Iterative damping growth model for mining

1.

1 Parameter fitting principle

after obtaining the relationship equation between single well water inflow and various geophysical logging parameters, it can be found that there are many undetermined constants, which are different in different places. In order to determine these coefficients, it is necessary to obtain single well water inflow and corresponding logging parameters in this area, Then, the undetermined parameters corresponding to this area are obtained by fitting, which is called parameter fitting. The fitting method used in this program is the modified damped least square method for multi parameter data fitting < sup > [14] < / sup >. The principle of fitting method is introced below

set the calculated I water volume of the first hole in the above model is q_i. The measured unit water inflow of pumping is Q < sub > J < / sub >. It can be seen from the above models that Q < sub > J < / sub > is a nonlinear multivariate function, so the following two functions are used as the objective function. It is appropriate to use the optimization method to obtain the undetermined coefficients of the selected model

(A+ λ< sup>2K) Δ P = B

(1) the objective function takes the square sum of the relative error of unit water inflow of each well

comprehensive geophysical prospecting technology for aquifer water content prediction

where: λ Is the damping coefficient

(2) the objective function is the sum of squares of the absolute error of unit water inflow of each well. If the difference of water inflow is large, and the prediction accuracy of smaller water inflow is higher, the sum of squares of relative error should be selected as the objective function. At this time, although the prediction accuracy of small water inflow drilling is improved, the prediction accuracy of large water inflow drilling is relatively reced. If the change of water inflow is small, and it is not required to have the same high prediction accuracy as that of large water inflow drilling, the square sum of absolute error is suitable as the objective function. The fitting process is shown in Figure 5-4

(2) model construction

using the least square criterion, the values of the model coefficients a, B, C, D, e, F, G, R to be calculated should make the objective function take the minimum value. Obviously, this is a problem of solving the least squares minimization of nonlinear multivariate function, which can be solved by Marquette's method (or damping least squares method), which is more effective in optimization methods. Usually, the undetermined coefficients of each model can be obtained after several iterations

Maquette method is a more effective algorithm to find the least square minimum solution in optimization. It converges faster than gradient method and conjugate gradient method, and is more stable than Gauss Newton method, so it has been widely used in many other inversion interpretation

in the classical Marquette algorithm, when the vector composed of model coefficients and the elements of its correction are very different from each other, the damping coefficient will be larger, which will increase the number of iterations and rece the operation speed. At the same time, it also requires that the initial value of model coefficients should be close to the minimum point, otherwise it is not easy to converge, that is, the stability is not ideal. Therefore, we use the method of weighted damping factor, that is, the unit matrix K in the classical Marquette equation is modified to the diagonal matrix K related to the size of the model coefficients, and the effect is that the model coefficients are large and the damping is small; The model coefficient is small and the damping is large. Thus, the coefficients of each model converge to the minimum point at the same speed, and the operation speed and stability of the algorithm are improved, The equation is

comprehensive geophysical prospecting technology for aquifer water content prediction

Fig. 5-4 flow chart of multi parameter fitting

comprehensive geophysical prospecting technology for aquifer water content prediction, The model of aquifer water content prediction using resistivity parameters of geophysical logging is obtained

2. The concept of damping is introced in the first chapter of mechanics of college physics. The effect of damping is force, but damping is not equal to resistance. If the damper is linear, the damping force F = damping coefficient * DX / dt. X is the moving distance. If the damper is rotating. Damping torque = damping coefficient * DW / dt. X is the angular velocity. for instance. A motor with a rotating shaft in the middle

3. Unknown_Error

4. Spring and block, because in the whole system, spring and block can store the total mechanical energy of the system

5. The next step is to select the amplitude curve below when defining the load. It should be clear that the amplitude curve needs to be closely connected with the load

7. Damping refers to the physical phenomenon that the body is subject to resistance in the process of motion, its energy graally decreases and the motion graally weakens

the physical model focuses on the damping coefficient, which has no fixed symbol.

8. Of course, spring can also be regarded as the damping of a motion process. The concept of damping is a little larger than that of spring. But it is more convenient to separate the spring, because the spring is not energy dissipation, it just stores the energy and can also be converted into other forms of energy, but the general damping is dissipation, energy is converted into heat, in the treatment of mechanical model, this part of energy can not be reused.

9. The mechanics library file of Visio is made by ourselves,