Full prefabricated refrigeration station compartment adaptive assembly method

Abstract

The application discloses a full-precast refrigeration station compartmentized self-adaptive assembling method, and belongs to the refrigeration station assembling technical field, aiming at solving the problems of the module interface butt joint depending on the repeated measurement and adjustment on site, the large on-site wiring and debugging workload, and the slow entering into stable working condition after assembling. The cold station is decoupled into a refrigeration host cabin, a water treatment cabin and a cold storage cabin module, the outer walls of the modules are integrated with a flange waterway interface seat, an aviation plug cable interface seat and an industrial Ethernet interface seat, and the panels are arranged in a coplanar manner; after the panels are subjected to the factory airtight and electrical tests, six-dimensional poses are written into Internet of Things tags; a digital twin model containing settlement and horizontal deviation is established by measuring and mapping a base platform on site; after hoisting and positioning, the tag pose and the base deviation are read, a target pose, a leg extension amount and a lateral fine adjustment amount are solved in a feedforward manner, and a hydraulic leveling leg and a lateral fine adjustment device are driven to cooperatively act, so that the waterway is sealed and connected, the cable joint is inserted and locked, and the communication network is connected, thereby reducing the on-site measurement and wiring and debugging work.

Description

Technical Field

[0001] This invention relates to the field of refrigeration plant assembly technology. More specifically, this invention relates to a modular adaptive assembly method for a fully pre-refrigerated plant. Background Technology

[0002] As data centers and other scenarios place increasing demands on the construction cycle and engineering quality of cooling sources, the practice of dividing cooling stations into prefabricated modules according to their functions, completing the installation of internal equipment and pipelines in the factory, and then transporting them to the site for assembly is gradually being adopted.

[0003] Application document with publication number CN119167472A discloses a modular installation method and system for prefabricated computer rooms based on digital twins. The scheme first establishes a building information model and divides the installation into modules based on it. Identifiable marks are set on the surface of the modules. The installation process of the modules is monitored on site with the help of the digital twin model. The data collected on site is compared with the digital twin model through three-dimensional laser scanning, and the installation position of the modules is adjusted accordingly. The pipelines between the modules are connected by a positioning quick-connect structure.

[0004] In the above scheme, the installation position of the module still needs to be repeatedly adjusted on site after scanning and comparison with the model, which is highly dependent on the on-site measurement conditions and working space; the water pipes, cables and communication lines between the modules are connected separately, and the workload of on-site wiring and debugging is large; after the module is in place, the setting of the cooling plant operating parameters usually lags behind the assembly process, and it takes a period of operation before it can enter a stable working condition.

[0005] Therefore, providing a fully pre-cooling station compartmentalized adaptive assembly method that can reduce on-site operation difficulty, improve module interface docking accuracy, and enable the refrigeration station to quickly enter stable operating conditions after assembly has become a pressing technical problem to be solved in this field. Summary of the Invention

[0006] One objective of this invention is to provide a modular adaptive assembly method for a fully pre-cooled station, in order to solve the problems in the prior art where module interface docking relies on repeated on-site measurements and adjustments, on-site wiring and debugging involve a large workload, and the station takes a long time to reach a stable operating condition after assembly.

[0007] To achieve the above objectives, the present invention provides a fully pre-cooled station compartmentalized adaptive assembly method, comprising the following steps:

[0008] S1. A prefabricated module decouples the chiller station into a chiller compartment, a water treatment compartment, and a cold storage compartment. Each module has an integrated panel on its outer wall. The panel is composed of a flange water interface seat, an aviation plug cable interface seat, and an industrial Ethernet interface seat. Each module is equipped with hydraulic leveling legs and a lateral fine-tuning device at the bottom. A locking mechanism is provided between adjacent modules.

[0009] S2. Before leaving the factory, the airtightness and electrical circuit tests of the panels are completed, and the six-dimensional pose of each panel is measured in the module reference coordinate system using a 3D scanner. R0 and t0 are the rotation matrix and translation vector of the panel coordinate system relative to the module reference coordinate system, respectively, and the test results are written into the IoT tag along with T0.

[0010] S3. Surveying and mapping basic platform, establishing a system including settlement d at installation points. i With levelness deviation θ i Digital twin model;

[0011] S4. After hoisting and positioning, read T0 and d. i θ i After a unified coordinate transformation, the two module panels are constrained by the coincidence of their normal directions and circumferential alignment at the interface. Feedforward calculation of the target pose of the panel and the extension / retraction of the j-th hydraulic leveling outrigger Along with the lateral fine-tuning, ΔT is the pose correction matrix, and Δz, Δα, and Δβ are the vertical axis displacement deviation and the rotation angle deviation around the x-axis and y-axis, respectively. j y j Using the coordinates of the outrigger bottom surface, the feedforward preset rear drive hydraulic leveling outrigger and the lateral fine adjustment device work together to ensure that the six-dimensional deviation of the mating surface of the two module panels does not exceed the set threshold.

[0012] S5. After the locking mechanism is tightened, the waterway is sealed, the cable connector is inserted and locked, and the communication network is connected.

[0013] Preferably, the extension and retraction of each hydraulic leveling outrigger is calculated based on the outrigger reaction force balance, provided that the target orientation is met. The result is the reaction force balance term ΔL. j,bal :

[0014] Let the vertical stiffness of the j-th hydraulic leveling outrigger be c. j The initial reaction force is F j0 The vertical stiffness is determined by the effective area of ​​the piston in the outrigger and the volumetric elastic modulus of the hydraulic oil, and the initial reaction force F j0 The reaction force was measured by pressure sensors located between each leg and the bottom surface of the module. ,by For constraints, solve for Take the minimum value of each ΔL j,bal Σ represents the summation over all hydraulically leveling outriggers, G represents the module's self-weight, and the origin of the coordinate system is the projection of the module's center of gravity onto the bottom surface.

[0015] Preferably, when the coordinate x of the module's center of gravity relative to the centroid of each hydraulic leveling outrigger is... G y G The determined eccentricity When the value exceeds the set threshold, the obtained expansion / contraction amounts ΔL are used.j For the attitude term ΔL j,pose Solve Take the minimum value of each expansion / contraction amount ΔL j,opt J is the objective function, λ is the weight, and λ0 and e0 are set constants, and e is the eccentricity of the module's center of gravity relative to the centroid of each hydraulic leveling outrigger. When e increases, λ increases and the weight of the reaction force balance term increases accordingly.

[0016] Preferably, after step S5 is completed, the pump frequency and valve opening are calculated using the real-time sensor data of the assembled chiller plant as input, in the following manner:

[0017] COP is the total energy efficiency ratio of a cooling plant. Taking the maximum value as the condition, Q is the real-time cooling load, and P is the maximum value. chiller P pump P fan These are the power ratings of the refrigeration unit, water pump, and fan, respectively.

[0018] Using pump frequency f and valve opening φ as decision variables, according to , The system is updated online, with η being the iteration step size. kf and kφ are the partial derivatives of COP with respect to the pump frequency f and valve opening φ, respectively. The controller outputs the converged pump frequency and valve opening to the frequency converter and electric regulating valve for execution. The converged pump frequency and valve opening under each load condition are stored in the controller's storage unit as the corresponding operating control parameters for that load condition.

[0019] Preferably, step S2, measuring the six-dimensional pose T0 of the panel, includes:

[0020] The panel is scanned with a 3D scanner in the module reference coordinate system to obtain the point cloud data of each feature point on the panel, and the coordinates of each feature point are extracted from the point cloud data.

[0021] On the substrate, select m feature points: the center of the flange end face of the flange water interface seat, the center of the positioning hole of the aviation plug cable interface seat, and the center of the industrial Ethernet interface seat. m is an integer from 3 to 8. Measure the coordinates p of each feature point in the module reference coordinate system. k k=1~m, take three non-collinear points p1, p2, and p3, which are located on the flange water interface, the aviation plug cable interface, and the industrial Ethernet interface respectively. , , Construct a panel coordinate system, where u, v, and w are three orthogonal unit basis vectors of the panel coordinate system. Let... , ,have to Solving makes Take the minimum values ​​of R0, t0, and q.k The known coordinates of the feature point in the panel coordinate system;

[0022] Write the panel number and the pairing number of adjacent modules into the IoT tag.

[0023] Preferably, after transportation and hoisting, the actual position of the module's panel relative to the factory-calibrated T0 undergoes deflection deformation caused by its own weight and outrigger constraints, which is mitigated by the reaction force F of each hydraulically leveling outrigger. j The load vector F and the equivalent stiffness matrix K of the module are calculated. The equivalent stiffness matrix K is obtained from the finite element model of the module structure. The deformation δ at the outrigger is obtained, and δ is converted into the pose correction ΔT of the panel mating surface. def ,make And based on the measured values ​​of the hydraulic leveling outrigger displacement sensors, the actual position T of the panel is calculated. c ,when When it exceeds the set threshold, use T c Update T0 in the IoT tag, otherwise retain T0. This threshold is determined based on the resolution of the hydraulic leveling outrigger displacement sensor and the allowable surface difference between the panel mating surface and the actual surface.

[0024] Preferably, the feedforward solution in step S4 includes:

[0025] Let the deviation of the module from the design pose given by the digital twin model after hoisting be . Δx and Δy are the displacement deviations along the x-axis and y-axis of the module's reference coordinate system, Δγ is the rotation angle deviation around the vertical axis, and Δα and Δβ are positive with the right-hand screw direction around the x-axis and y-axis.

[0026] Let the factory calibration pose of the i-th module panel be T. 0i The pose correction matrix is ​​ΔT i The relative pose of the target between adjacent i-th and i+1-th module panels Actual relative pose , with T′ rel Given that the normal directions of the docking surfaces coincide and the circumferential directions are aligned, we solve the following simultaneous equations for Δx, Δy, Δz, Δα, Δβ, and Δγ, which are constraints.

[0027] Lateral fine-tuning measures Δx and Δy without additional displacement due to rotation, rotation measures Δγ, and hydraulic leveling outrigger extension and retraction measures are solved by inversely solving Δz, Δα, and Δβ.

[0028] Preferably, when the hydraulic leveling outriggers cause the module to rotate at angles Δα and Δβ, the panel located at a height h above the bottom reference of the module will subsequently experience additional horizontal displacement. h is the height from the panel to the bottom surface of the module; the lateral fine-tuning amount is as follows: ΔX and ΔY are defined as the components of the lateral fine-tuning amount in the x-axis and y-axis directions of the module's reference coordinate system, respectively. The lateral fine-tuning device is executed first, followed by the hydraulic leveling outrigger. During the movement of the hydraulic leveling outrigger, the lateral fine-tuning device compensates synchronously according to ΔX and ΔY.

[0029] Preferably, the establishment of the digital twin model in step S3 includes: measuring the elevation z(x,y) of each measuring point on the base platform using a 3D scanner according to a grid, and comparing z(x,y) with the design elevation z c Subtracting (x,y) gives the settlement. , z c (x, y) represents the design elevation of the basic platform; the average settlement of each measuring point in the installation area of ​​the i-th module is taken. Σ represents the summation of all measuring points within the installation area, and N represents the total number of measuring points within the installation area; the length L of the line connecting installation point i and its adjacent installation point i+1 is... i levelness deviation L i Let d be the length of the line connecting installation point i and installation point i+1; let d be the length of the line connecting each installation point. i θ i The location of the data is written into the digital twin model.

[0030] Preferably, the panel used in the method includes a base plate rigidly connected to the outer wall of the module. A flange water interface seat, an aviation plug cable interface seat, and an industrial Ethernet interface seat are fixed on the same mounting surface of the base plate. The axes of the three interface seats are parallel to each other and their outer end faces are aligned with the same plane. A self-sealing shut-off valve is provided inside the flange water interface seat. The aviation plug cable interface seat and the industrial Ethernet interface seat are respectively provided with floating seats. The floating seats are clearance-fitted with the base plate, and their movable direction is perpendicular to the axis of the interface seat.

[0031] The present invention has at least the following beneficial effects:

[0032] First, this invention calibrates the six-dimensional pose of each panel at the factory stage and writes it into the module's IoT tag. On-site, only a 3D scanner is used to map the basic platform. By reading the pose reference in the tag and the settlement and levelness deviation of the basic platform, the target pose of the panel is obtained. This reduces the work of measuring the module interface on-site and reduces the dependence on on-site measurement conditions and working space. At the same time, the measured values ​​of the hydraulic leveling outrigger reaction force and outrigger displacement sensor are used to correct the panel deformation caused by transportation and hoisting, so that the panel pose reference remains applicable after the module is in place.

[0033] Secondly, this invention uses the normal overlap and circumferential alignment of the mating surfaces of two adjacent module panels as constraints, according to T=T0×ΔT(d i ,θ iThe feedforward calculation of the target pose of the panel and the extension and lateral fine-tuning of the hydraulic leveling legs are used. The feedforward is then preset and driven to work in coordination with the hydraulic leveling legs and the lateral fine-tuning device. This allows the docking of two adjacent module panels to meet the docking requirements in fewer actions. The extension of each leg is also solved according to the reaction force balance. When the center of gravity of the module is off-center, the weight between pose compensation and reaction force balance is adjusted. This can reduce the situation where the reaction force of a single leg is too large, and make it easier for the module to be subjected to more uniform force during the leveling process.

[0034] Third, this invention constructs a panel coordinate system using three non-collinear feature points and fits the panel pose using least squares, calculates flexural deformation using leg reaction force and equivalent stiffness matrix, solves for deviations in each direction by combining the target relative pose and the actual relative pose, and includes the horizontal displacement of the panel caused by rotation in the lateral fine-tuning amount according to ΔX=Δx+Δβ×h and ΔY=Δy-Δα×h, so that the panel pose calibration, deformation correction and deviation distribution are interconnected.

[0035] Fourth, the intelligent quick-connect panel of the present invention arranges the flange water interface seat, aviation plug cable interface seat and industrial Ethernet interface seat on the same plane. After the locking mechanism is locked, the water interface is sealed and connected, the cable connector is inserted and locked, and the communication network is connected, which can reduce the workload of wiring and debugging water pipes, cables and communication lines separately on site.

[0036] Fifth, after the assembly is completed, the present invention takes the maximum value of the total energy efficiency ratio of the chiller plant as a condition, iteratively updates the pump frequency and valve opening according to real-time sensor data, and stores the converged operating parameters under each load condition into the controller storage unit, which enables the chiller plant to enter a stable operating condition earlier after the assembly is completed.

[0037] Other advantages, objectives and features of the present invention will become apparent in part from the following description, and in part from those skilled in the art through study and practice of the invention. Attached Figure Description

[0038] Figure 1 This is an overall flowchart of the method provided by the present invention;

[0039] Figure 2 This is a schematic diagram of the structure of the intelligent quick-connect panel in the method provided by the present invention.

[0040] The markings in each of the attached figures are as follows:

[0041] 1. Base plate; 2. Flange water interface socket; 3. Aviation plug cable interface socket; 4. Industrial Ethernet interface socket; 5. Self-sealing shut-off valve; 6. Floating seat; 7. Return spring; 8. Module outer wall. Detailed Implementation

[0042] The present invention will now be described in further detail with reference to the accompanying drawings, so that those skilled in the art can implement it based on the description.

[0043] It should be noted that, unless otherwise specified, the methods described in the following embodiments are conventional methods; in the description of the present invention, the orientation or positional relationship indicated by the terms is based on the orientation or positional relationship shown in the accompanying drawings, and is only for the convenience of describing the present invention and simplifying the description, and does not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the present invention.

[0044] like Figures 1-2 As shown, the present invention provides a fully pre-cooled station compartmentalized adaptive assembly method, including the following steps:

[0045] The chiller plant is decoupled into prefabricated modules comprising a chiller compartment, a water treatment compartment, and a cold storage compartment. Each module has an integrated panel on its outer wall 8, which consists of flange water inlets, aviation plug cable inlets, and industrial Ethernet inlets arranged coplanarly. Specifically, the pre-cooling plant is functionally decoupled into three types of prefabricated modules: a chiller compartment, a water treatment compartment, and a cold storage compartment. The chiller compartment can house chiller units, the water treatment compartment can house water treatment equipment and a constant pressure water supply device, and the cold storage compartment can house cold storage equipment. Each prefabricated module has a box-shaped structure composed of a steel frame and sandwich insulation panels. The dimensions of the compartment can be determined according to transportation clearances and the layout of equipment within the module. The internal equipment and connecting pipelines of each module are installed in the factory. Intelligent systems are installed on the outer wall 8 of each module. The quick-connect panel includes a base plate 11 rigidly connected to the outer wall 8 of the module. The base plate 11 can be made of stainless steel or carbon steel. On the same mounting surface of the base plate 11, there are flange water interface 2, aviation plug cable interface 3 and industrial Ethernet interface 4 arranged coplanarly. The nominal diameter of the flange water interface 2 can be 50 to 200 mm and the nominal pressure can be 1.6 MPa. The aviation plug cable interface 3 can be a three-phase five-core structure. The protection level of the industrial Ethernet interface 4 can be IP65 or higher dustproof and water-resistant. The axes of the three interface assemblies are parallel to each other and their outer end faces are aligned with the same plane. Each module is equipped with hydraulic leveling legs and a lateral fine adjustment device at the bottom. A locking mechanism is provided between adjacent modules.

[0046] Before leaving the factory, while completing the airtightness and electrical continuity tests of the panels, a 3D scanner is used to measure the six-dimensional pose T0=[R0,t0] of each panel in the module reference coordinate system. R0 is the rotation matrix of the panel coordinate system relative to the module reference coordinate system, and t0 is the translation vector of the origin of the panel coordinate system. The airtightness test results, electrical continuity test results and T0 are written into the IoT tag. Specifically, during the airtightness test, compressed air of 0.6 to 1.0 MPa can be introduced into the flange water interface 2 and maintained for 30 minutes. During the electrical continuity test, the core wires of the aviation plug cable interface 3 and the industrial Ethernet interface 4 can be tested for continuity one by one. The pose calibration can be performed by using a 3D scanner to measure the six-dimensional pose T0 of the panel in the module reference coordinate system. The IoT tag can be an RFID tag or a QR code tag and can be fixed on the outer wall 8 of the module near the panel.

[0047] Using a 3D scanner as the basic mapping platform, an establishment was made that included the settlement d at each installation point. i With levelness deviation θ i A digital twin model is used; specifically, after the module is transported to the site, a 3D scanner is used to survey the foundation platform. The survey points can be arranged at a grid spacing of 0.5 to 1 meter. The settlement d of each installation point is obtained by subtracting the design elevation from the elevation of each survey point. i The horizontal deviation θ is obtained from the settlement difference and spacing between adjacent installation points. i ;

[0048] Read T0 and d after hoisting into place i θ i After unified coordinate transformation, with the normal direction of the mating surface of adjacent module panels coinciding and the circumferential alignment as constraints, according to T=T0×ΔT(d i ,θ i The feedforward calculation determines the target pose of the panel and the extension / retraction ΔL of the j-th hydraulic leveling outrigger. j =Δz+Δα×y j -Δβ×x j And the lateral fine-tuning amount, where ΔT is determined by d i θ i After unified coordinate transformation, its displacement and rotation components are analytically obtained based on the difference between the design elevation and the measured elevation of the foundation platform. ΔT is the pose correction matrix, and Δz, Δα, and Δβ are the vertical axis displacement deviation and the rotation deviation around the x-axis and y-axis, respectively. Δα and Δβ are positive with the right-hand screw direction around the x-axis and y-axis, respectively. j y j The bottom coordinates of the outrigger are used as the reference coordinates. The feedforward preset and then driven hydraulic leveling outrigger works in conjunction with the lateral fine-tuning device until docking is complete. Specifically, the T0 value from the IoT tag and the d value from the digital twin model are read. i θ iThe coordinates are then uniformly transformed to the on-site reference coordinate system, and the extension / retraction amount ΔL of the j-th hydraulic leveling outrigger is determined from the target pose. j The hydraulic leveling outriggers can be hydraulic cylinder type leveling outriggers, and the lateral fine adjustment device can be an electric screw fine adjustment device. The number of hydraulic leveling outriggers can be set according to the self-weight and distribution of the module, and can be 4 or 6, until the six-dimensional deviation of the mating surface of two adjacent module panels is no greater than 1mm. This threshold is determined according to the maximum allowable surface difference of the panel mating surface and the compression of the flange seal, that is, the mating is completed.

[0049] After the locking mechanism is tightened, the water interface is sealed, the cable connector is inserted and locked, and the communication network is connected. Specifically, the locking mechanism can be a hydraulic locking mechanism or a high-strength bolt locking mechanism. After tightening, the self-sealing shut-off valve 5 in the flange water interface seat 2 is opened and a sealed connection is formed. The connector of the aviation plug cable interface is inserted and locked, and the industrial Ethernet interface is connected.

[0050] The above methods can reduce the work of measuring module interfaces on site and reduce the difficulty of on-site docking, so that the interface docking of two adjacent modules can be completed in fewer actions.

[0051] In another technical solution, the extension and retraction of each hydraulic leveling outrigger is calculated based on the outrigger reaction force balance, while satisfying the aforementioned target orientation. The result is the reaction force balance term ΔL. j,bal Let the vertical stiffness of the j-th hydraulic leveling outrigger be c. j The initial reaction force is F j0 Its reaction force F j =F j0 +c j ×ΔL j,bal , with ΣF j =G、ΣF j ×x j =0、ΣF j ×y j =0 is a constraint; solve for max(F) j )-min(F j ) Take the minimum value of each ΔL j,bal Let Σ be the summation over all hydraulic leveling outriggers, G be the module's self-weight, and the origin of the coordinate system be the projection of the module's center of gravity onto the bottom surface. The above three equations limit the solution set of the outrigger reaction forces to one dimension. Within this solution set, and under the conditions that the extension / retraction of each outrigger does not exceed its upper stroke limit and the reaction force of each outrigger is not less than 10% of the module's self-weight, the solution that minimizes the range of outrigger reaction forces is unique. Specifically, each hydraulic leveling outrigger is symmetrically arranged along the long and short sides of the module's bottom surface. The number of outriggers can be 4 or 6, and the vertical stiffness c of each outrigger... j You can take 1×10 5 Up to 1×10 6The distance between the centers of two adjacent legs can be 0.6 to 0.8 times the length of the shorter side of the module, and the initial reaction force F j0 The measured reaction force of each leg after the module is in place and before leveling can be obtained. The reaction force of each leg can be measured by a pressure sensor installed between the leg and the bottom surface of the module. The pressure sensor can be a spoke-type force sensor. The least squares iterative solution is used with the extension and retraction of each leg as the variable. The iteration is terminated when the difference between the ranges of the reaction forces of each leg in two adjacent iterations is less than 0.5 kN. After the reaction forces of each leg tend to be balanced, the situation of excessive force on a single leg is reduced accordingly.

[0052] In another technical solution, based on the above solution of balancing the reaction forces of the outriggers, the coordinate x of the module's center of gravity relative to the centroid of each hydraulically leveling outrigger is... G y G The determined eccentricity When the value exceeds the set threshold, the obtained expansion / contraction amounts ΔL are used. j For the attitude term ΔL j,pose Solve for J = Σ(ΔL) j,opt -ΔL j,pose ) 2 +λ×[max(F j )-min(F j The minimum value of each expansion / contraction amount ΔL is taken. j,opt J is the objective function, λ is the weight and λ = λ0 × e / (e0 + e), λ0 and e0 are set constants, and e is the eccentricity of the module's center of gravity relative to the centroids of each hydraulic leveling outrigger. When e increases, λ increases and the weight of the reaction force balance term increases accordingly. Specifically, the coordinate x of the module's center of gravity on the bottom surface is... G y G The threshold value can be calculated from the mass and installation position of each device within the module. This threshold can be taken as 50mm and determined according to the allowable surface difference of the panel mating surface and the leg stroke. λ0 can be taken as 1 to 5, and e0 can be taken as 0.02 to 0.05 meters. When the eccentricity e is not greater than this threshold, the attitude term is directly taken as each ΔL. j When the eccentricity is large, the weight increases accordingly, thus ensuring a balance between panel docking and leg stress distribution.

[0053] In another technical solution, after the locking step is completed, the pump frequency and valve opening are calculated using real-time sensor data from the assembled chiller plant as input, in the following manner:

[0054] COP is the total energy efficiency ratio of a cooling plant. Taking the maximum value as the condition, Q is the real-time cooling load, and P is the maximum value. chiller P pump P fan The power of the chiller, water pump, and fan are respectively; with the water pump frequency f and valve opening φ as decision variables, the formula is f = f + η × f、φ=φ+η× The online iterative update φ is the iteration step size, kf and kφ are the partial derivatives of COP with respect to pump frequency f and valve opening φ, respectively. The controller outputs the converged pump frequency and valve opening to the frequency converter and electric regulating valve for execution. The converged pump frequency and valve opening under each load condition are stored in the controller's storage unit as the corresponding operating control parameters for that load condition. Specifically, the iteration step size η can be 0.001 to 0.01 and determined according to the iteration convergence speed and the allowable adjustment range of the operating parameters. The real-time cooling load Q can be measured by the flow meter and supply and return water temperature difference sensor installed on the chilled water supply main. Each power can be measured by the corresponding power transmitter. The sampling period for each power and cooling load can be 10 to 60 seconds. The moving average is then substituted into the iterative formula. The load condition can be divided according to the cooling load range. The range width can be 10% to 20% of the chiller station's design load. After assembly, the operating parameters are tuned, and the time for the chiller station to enter a stable operating condition is correspondingly shortened.

[0055] In another technical solution, measuring the six-dimensional pose T0 of the panel includes:

[0056] The panel is scanned with a 3D scanner in the module reference coordinate system to obtain the point cloud data of each feature point on the panel, and the coordinates of each feature point are extracted from the point cloud data.

[0057] On the substrate 11, take m feature points, including the center of the flange end face of the flange water interface seat 2, the center of the positioning hole of the aviation plug cable interface seat 3, and the center of the industrial Ethernet interface seat 4. m is an integer from 3 to 8. Measure the coordinates p of each feature point in the module reference coordinate system. k Let k = 1~m. Take three non-collinear points p1, p2, and p3, located on flange water interface 2, aviation plug cable interface 3, and industrial Ethernet interface 4, respectively. Construct a panel coordinate system according to u = (p2-p1) / |p2-p1|, v = (p3-p1) / |p3-p1|, and w = (u×v) / |u×v|, where u, v, and w are the three orthogonal unit basis vectors of the panel coordinate system. Let R0 = [u,v,w] and t0 = p1, then T0 = [R0,t0]. Solve for Σ|(p k -t0)-R0×q k | 2 Take the minimum values ​​of R0, t0, and q. k The known coordinates of the feature points in the panel coordinate system are given. Specifically, the solution is obtained by least squares iteration using the aforementioned R0 and t0 as initial values. When the difference between the target values ​​of two adjacent iterations is less than 1 × 10... -6 The iteration terminates when the convergence threshold is reached; this threshold can be 1×10. -6 Up to 1×10 -5After calibration, the residuals of each feature point are checked with a 3D scanner. The calibration is considered valid when the residuals of each feature point are not greater than 0.05mm. The panel number and the docking pairing number of adjacent modules are written into the IoT tag. The panel pose obtained by fitting multiple feature points is consistent with the actual installation state of the panel.

[0058] In another technical solution, after the panel is factory-calibrated, the module undergoes transportation and hoisting. The actual orientation of the panel relative to the factory-calibrated T0 undergoes deflection deformation caused by its own weight and the constraints of the outriggers. This deformation is mitigated by the reaction force F of each hydraulically leveling outrigger. j The resulting load vector F and the module equivalent stiffness matrix K are related by δ=K -1 The deformation δ at the support leg is obtained by multiplying by F, and then δ is converted into the pose correction amount ΔT of the panel mating surface. def Let T = T0 × ΔT def ×ΔT(d i ,θ i The actual position T of the panel is calculated by inversely based on the measured values ​​of the hydraulic leveling outrigger displacement sensor. c When |T c -T0| When it is greater than the set threshold, T c Update T0 within the IoT tag; otherwise, retain T0. Specifically, the module's equivalent stiffness matrix K can be calculated from the finite element model of the module structure. The conversion from δ to the panel mating surface can be based on the geometric relationship between the panel and the support leg. This threshold can be 0.5mm and determined according to the resolution of the hydraulic leveling support leg displacement sensor and the allowable surface difference of the panel mating surface. The displacement sensor can be a magnetostrictive displacement sensor installed on the hydraulic leveling support leg. The panel pose reference remains usable after transportation and hoisting.

[0059] In another technical solution, the feedforward solution includes:

[0060] Let the deviation of the module from the designed pose, as given by the digital twin model after hoisting, be Δ=[Δx,Δy,Δz,Δα,Δβ,Δγ]. T Δx and Δy are the displacement deviations along the x-axis and y-axis of the module's reference coordinate system, respectively; Δγ is the rotation angle deviation about the vertical axis; Δα and Δβ are positive with the right-hand screw direction about the x-axis and y-axis, respectively. Let the factory calibration pose of the i-th module panel be T. 0i The pose correction matrix is ​​ΔT i The relative pose T of the target between two adjacent module panels i and i+1 rel =(T 0i ) -1 ×T 0(i+1) The actual relative pose T′ rel =(T 0i ×ΔT i ) -1 ×(T0(i+1) ×ΔT i+1 ), with T′ rel The normal alignment and circumferential alignment of the docking surfaces are used as constraints to solve for Δx, Δy, Δz, Δα, Δβ, and Δγ. Lateral fine-tuning measures Δx and Δy without additional displacement due to rotation, while rotation measures Δγ. The extension and retraction of the hydraulic leveling outriggers are solved by inversely applying Δz, Δα, and Δβ. Specifically, the above simultaneous solution can be performed according to rigid body coordinate transformation relationships. The portion of |Δx| and |Δy| not exceeding 0.5mm to 2mm can be handled by the lateral fine-tuning device, while the excess portion can be compensated by the rotational movement of the hydraulic leveling outriggers. Δz, Δα, and Δβ can be distributed to each hydraulic leveling outrigger according to the principle that the difference in extension and retraction between each outrigger does not exceed 5% of the outrigger's stroke. The above calculations and actions can be executed cyclically with a control cycle not exceeding 2 seconds, stopping when the six-dimensional deviations of the docking surfaces of the two module panels are not greater than the set values. After the target relative pose and the actual relative pose are combined, the deviations in each direction can be directly distributed to the corresponding actuators.

[0061] In another technical solution, based on the deviations obtained from the feedforward calculation, when the hydraulic leveling outrigger causes the module to rotate at angles Δα and Δβ, the panel located at a height h above the module's bottom reference will experience an additional horizontal displacement (Δβ×h, -Δα×h), where h is the height of the panel from the module's bottom reference. The lateral fine-tuning is set according to ΔX=Δx+Δβ×h and ΔY=Δy-Δα×h, where ΔX and ΔY are the components of the lateral fine-tuning in the x-axis and y-axis directions of the module's reference coordinate system, respectively. The lateral fine-tuning is performed in sequence, with the hydraulic leveling outrigger following suit. During the movement of the hydraulic leveling outrigger, the lateral fine-tuning device compensates synchronously for ΔX and ΔY. Specifically, the height h can be 1 to 3 meters. After the horizontal displacement of the panel caused by the rotation is included in the lateral fine-tuning, the lateral fine-tuning and the rotation of the outrigger are coordinated.

[0062] In another technical solution, establishing a digital twin model includes: using a 3D scanner to measure the elevation z(x,y) of each measuring point on the base platform according to a grid, and comparing z(x,y) with the design elevation z c Subtracting (x,y) from z, we get the settlement d(x,y) = z(x,y) - z c (x,y), z c (x, y) represents the design elevation of the basic platform, and d is the average settlement of each measuring point in the installation area of ​​the i-th module. i =Σd(x,y) / N, where Σ is the summation of all measuring points within the installation area, and N is the total number of measuring points within the installation area. The length L of the line connecting installation point i and its adjacent installation point i+1 is... i Levelness deviation θ i =arctan[(d i+1 -d i ) / L i], L i Let d be the length of the line connecting installation point i and installation point i+1. i θ i The location of the data is written into the digital twin model; specifically, the measurement points of the 3D scanner can be arranged at a grid spacing of 0.5 to 1 meter. The settlement and levelness deviations calculated from the mapping data ensure that the digital twin model matches the actual state of the base platform.

[0063] This invention also provides an intelligent quick-connect panel for implementing the above-mentioned fully pre-cooled station compartmentalized adaptive assembly method, including a base plate 11 rigidly connected to the outer wall 8 of the module. A flange water interface seat 2, an aviation plug cable interface seat 3, and an industrial Ethernet interface seat 4 are fixed on the same mounting surface of the base plate 11. The axes of the three interface seats are parallel to each other, and their outer end faces are aligned with the same plane. A self-sealing shut-off valve 5 is provided inside the flange water interface seat 2. A floating seat 6 is provided for the aviation plug cable interface seat 3 and the industrial Ethernet interface seat 4, respectively. The floating seat 6 is clearance-fitted with the base plate 11, and its movable direction is perpendicular to the axis of the interface seat. Specifically, the base plate 11 can... The plates used are stainless steel or carbon steel, with gaps ranging from 0.5 to 2 mm. The flange water interface seat 2 can be welded to the base plate 11. The aviation plug cable interface seat 3 and the industrial Ethernet interface seat 4 can be fixed to the base plate 11 by screws. The base plate 11 can be connected to the outer wall 8 of the module by welding or bolts. A return spring 7 can be provided between the floating seat 6 of the aviation plug cable interface seat 3 and the base plate 11 to allow the floating seat 6 to return to the middle position after the plug is inserted. In the above technical solution, the three interface seats are arranged in the same plane, and the aviation plug cable interface seat 3 and the industrial Ethernet interface seat 4 have a certain amount of mobility to facilitate insertion during docking.

[0064] To verify the above implementation method, an experiment can be conducted using the panel docking of two adjacent prefabricated modules. After docking, a coordinate measuring machine (CMM) is used to measure the relative positions of each measuring point on the docking surface of the two panels, obtaining the normal and circumferential deviations of the docking surface. The docking is repeated five times, and the deviation values ​​are recorded each time, and their mean and standard deviation are calculated. For the assembled chiller plant, the pump frequency, valve opening, and overall energy efficiency ratio of the chiller plant are recorded under several load conditions to verify the convergence of the operating parameters. All testing instruments used in the experiment can be obtained commercially.

[0065] After the docking was completed, a coordinate measuring machine was used to measure the relative positions of each measuring point on the docking surfaces of the two panels, and the normal deviation and circumferential deviation of the docking surfaces were obtained. The results of the docking were repeated 5 times, as shown in Table 1.

[0066] Table 1 Panel mating deviation test data

[0067]

[0068] As shown in Table 1, the normal and circumferential deviations of each panel docking are no greater than 1 mm, and the repeatability of feedforward calculation and coordinated action meets the docking accuracy requirements.

[0069] For the assembled chiller plant, the pump frequency, valve opening, and overall energy efficiency ratio were recorded under several load conditions. The results are shown in Table 2.

[0070] Table 2 Test data of load conditions and operating parameters

[0071]

[0072] As shown in Table 2, the pump frequency and valve opening under each load condition converge after iteration. The total energy efficiency ratio of the chiller plant is highest at around 80% of the design load, indicating that the operating control parameters stored in the controller storage unit match the load conditions.

[0073] Although embodiments of the present invention have been disclosed above, they are not limited to the applications listed in the specification and embodiments. They can be applied to various fields suitable for the present invention. For those skilled in the art, other modifications can be easily made. Therefore, without departing from the general concept defined by the claims and their equivalents, the present invention is not limited to the specific details and illustrations shown and described herein.

Claims

1. A fully pre-cooled station modular adaptive assembly method, characterized in that, Includes the following steps: S1. A prefabricated module decouples the chiller station into a chiller compartment, a water treatment compartment, and a cold storage compartment. Each module has an integrated panel on its outer wall. The panel is composed of a flange water interface seat, an aviation plug cable interface seat, and an industrial Ethernet interface seat. Each module is equipped with hydraulic leveling legs and a lateral fine-tuning device at the bottom. A locking mechanism is provided between adjacent modules. S2. Before leaving the factory, the airtightness and electrical circuit tests of the panels are completed, and the six-dimensional pose of each panel is measured in the module reference coordinate system using a 3D scanner. R0 and t0 are the rotation matrix and translation vector of the panel coordinate system relative to the module reference coordinate system, respectively, and the test results are written into the IoT tag along with T0. S3. Surveying and mapping basic platform, establishing a system including settlement d at installation points. i With levelness deviation θ i Digital twin model; S4. After hoisting and positioning, read T0 and d. i θ i After a unified coordinate transformation, the two module panels are constrained by the coincidence of their normal directions and circumferential alignment at the interface. Feedforward calculation of the target pose of the panel and the extension / retraction of the j-th hydraulic leveling outrigger Along with the lateral fine-tuning, ΔT is the pose correction matrix, and Δz, Δα, and Δβ are the vertical axis displacement deviation and the rotation angle deviation around the x-axis and y-axis, respectively. j y j Using the coordinates of the outrigger bottom surface, the feedforward preset rear drive hydraulic leveling outrigger and the lateral fine adjustment device work together to ensure that the six-dimensional deviation of the mating surface of the two module panels does not exceed the set threshold. S5. After the locking mechanism is tightened, the waterway is sealed, the cable connector is inserted and locked, and the communication network is connected.

2. The fully pre-cooled station compartmentalized adaptive assembly method as described in claim 1, characterized in that, The extension and retraction of each hydraulic leveling outrigger are calculated based on the outrigger reaction force balance, provided that the target pose is met. The result is the reaction force balance term ΔL. j,bal : Let the vertical stiffness of the j-th hydraulic leveling outrigger be c. j The initial reaction force is F j0 The vertical stiffness is determined by the effective area of ​​the piston in the outrigger and the volumetric elastic modulus of the hydraulic oil, and the initial reaction force F j0 The reaction force was measured by pressure sensors located between each leg and the bottom surface of the module. ,by For constraints, solve for Take the minimum value of each ΔL j,bal Σ represents the summation over all hydraulically leveling outriggers, G represents the module's self-weight, and the origin of the coordinate system is the projection of the module's center of gravity onto the bottom surface.

3. The fully pre-cooled station compartmentalized adaptive assembly method as described in claim 2, characterized in that, When the coordinate x of the module's center of gravity relative to the centroid of each hydraulic leveling outrigger... G y G The determined eccentricity When the value exceeds the set threshold, the obtained expansion / contraction amounts ΔL are used. j For the attitude term ΔL j,pose Solve Take the minimum value of each expansion / contraction amount ΔL j,opt J is the objective function, λ is the weight, and λ0 and e0 are set constants, and e is the eccentricity of the module's center of gravity relative to the centroid of each hydraulic leveling outrigger. When e increases, λ increases and the weight of the reaction force balance term increases accordingly.

4. The fully pre-cooled station compartmentalized adaptive assembly method as described in claim 1, characterized in that, After step S5 is completed, using the real-time sensor data of the assembled cooling plant as input, the pump frequency and valve opening are calculated as follows: COP is the total energy efficiency ratio of a cooling plant. Taking the maximum value as the condition, Q is the real-time cooling load, and P is the maximum value. chiller P pump P fan These are the power ratings of the refrigeration unit, water pump, and fan, respectively. Using pump frequency f and valve opening φ as decision variables, according to , The system is updated online, with η being the iteration step size. kf and kφ are the partial derivatives of COP with respect to the pump frequency f and valve opening φ, respectively. The controller outputs the converged pump frequency and valve opening to the frequency converter and electric regulating valve for execution. The converged pump frequency and valve opening under each load condition are stored in the controller's storage unit as the corresponding operating control parameters for that load condition.

5. The fully pre-cooled station compartmentalized adaptive assembly method as described in claim 1, characterized in that, Step S2, measuring the six-dimensional pose T0 of the panel, includes: The panel is scanned with a 3D scanner in the module reference coordinate system to obtain the point cloud data of each feature point on the panel, and the coordinates of each feature point are extracted from the point cloud data. On the substrate, select m feature points: the center of the flange end face of the flange water interface seat, the center of the positioning hole of the aviation plug cable interface seat, and the center of the industrial Ethernet interface seat. m is an integer from 3 to 8. Measure the coordinates p of each feature point in the module reference coordinate system. k Let k = 1~m. Take three non-collinear points p1, p2, and p3, located on the flange water interface, the aviation plug cable interface, and the industrial Ethernet interface, respectively. , , Construct a panel coordinate system, where u, v, and w are three orthogonal unit basis vectors of the panel coordinate system. Let... , ,have to Solving makes Take the minimum values ​​of R0, t0, and q. k The known coordinates of the feature point in the panel coordinate system; Write the panel number and the pairing number of adjacent modules into the IoT tag.

6. The fully pre-cooled station compartmentalized adaptive assembly method as described in claim 5, characterized in that, After transportation and hoisting, the actual position of the module's panel relative to the factory-specified T0 undergoes deflection deformation caused by its own weight and outrigger constraints, which is addressed by the reaction force F of each hydraulically leveling outrigger. j The load vector F and the equivalent stiffness matrix K of the module are calculated. The equivalent stiffness matrix K is obtained from the finite element model of the module structure. The deformation δ at the outrigger is obtained, and δ is converted into the pose correction ΔT of the panel mating surface. def ,make And based on the measured values ​​of the hydraulic leveling outrigger displacement sensors, the actual position T of the panel is calculated. c ,when When it exceeds the set threshold, use T c Update T0 in the IoT tag, otherwise retain T0. This threshold is determined based on the resolution of the hydraulic leveling outrigger displacement sensor and the allowable surface difference between the panel mating surface and the actual surface.

7. The fully pre-cooled station compartmentalized adaptive assembly method as described in claim 1, characterized in that, The feedforward solution in step S4 includes: Let the deviation of the module from the design pose given by the digital twin model after hoisting be . Δx and Δy are the displacement deviations along the x-axis and y-axis of the module's reference coordinate system, Δγ is the rotation angle deviation around the vertical axis, and Δα and Δβ are positive with the right-hand screw direction around the x-axis and y-axis. Let the factory calibration pose of the i-th module panel be T. 0i The pose correction matrix is ​​ΔT i The relative pose of the target between adjacent i-th and i+1-th module panels Actual relative pose , with T′ rel Given that the normal directions of the docking surfaces coincide and the circumferential directions are aligned, we solve the following simultaneous equations for Δx, Δy, Δz, Δα, Δβ, and Δγ, which are constraints. Lateral fine-tuning measures Δx and Δy without additional displacement due to rotation, rotation measures Δγ, and hydraulic leveling outrigger extension and retraction measures are solved by inversely solving Δz, Δα, and Δβ.

8. The fully pre-cooled station compartmentalized adaptive assembly method as described in claim 7, characterized in that, When the hydraulic leveling outriggers cause the module to rotate at angles Δα and Δβ, the panel located at a height h above the bottom reference of the module will experience additional horizontal displacement. h is the height from the panel to the bottom surface of the module; the lateral fine-tuning amount is as follows: ΔX and ΔY are defined as the components of the lateral fine-tuning amount in the x-axis and y-axis directions of the module's reference coordinate system, respectively. The lateral fine-tuning device is executed first, followed by the hydraulic leveling outrigger. During the movement of the hydraulic leveling outrigger, the lateral fine-tuning device compensates synchronously according to ΔX and ΔY.

9. The fully pre-cooled station compartmentalized adaptive assembly method as described in claim 1, characterized in that, Step S3, establishing the digital twin model, includes: using a 3D scanner to measure the elevation z(x,y) of each measuring point on the base platform according to the grid, and comparing z(x,y) with the design elevation z c Subtracting (x,y) gives the settlement. , z c (x, y) represents the design elevation of the basic platform; the average settlement of each measuring point in the installation area of ​​the i-th module is taken. Σ represents the summation of all measuring points within the installation area, and N represents the total number of measuring points within the installation area; the length L of the line connecting installation point i and its adjacent installation point i+1 is... i levelness deviation L i Let d be the length of the line connecting installation point i and installation point i+1; let d be the length of the line connecting each installation point. i θ i The location of the data is written into the digital twin model.

10. The fully pre-cooled station compartmentalized adaptive assembly method as described in claim 1, characterized in that, The panel used in the method includes a base plate rigidly connected to the outer wall of the module. A flange water interface seat, an aviation plug cable interface seat, and an industrial Ethernet interface seat are fixed on the same mounting surface of the base plate. The axes of the three interface seats are parallel to each other and their outer end faces are aligned with the same plane. A self-sealing shut-off valve is provided inside the flange water interface seat. The aviation plug cable interface seat and the industrial Ethernet interface seat are respectively provided with floating seats. The floating seats are clearance-fitted with the base plate, and their movable direction is perpendicular to the axis of the interface seat.

Images