adjustment value of the priority adjustment specification and the impact value of the affected specification, updating the specification array of the current rolled piece, and completing the first adjustment;

[0078] S25, judging whether the updated specification array of the current rolled piece is qualified or whether a predetermined number of adjustments have been made; when the judgment is yes, the adjustment is ended; when the judgment is no, steps S21 to S24 are repeated until the updated specification array of the current rolled piece is qualified or a predetermined number of adjustments have been made. The predetermined number of times can be, for example, 2 to the power of n, where n is the number of specification items (e.g., 14).

[0079] In some embodiments, step S21 includes:

[0080] Subtract the specification array [QR] of the current rolled product from the standard specification array [Sav] to obtain the adjustment vector;

[0081] The adjustment vector is compared with the upper limit exceeding specification array [Url] and the lower limit exceeding specification array [Ll]. When a certain specification data in the adjustment vector does not meet the upper limit exceeding specification array and the lower limit exceeding specification array, the specification is determined to be a specification that needs to be adjusted, and the specification that needs to be adjusted can be marked.

[0082] In some embodiments, step S21 includes:

[0083] If the specification to be adjusted is closer to the upper limit, the difference between the current value of the specification to be adjusted and 50% to 100% of the upper limit (for example, for some specifications, 50% to 80% of the upper limit can be used; for other specifications (such as bottom width), 80% to 100% of the upper limit can be used) is calculated as the adjustment value of the specification to be adjusted;

[0084] If the specification that needs to be adjusted is closer to the lower limit, calculate the difference between the current value of the specification that needs to be adjusted and 50% to 100% of the lower limit (for example, for some specifications, 50% to 80% of the upper limit can be used; for other specifications (such as bottom width), 80% to 100% of the upper limit can be used) as the adjustment value of the specification that needs to be adjusted.

[0085] In some embodiments, step S22 includes: sorting multiple specifications that need to be adjusted according to the adjustment difference percentage, and determining the specification with the largest adjustment difference percentage as the priority adjustment specification; wherein, if the specification that needs to be adjusted is closer to the upper limit, the adjustment difference percentage = adjustment value / upper limit excess data*100%; if the specification that needs to be adjusted is closer to the lower limit, the adjustment difference percentage = adjustment value / lower limit excess data*100%.

[0086] refer to Figure 4 In some embodiments, the rail specification adjustment method includes the following steps:

[0087] (1) Read standard and actual data. The actual values ​​include 14 specifications: rail height, waist thickness, rail crown fullness, head width, bottom width, upper leg tip thickness, lower leg tip thickness, upper abdominal cavity height, lower abdominal cavity height, upper symmetry, lower symmetry, head deviation, upper profile, and lower profile.

[0088] (2) Determine the steel production situation and select a more suitable solution.

[0089] (3) Read the standard value from the database, subtract it from the actual value to obtain the difference, and determine whether the actual value of the specification is biased towards the upper limit or the lower limit.

[0090] (4) Calculate the required adjustment value, that is, obtain the specification adjustment difference vector. If the actual specification value exceeds 60% of the upper limit or 60% of the lower limit, calculate the difference between the difference value and 60% of the upper limit or 60% of the lower limit to obtain the adjustment value. If the actual specification value does not exceed 60% of the upper limit or 60% of the lower limit, no calculation is performed.

[0091] (5) Determine the ratio of all adjustment values ​​to the corresponding limit values ​​and sort them. The values ​​with the largest excess ratio are adjusted first. Adjust the actual value only once, then modify the actual value (add the adjustment value and the impact value to the original actual value, the new actual value = the original actual value + adjustment value + other specification impact values), and then repeat the cycle.

[0092] (6) Determine whether all specifications are qualified or whether the upper limit of the adjustment times is reached. If the judgment is yes, end the adjustment. If the judgment is no, record the adjustment value, impact value, impact specification, and the number of adjustment plans (accumulated), and return to step (3), and repeat the following steps from step (3).

[0093] The following describes a specific embodiment as an example.

[0094] The adjustment method of the present invention is implemented in the rolling process of the 60 rails in the second line of the rail beam plant:

[0095] Step 1: Determine the standard value, upper and lower limits. The relevant data are shown in Table 1.

[0096] Table 1: Standard values, upper and lower limits of rolled product specifications

[0097]

[0098] Step 2: Input the sampling deviation, the model starts to calculate, and obtains the roll gap reduction of the stand. The sampling deviation is shown in Table 2.

[0099] Table 2: Sampling bias

[0100]

[0101] Step 3: Model calculation logic

[0102] Specification array of rolled product before adjustment [QR1] = [rail height: 176.4, waist thickness: 16.25, rail crown fullness: 9.8, upper head width: 70.9, bottom width: 150.1, upper leg tip thickness: 13.5, lower leg tip thickness: 13.2, upper abdominal cavity height: 115.4, lower abdominal cavity height: 115.4, upper symmetry: 0.8, lower symmetry: -0.7, head deviation: 0, profile upper: -0.1, profile lower: 0]

[0103] Adjustment vector [ΔSam1] = standard specification array [Sav] - specification array of rolled product before adjustment [QR1] = [rail height: 0.3, waist thickness: 0.25, rail crown fullness: -0.2, upper head width: -0.1, bottom width: -0.1, upper leg tip thickness: -0.1, lower leg tip thickness: -0.4, upper abdominal cavity height: -0.1, lower abdominal cavity height: -0.1, upper symmetry: -0.8, lower symmetry: 0.7, head deviation: 0, profile upper: 0.1, profile lower: 0]

[0104] The adjustment vector [ΔSam1] and the upper limit specification [Url] or the lower limit specification [Ll] are used to make specification judgment:

[0105] Build adjustment table: [rail height: 0.3, waist thickness: 0.25, upper symmetry: -0.3, lower symmetry: 0.2]

[0106] Calculate the adjustment difference percentage according to the adjustment table and construct the adjustment sequence table: [① rail height: -0.3, ② waist thickness: -0.25, ③ upper symmetry: 0.3, ④ lower symmetry: -0.2]

[0107] Adjust the specifications in order according to the adjustment sequence table. For example, if the rail height adjustment needs to be increased by 0.3, substitute [adjustment specification: rail height, adjustment parameter: 0.3] into the model for calculation, and obtain the adjustment plan [U2 3rd pass operating side roll gap: adjustment 0.5; UF 1st pass operating side roll gap: adjustment 0.2] and influence value data [upper leg tip thickness: influence value 0.15; lower leg tip thickness: influence value 0.15]. Substitute the influence value and adjustment value into the specification array [QR1] of the rolled piece before adjustment for calculation, and the first round of adjustment is completed. Reconstruct the adjustment sequence table again and select the specifications that exceed the maximum percentage for adjustment until all the adjusted specifications are qualified or the upper limit of the number of solutions is reached. The reconciliation record is shown in Table 3.

[0108] Table 3: Adjustment Record

[0109]

[0110] Judging from the analysis results, the rolled rails are adjusted based on the adjustment plan obtained after model calculation, and the final adjusted rail specifications and dimensions meet the standard values ​​and upper and lower limits.

[0111] It should be noted that a person skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing related hardware through a computer program, and the program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods.

[0112] A person skilled in the art should understand that the discussion of any of the above embodiments is only exemplary and is not intended to imply that the scope of the disclosure of the embodiments of the present invention (including the claims) is limited to these examples; under the concept of the embodiments of the present invention, the technical features in the above embodiments or different embodiments can also be combined, and there are many other changes in different aspects of the embodiments of the present invention as described above, which are not provided in detail for the sake of simplicity. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the embodiments of the present invention should be included in the protection scope of the embodiments of the present invention.

Claims

1. A method for constructing a rail specification adjustment model, It is characterized in that include: Acquire rail rolled product specification data, wherein the rail rolled product specification data includes a standard specification array and a plurality of corresponding adjustment data groups, each adjustment data group includes a specification array of the rolled product before adjustment and a specification array of the rolled product after adjustment; Calculate a first difference vector, a second difference vector and a third difference vector, wherein the first difference vector is equal to the standard specification array minus the specification array of the rolled piece before adjustment; the second difference vector is equal to the specification array of the rolled piece after adjustment minus the specification array of the rolled piece before adjustment; and the third difference vector is equal to the standard specification array minus the specification array of the rolled piece after adjustment; Based on the first difference vector, the second difference vector and the third difference vector, determining whether each set of adjustment data is valid, thereby screening out multiple sets of valid adjustment data from the multiple sets of adjustment data; The rail specification adjustment model is constructed based on the multiple groups of valid adjustment data, wherein the model is configured to determine corresponding adjustment schemes and affected specifications and affected values ​​corresponding to the adjustment schemes based on specific specifications that need to be adjusted and their adjustment values.

2. The method according to claim 1, It is characterized in that The specification array includes data of multiple specifications of the rolled product, and the multiple specifications include multiple of the following: rail height, waist thickness, rail crown fullness, upper head width, bottom width, upper leg tip thickness, lower leg tip thickness, upper abdominal cavity height, lower abdominal cavity height, upper symmetry, lower symmetry, head deviation, upper profile, and lower profile.

3. The method according to claim 1, It is characterized in that The determining whether each set of adjustment data is valid based on the first difference vector, the second difference vector, and the third difference vector includes: For each set of adjustment data, let SUM[ΔSam1]-SUM[ΔXY]=X, If lim X→0 f([ΔSam2])=0, then the set of adjustment data is valid; Among them, [ΔSam1] is the first difference vector; [ΔXY] is the second difference vector; [ΔSam2] is the third difference vector.

4. The method according to claim 1, It is characterized in that The constructing of the rail specification adjustment model based on the multiple groups of effective adjustment data includes: using a decision tree algorithm and a K-nearest neighbor algorithm to perform data classification to determine the affected specifications and their impact values ​​corresponding to different adjustment schemes.

5. The method according to claim 1, It is characterized in that Also includes: Based on the rolling force and the pressing amount of each rolling stand during rolling, judging whether the rolling stand is in a normal state; when judging that the rolling stand is in a normal state, obtaining the specification data of the rail rolled piece; Among them, when the following formula is satisfied, the rolling stand is judged to be in normal state: (Z 1 -Z 2 ) / (S 1 -S 2 )≤Td Among them, Z 1 is the rolling force of the first rolled piece without considering the fluctuation value; Z 2 is the rolling force of the second rolling piece without considering the fluctuation value; S 1 is the downward pressure of the first rolled piece; S 2 is the pressing amount of the second rolled piece; Td is the proportional value of the pressing amount controlled by the rolling force.

6. A device for constructing a model for adjusting rail specifications, It is characterized in that include: An acquisition module is configured to acquire specification data of a rail rolled piece, wherein the specification data of the rail rolled piece includes a standard specification array and a plurality of corresponding groups of adjustment data, each group of adjustment data includes a specification array of the rolled piece before adjustment and a specification array of the rolled piece after adjustment; A calculation module, configured to calculate a first difference vector, a second difference vector and a third difference vector, wherein the first difference vector is equal to the standard specification array minus the specification array of the rolled piece before adjustment; the second difference vector is equal to the specification array of the rolled piece after adjustment minus the specification array of the rolled piece before adjustment; the third difference vector is equal to the standard specification array minus the specification array of the rolled piece after adjustment; a screening module configured to determine whether each set of adjustment data is valid based on the first difference vector, the second difference vector and the third difference vector, so as to screen out multiple sets of valid adjustment data from the multiple sets of adjustment data; A construction module is configured to construct the rail specification adjustment model based on the multiple sets of valid adjustment data, wherein the model is configured to determine a corresponding adjustment scheme and an affected specification and an affected value corresponding to the adjustment scheme based on a specific specification that needs to be adjusted and its adjustment value.

7. A method for adjusting rail specifications, It is characterized in that The following steps are involved: S21, determining a plurality of specifications that need to be adjusted and their adjustment values ​​based on the specification array of the current rolled piece and the corresponding standard specification array, the upper limit exceeding specification array, and the lower limit exceeding specification array; S22, determining a priority adjustment specification from the multiple specifications that need to be adjusted; S23, substituting the priority adjustment specification and its adjustment value into a rail specification adjustment model to obtain a corresponding adjustment scheme and affected specifications and their impact values; wherein the rail specification adjustment model is constructed according to the method described in any one of claims 1-5; S24, based on the adjustment value of the priority adjustment specification and the impact value of the affected specification, updating the specification array of the current rolled piece to complete the first adjustment; S25, judging whether the updated specification array of the current rolled piece is qualified or whether a predetermined number of adjustments have been made; when the judgment is yes, the adjustment is ended; when the judgment is no, repeating steps S21 to S24.

8. The method according to claim 7, It is characterized in that Step S21 includes: Subtract the specification array of the current rolled piece from the standard specification array to obtain an adjustment vector; The adjustment vector is compared with the upper limit exceeding specification array and the lower limit exceeding specification array. When a certain specification data in the adjustment vector does not conform to the upper limit exceeding specification array and the lower limit exceeding specification array, the specification is determined to be a specification that needs to be adjusted.

9. The method according to claim 8, It is characterized in that Step S21 includes: If the specification to be adjusted is closer to the upper limit, the difference between the current value of the specification to be adjusted and 50% to 100% of the upper limit is calculated as the adjustment value of the specification to be adjusted; If the specification that needs to be adjusted is closer to the lower limit, the difference between the current value of the specification that needs to be adjusted and 50% to 100% of the lower limit is calculated as the adjustment value of the specification that needs to be adjusted.

10. The method according to claim 9, It is characterized in that Step S22 includes: sorting the multiple specifications that need to be adjusted according to the adjustment difference percentage, and determining the specification with the largest adjustment difference percentage as the priority adjustment specification; wherein, if the specification that needs to be adjusted is closer to the upper limit, the adjustment difference percentage = adjustment value / upper limit excess data*100%; if the specification that needs to be adjusted is closer to the lower limit, the adjustment difference percentage = adjustment value / lower limit excess data*100%.

Images