CN104503396A - Multi-metal balance yield algorithm - Google Patents

Abstract

The invention discloses a multi-metal balance yield algorithm, relates to the yield algorithm of multiple types of valuable elements in mines, mineral dressing and the metallurgical flow production process, and belongs to the technology of the field of metal recovery. According to the algorithm, a nonlinear multi-constrained optimization calculation evaluation model is established according to a metal balance equation and a yield balance equation, and a concrete solving algorithm is designed. The algorithm is enabled to possess advantages of high yield calculation precision, low calculation amount and high speed. Besides, the algorithm is high in generality, suitable for various processes of material yield calculation, ingenious in conception, accurate and practical, high in cost performance and beneficial for large-scale popularization and application.

Description

A kind of algorithm of many metal balances productive rate

Technical field

The present invention relates to the algorithm of multiple types valuable element productive rate in mine, ore dressing and metallurgical process production run, belong to metal recovery art, specifically a kind of algorithm of many metal balances productive rate.

Background technology

The valuable element balance statistics measuring and calculating of mining production, ore-dressing practice, metallurgical process usually needs the concrete trend of inputoutput when material distribution calculating valuable element.Due in production operation process, the randomness that laboratory analysis of data obtains, the existence of the factors such as sampling process, sample making course, data analysis procedure, technological process of production fluctuation, the nonrepeatability of sampling process and the run, drip, leak of flow process, the various error accumulations that this uncertainty by analytical test data, inexactness produce, final theoretical yield is the external manifestation of the ornamental at the accumulation center of various error.But due to the restriction of measuring condition, the data volume that general scene can record is very limited, and due to the problem of measuring method and sample representation, measurement result also often exists no small error, the data with error utilizing these limited are difficult to play effective directive function to production.

For this abnormal problem occurred in the metal balance computation process of multiple types valuable element, from nineteen seventies, external main mining company and scientific research institutions drop into the research and development that a large amount of manpower and materials carry out related algorithm model, through repeatedly rig-site utilization checking, all achieve certain achievement and be integrated in the mining industry Advanced Control Software platform released separately, be concerned about immeasurablel data are extrapolated by measurable data, and due to measurable data because the impact of error might not be credible, therefore the correction carrying out to a certain degree to it is needed if desired.The key of this problem is, the reckoning carried out must be rational with revising, otherwise may cause the pernicious result misleading and produce.Therefore, how reasonably processing compared with the measured data of big error existence, extrapolate other be concerned about data, and real result being credible, is a problem highly studied.

For the balance on-line realizing valuable element calculates, existing MaterialBalance Computation method cannot reduce or eliminate the conventional materials level theory calculation of yield abnormal variation because production procedure fluctuation and human factor occur and utilize multiple constraint target to attempt beneficiation flowsheet nonlinear solution method, makes solving complexity quite high.For realize mass balance model in line computation, reduce the complexity of algorithm model, improve the online counting yield of mass balance model, be necessary to utilize the restrained boundary condition of material balance expression formula solution nonuniqueness and production procedure to be revised analytical test data by multi-target non-linear contained optimization, carry out material balance and the calculating of productive rate balance model on this basis.

Summary of the invention

Technical matters to be solved by this invention sets up the quick Calculation Estimation function model of material balance and constraint condition, when DATA REASONING exists compared with big error at the scene, can automatically repair irrational measured value, and provide the quick calculation method solving material balance productive rate.

For achieving the above object, the present invention takes following technical scheme: a kind of algorithm of many metal balances productive rate, and it comprises the following steps:

1. this computation process is according to metal balance equation $\begin{array}{c}{\hat{x}}_{21}{\mathrm{\γ}}_2+{\hat{x}}_{11}{\mathrm{\γ}}_1={\hat{x}}_1\\ {\hat{x}}_{22}{\mathrm{\γ}}_2+{\hat{x}}_{12}{\mathrm{\γ}}_1={\hat{x}}_2\end{array}$ With productive rate balance equation γ ₂+ γ ₁=1, and one section of concentrate+two sections of raw ore=mono-section raw ore

One section of concentrate+two sections of concentrate=high concentrate

(hereinafter: one section of concentrate is called for short an essence, two sections of raw ores are called for short two former two sections of concentrate+low concentrate+total mine tailing=bis-section raw ore, and one section of raw ore is called for short one former, two sections of concentrate are called for short two essences, high concentrate is called for short high-precision, and low concentrate is called for short low essence, and total mine tailing is called for short total tail); Following relational expression can be obtained:

$\begin{array}{c}{\mathrm{\γ}}_1+{\mathrm{\γ}}_2=1\\ x_{11}{\mathrm{\γ}}_1+x_{12}{\mathrm{\γ}}_2=x_1\\ x_{21}{\mathrm{\γ}}_1+x_{22}{\mathrm{\γ}}_2=x_2\end{array}---\left(1\right)$

In formula, γ ₁it is concentrate yield;

γ ₂it is mine tailing productive rate;

X ₁₁it is actual plumbous grade in concentrate;

X ₁₂it is actual plumbous grade in mine tailing;

X ₂₁it is actual zinc grade in concentrate;

X ₂₂it is actual zinc grade in mine tailing;

X ₁it is actual plumbous grade in raw ore;

X ₂it is actual zinc grade in raw ore;

To in expression formula (1) any two all can obtain yield results, if the valuable element of raw ore, concentrate, mine tailing is plumbous, zinc grade is actual value, then combination of two acquired results is identical.Then often not identical according to the yield results that calculates of reality chemical examination grade.For this reason, the evaluation model optimized and calculate is set up.

The described evaluation model equation calculated of optimizing is:

$\begin{array}{c}\min {\left(\frac{x_1-{\hat{x}}_1}{{\hat{x}}_1}\right)}^2+{\left(\frac{x_2-{\hat{x}}_2}{{\hat{x}}_2}\right)}^2+{\left(\frac{x_{11}+{\hat{x}}_{11}}{{\hat{x}}_{11}}\right)}^2+{\left(\frac{x_{12}-{\hat{x}}_{12}}{{\hat{x}}_{12}}\right)}^2\\ +{\left(\frac{x_{21}-{\hat{x}}_{21}}{{\hat{x}}_{21}}\right)}^2+{\left(\frac{x_{22}-{\hat{x}}_{22}}{{\hat{x}}_{22}}\right)}^2\end{array}$

In formula, γ ₁it is concentrate yield;

γ ₂it is mine tailing productive rate;

X ₁₁it is actual plumbous grade in concentrate;

it is the plumbous grade of concentrate after adjustment;

X ₁₂it is actual plumbous grade in mine tailing;

it is the plumbous grade of mine tailing after adjustment;

X ₂₁it is actual zinc grade in concentrate;

it is the concentrate zinc grade after adjustment;

X ₂₂it is actual zinc grade in mine tailing;

it is the mine tailing zinc grade after adjustment;

X ₁it is actual plumbous grade in raw ore;

it is the plumbous grade of raw ore after adjustment;

X ₂it is actual zinc grade in raw ore;

it is the raw ore zinc grade after adjustment;

C ₁, C ₂, C ₃, C ₄, C ₅, C ₆it is the limit D-value between mineral processing production flow process raw ore, concentrate, mine tailing lead, zinc grade;

A ₁, B ₁, A ₂, B ₂be respectively the domain of variation bound of concentrate and tailings productive rate;

A ₁, a ₂be respectively the lower limit set value of raw ore lead, zinc adjustment grade;

B ₁, b ₂be respectively the upper limit set value of raw ore lead, zinc adjustment grade;

A ₁₁, a ₂₁be respectively the lower limit set value of concentrate lead, zinc adjustment grade;

B ₁₁, b ₂₁be respectively the upper limit set value of concentrate lead, zinc adjustment grade;

A ₁₂, a ₂₂be respectively the lower limit set value of mine tailing lead, zinc adjustment grade;

B ₁₂, b ₂₂be respectively the upper limit set value of mine tailing lead, zinc adjustment grade.

The algorithm of a kind of many metal balances productive rate of the present invention, adopt C/S framework, optimizing computation model operates on Terminal Server Client, production process data by OPC agreement from Process Control System OPCServer server obtain, for model calculate analytical test data from on-stream analyzer for ore grade or laboratory.When from on-stream analyzer for ore grade, grade analysis data are obtained by ModBus rtu protocol; When being inputted by typing window from during laboratory.

The device that the algorithm of a kind of many metal balances productive rate of the present invention uses, comprises production run Control Server (OPCServer), on-stream analyzer for ore grade, off-line data typing terminal, the network switch, model calculates host computer and communication line composition; Model calculation constraint condition parameter calculates the input of host computer human window by model; Take valuable element as lead, the mixing beneficiation production procedure raw ore of zinc, concentrate, mine tailing carry out model construction arthmetic statement:

The algorithm of described a kind of many metal balances productive rate, the detailed computation process of its model algorithm is as follows:

Step1: provide feasible initial point x ⁽⁰⁾, ε ₀> 0, ε > 0, k=0, the basic thought that feasible point solves is as follows: ensureing under the precondition that equality constraint meets, continuous correction initial analysis chemical examination hard goods place value, makes to calculate metal productive rate and meets prior-constrained condition, namely meet

$A_1\≤{\mathrm{\γ}}_1^{\left(0\right)}=\frac{x_1^{\left(0\right)}-x_{12}^{\left(0\right)}}{x_{11}^{\left(0\right)}-x_{12}^{\left(0\right)}}\≤B_1---\left(3\right)$

$A_2\≤{\mathrm{\γ}}_2^{\left(0\right)}=\frac{x_2^{\left(0\right)}-x_{22}^{\left(0\right)}}{x_{21}^{\left(0\right)}-x_{22}^{\left(0\right)}}\≤B_2---\left(4\right)$

1.1: according to equality constraint, calculating starting point is $x^{\left(0\right)}=\left[\begin{array}{c}{\hat{x}}_1\\ {\hat{x}}_2\\ {\hat{x}}_{11}\\ {\hat{x}}_{12}\\ {\hat{x}}_{21}\\ {\hat{x}}_{22}\\ \frac{{\hat{x}}_1-{\hat{x}}_{12}}{{\hat{x}}_{11}-{\hat{x}}_{12}}\\ \frac{{\hat{x}}_2-{\hat{x}}_{22}}{{\hat{x}}_{21}-{\hat{x}}_{22}}\end{array}\right].$

Step2: initial feasible solution x can be drawn according to Step1 ⁽⁰⁾, given ε ₀=0.1, according to the goal constraint collection of Optimization Solution problem, namely (2) constraint portions, determines ε _kactive constraint index set I (x ^(k), ε _k)={ i|0≤g _i(x ^(k))≤ε _k, 1≤i≤m}, calculates constraint set matrix with the downward gradient of objective function

$\▿g\left(x^{\left(k\right)}\right)=\left[\begin{array}{cccccccccccccccccccccccc}0& 0& -1& 0& -1& 0& 0& 0& 1& -1& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0\\ 0& 0& 0& -1& 0& -1& 0& 0& 0& 0& 1& -1& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0\\ 0& 0& 1& 0& 0& 0& 1& 0& 0& 0& 0& 0& 1& -1& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0\\ 0& 0& 0& 0& 1& 0& -1& 0& 0& 0& 0& 0& 0& 0& 1& -1& 0& 0& 0& 0& 0& 0& 0& 0\\ 0& 0& 0& 1& 0& 0& 0& 1& 0& 0& 0& 0& 0& 0& 0& 0& 1& -1& 0& 0& 0& 0& 0& 0\\ 0& 0& 0& 0& 0& 1& 0& -1& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 1& -1& 0& 0& 0& 0\\ 1& -1& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 1& -1& 0& 0\\ 1& -1& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 1& -1\end{array}\right]$

$\▿f\left(x^{\left(k\right)}\right)=\left[\begin{array}{c}\frac{2}{{\hat{x}}_1^2}{x}_1^{\left(k\right)}-\frac{2}{{\hat{x}}_1}\\ \frac{2}{{\hat{x}}_2^2}{x}_2^{\left(k\right)}-\frac{2}{{\hat{x}}_2}\\ \frac{2}{{\hat{x}}_{11}^2}{x}_{11}^{\left(k\right)}-\frac{2}{{\hat{x}}_{11}}\\ \frac{2}{{\hat{x}}_{12}^2}{x}_{12}^{\left(k\right)}-\frac{2}{{\hat{x}}_{12}}\\ \frac{2}{{\hat{x}}_{21}^2}{x}_{21}^{\left(k\right)}-\frac{2}{{\hat{x}}_{21}}\\ \frac{2}{{\hat{x}}_{22}^2}{x}_{22}^{\left(k\right)}-\frac{2}{{\hat{x}}_{22}}\\ 0\\ 0\end{array}\right]$

Step3: test condition I (g _i(x ^(k)), ε _k)=φ, if meet, and then stop iteration, x ^(k)for approximate minimal value; If then make σ _k=-1, turn Step6; If I is (g _i(x ^(k)), ε _k) ≠ φ, then turn Step4;

Step4: solve linear programming problem

min σ (5)

$s.t.\left\{\begin{array}{c}{d}^T\▿f\left(x^{\left(k\right)}\right)\≤\mathrm{\σ}\\ {-d}^T{\▿g}_i\left(x^{\left(k\right)}\right)\\ \left|d_j\right|\≤1\end{array},\≤{\mathrm{\η}}_i\mathrm{\σ},i\∈I(g_j\left(x^{\left(k\right)}\right),{\mathrm{\ϵ}}_k),j=1,\·\·\·,m\right.---\left(6\right)$

When ε active constraint is linear restriction, η _i=0; When ε active constraint is non-linear constrain, η _i=1, solution can obtain d ^(k)=d, σ _k.From the condition in constraint equation (1), η when the first two inequality constrain that and if only if is worked _i=1, i.e. the constraint condition of linear programming comprises three parts: objective function constraint, first stage ε _kthe constraint of nonlinear optimization nonvoid subset, the constraint of linear optimization vector domain of variation, wherein first stage ε _kthe constraint of nonlinear optimization nonvoid subset is only the subset I (g that nonlinear constrained optimization solves part _i(x ^(k)), ε _k);

Above-mentioned linear restriction planning can be converted into

min σ (7)

$s.t.\left\{\begin{array}{c}(\frac{2}{{\hat{x}}_1^2}{x}_1^{\left(k\right)}-\frac{2}{{\hat{x}}_1})d_1+(\frac{2}{{\hat{x}}_2^2}{x}_2-\frac{2}{{\hat{x}}_2})d_2+(\frac{2}{{\hat{x}}_3^2}{x}_3^{\left(k\right)}-\frac{2}{{\hat{x}}_3})d_3\≤\mathrm{\σ}\\ {-d}^T{\▿g}_i\left(x^{\left(k\right)}\right)\≤{\mathrm{\η}}_i\mathrm{\σ},i\∈I(g_j\left(x^{\left(k\right)}\right),{\mathrm{\ϵ}}_k),j=1,\·\·\·,m\\ \left|d_1\right|\≤1\\ \left|d_2\right|\≤1\\ \left|d_3\right|\≤1\\ \left|d_8\right|\≤1\\ \left|d_9\right|\≤1\end{array}\right.---\left(8\right)$

Above-mentioned linear programming is transformed further

min d ₁₄＝σ (9)

$s.t.\left\{\begin{array}{c}-(\frac{2}{{\hat{x}}_1^2}{x}_1^{\left(k\right)}-\frac{2}{{\hat{x}}_1})d_1-(\frac{2}{{\hat{x}}_2^2}{x}_2^{\left(k\right)}-\frac{2}{{\hat{x}}_2})d_2-(\frac{2}{{\hat{x}}_{11}^2}{x}_{11}^{\left(k\right)}-\frac{2}{{\hat{x}}_{11}})d_3+(\frac{2}{{\hat{x}}_{12}^2}{x}_{12}^{\left(k\right)}-\frac{2}{{\hat{x}}_{12}})d_4\\ +(\frac{2}{{\hat{x}}_{21}^2}{x}_{21}^{\left(k\right)}-\frac{2}{{\hat{x}}_{21}})d_5+(\frac{2}{{\hat{x}}_{22}^2}{x}_{22}^{\left(k\right)}-\frac{2}{{\hat{x}}_{22}})d_6+d_{14}\≥0\\ d_i^T{\▿g}_i\left(x^{\left(k\right)}\right)+{\mathrm{\η}}_i{d}_i\≥0,i\∈I(g_j\left(x^{\left(k\right)}\right),{\mathrm{\ϵ}}_k),j=1,\·\·\·,m\\ d_1\≥-1\\ {-d}_1\≥-1\\ d_2\≥-1\\ -d_2\≥-1\\ d_3\≥-1\\ {-d}_3\≥-1\\ d_4\≥-1\\ {-d}_4\≥-1\\ d_5\≥-1\\ {-d}_5\≥-1\\ d_6\≥-1\\ {-d}_6\≥-1\end{array}\right.---\left(10\right)$

? under known condition, no matter σ and η _ivalue how, above-mentioned optimization problem with linear constraints can adopt projection gradient method to solve.Make d ₁₄=σ, concrete solution procedure is as follows:

4.1: the constrained optimization problem be made up of formula (9), (10) can obtain

$\▿f\left(x\right)={\left[\begin{array}{c}0\\ 0\\ 0\\ 0\\ 0\\ 0\\ 1\end{array}\right]}^T,b={\left[\begin{array}{cccccccccccccc}0& \·\·\·& -1& -1& -1& -1& -1& -1& -1& -1& -1& -1& -1& -1\end{array}\right]}^T$

$A_{9\×(19+i)}=\left[\begin{array}{cccccccccccccccccccc}-(\frac{2}{{\hat{x}}_1^2}{x}_1^{\left(k\right)}-\frac{2}{{\hat{x}}_1})& \·\·\·& 1& -1& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0\\ -(\frac{2}{{\hat{x}}_2^2}{x}_2^{\left(k\right)}-\frac{2}{{\hat{x}}_2})& \·\·\·& 0& 0& 1& -1& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0\\ -(\frac{2}{{\hat{x}}_{11}^2}{x}_{11}^{\left(k\right)}-\frac{2}{{\hat{x}}_{11}})& \·\·\·& 0& 0& 0& 0& 1& -1& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0\\ -(\frac{2}{{\hat{x}}_{12}^2}{x}_{12}^{\left(k\right)}-\frac{2}{{\hat{x}}_{12}})& \·\·\·& 0& 0& 0& 0& 0& 0& 1& -1& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0\\ -(\frac{2}{{\hat{x}}_{21}^2}{x}_{21}^{\left(k\right)}-\frac{2}{{\hat{x}}_{21}})& \▿g_i\left(x^{\left(k\right)}\right)& 0& 0& 0& 0& 0& 0& 0& 0& 1& -1& 0& 0& 0& 0& 0& 0& 0& 0\\ -(\frac{2}{x_{22}^2}{x}_{22}^{\left(k\right)}-\frac{2}{{\hat{x}}_{22}})& \·\·\·& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 1& -1& 0& 0& 0& 0& 0& 0\\ 0& \·\·\·& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 1& -1& 0& 0& 0& 0\\ 0& \·\·\·& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 1& -1& 0& 0\\ 1& {\mathrm{\η}}_i& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 0& 1& -1\end{array}\right]$

4.2: get feasible initial point d ⁽¹⁾, make

$\begin{array}{c}{a}_i^T{d}^{\left(1\right)}-b_i=0,i=1,\·\·\·,l\\ a_i^T{d}^{\left(1\right)}-b_i\≥0,i=1,\·\·\·,m\end{array}---\left(11\right)$

Put k=1;

4.3: determine d ^(k)place's active constraint index set

$I\left(d^{\left(k\right)}\right)=\left\{i\right|a_i^T{d}^{\left(k\right)}-b_i=0,i=1,\·\·\·,n\}---\left(12\right)$

Wherein, n >=l.

4.4: if n=0, namely without equality constraint, and I (d ^(k)during)=φ, then make Q ^k=I, can calculate

$P^k={-Q}^k\▿f\left(d^{\left(k\right)}\right)---\left(13\right)$

If n=0, namely without equality constraint, but I (d ^(k)during) ≠ φ, then make

$N^k=(a_{i_1},\·\·\·,a_{i_r}),i_j\∈\{1,\·\·\·,l\}\∪I\left(d^{\left(k\right)}\right)$

Q ^k＝I-N ^k((N ^k) ^TN ^k) ^-1(N ^k) ^T(14)

$P^k={-Q}^k\▿f\left(d^{\left(k\right)}\right)---\left(15\right)$

4.5: if P ^k=0, be generally || P ^k||≤ε,

4.5.1: if n=0, i.e. I (d ^(k))=φ, then stop calculating, d ^(k)for K-T point.

4.5.2: if n ≠ 0, i.e. I (d ^(k)) ≠ φ, calculates

$q^{\left(k\right)}{\left({\left(N^k\right)}^T{N}^k\right)}^{-1}{\left(N^k\right)}^T\▿f\left(d^{\left(k\right)}\right)---\left(16\right)$

If meet then stop calculating, d ^(k)for K-T point.

4.6: if as i ∈ I (d ^(k)) time, exist $q_i^{\left(k\right)}<0,$ Order $q_k^v=\min \left\{q_i^{\left(k\right)}\right|i\&Element;I\left(d^{\left(k\right)}\right)\},$ Put N ^k=(N ^k) ', (N ^k) ' be N ^kremove corresponding row a _q;

Q ^k＝I-N ^k((N ^k) ^TN ^k) ^-1(N ^k) ^T(17)

$P^k=-Q^k\&dtri;f\left(d^{\left(k\right)}\right)---\left(18\right)$

4.7: if P ^k≠ 0, solve one-dimensional problem

min f(λ _k)＝f(x ^k+λ _kP ^k) (19)

s.t. 0≤λ _k≤λ _max(20)

Wherein,

${\mathrm{\&lambda;}}_{\max }=\left\{\begin{array}{c}\min \{-\frac{a_i^T{d}^{\left(k\right)}-b_i}{a_i^T{d}^{\left(k\right)}}|a_i^T{d}^{\left(k\right)}<0,i\&NotElement;I\left(d^{\left(k\right)}\right)\}\\ +\&infin;,\{a_i^T{d}^{\left(k\right)}<0,i\&NotElement;I\left(d^{\left(k\right)}\right)\}=\mathrm{\&phi;}\end{array}\right\}---\left(21\right)$

Solve and can obtain λ _k, put k:=k+1, d ^k+1=d ^k+ λ _kp ^k, proceed to 4.2.

Step5: with d ^(k)for linear search is carried out in direction, first right ask t _i> 0, makes g _i(x ^(k)+ t _id ^(k))=0 is (if t _imore than one of > 0, then get its minimum value, namely

${\mathrm{\&alpha;}}_{\max }=\min \left\{t_i\right|g_i(x^{\left(k\right)}+t_i{d}^{\left(k\right)})=0,t_i>0,i\&NotElement;I(g_i\left(x^{\left(k\right)}\right),{\mathrm{\&epsiv;}}_k)\},$

Solve simultaneously

$\left\{\begin{array}{c}\begin{array}{cc}\min & f(x^{\left(k\right)}+{\mathrm{\&alpha;}}_k{d}^{\left(k\right)})\end{array}\\ 0\&le;{\mathrm{\&alpha;}}_k\&le;{\mathrm{\&alpha;}}_{\max }\end{array}\right.$

Solution can obtain α _k; Then x ^(k+1)=x ^(k)+ α _kd ^(k);

5.1: if α ^(k)∈ (-∞ ,+∞), solves about objective function extreme point f (x ^(k)+ α _kd ^(k)) corresponding to α ^(k)dependent variable, namely

$\begin{array}{c}f(x^{\left(k\right)}+{\mathrm{\&alpha;}}_k{d}^{\left(k\right)})=f\left({\mathrm{\&alpha;}}^{\left(k\right)}\right)\\ ={\left(\frac{d_1^{\left(k\right)}{\mathrm{\&alpha;}}^{\left(k\right)}+x_1-{\hat{x}}_1}{{\hat{x}}_1}\right)}^2+{\left(\frac{d_2^{\left(k\right)}{\mathrm{\&alpha;}}^{\left(k\right)}+x_2-{\hat{x}}_2}{{\hat{x}}_2}\right)}^2+{\left(\frac{d_3^{\left(k\right)}{\mathrm{\&alpha;}}^{\left(k\right)}+x_{11}-{\hat{x}}_{11}}{{\hat{x}}_{11}}\right)}^2\\ +{\left(\frac{d_4^{\left(k\right)}{\mathrm{\&alpha;}}^{\left(k\right)}+x_{12}-{\hat{x}}_{12}}{{\hat{x}}_{12}}\right)}^2+{\left(\frac{d_5^{\left(k\right)}{\mathrm{\&alpha;}}^{\left(k\right)}+x_{21}-{\hat{x}}_{21}}{{\hat{x}}_{21}}\right)}^2+{\left(\frac{d_6^{\left(k\right)}{\mathrm{\&alpha;}}^{\left(k\right)}+x_{22}-{\hat{x}}_{221}}{{\hat{x}}_{22}}\right)}^2\\ =({\left(\frac{d_1^{\left(k\right)}}{{\hat{x}}_1}\right)}^2+{\left(\frac{d_2^{\left(k\right)}}{{\hat{x}}_2}\right)}^2+{\left(\frac{d_3^{\left(k\right)}}{{\hat{x}}_{11}}\right)}^2+{\left(\frac{d_4^{\left(k\right)}}{{\hat{x}}_{12}}\right)}^2+{\left(\frac{d_5^{\left(k\right)}}{{\hat{x}}_{21}}\right)}^2+{\left(\frac{d_6^{\left(k\right)}}{{\hat{x}}_{22}}\right)}^2){\left({\mathrm{\&alpha;}}^{\left(k\right)}\right)}^2\end{array}$ When

$\begin{array}{c}+2{(\frac{d_1^{\left(k\right)}(x_1-{\hat{x}}_1)}{{\hat{x}}_1}+\frac{d_2^{\left(k\right)}(x_2-{\hat{x}}_2)}{{\hat{x}}_2}+\frac{d_3^{\left(k\right)}(x_{11}-{\hat{x}}_{11})}{{\hat{x}}_{11}}+\frac{d_4^{\left(k\right)}(x_{12}-{\hat{x}}_{12})}{{\hat{x}}_{12}}+\frac{d_5^{\left(k\right)}(x_{21}-{\hat{x}}_{21})}{{\hat{x}}_{21}}+\frac{d_6^{\left(k\right)}(x_{22}-{\hat{x}}_{22})}{{\hat{x}}_{22}})}^2\\ +{\left(\frac{x_1-{\hat{x}}_1}{{\hat{x}}_1}\right)}^2+{\left(\frac{x_2-{\hat{x}}_2}{{\hat{x}}_2}\right)}^2+{\left(\frac{x_{11}-{\hat{x}}_{11}}{{\hat{x}}_{11}}\right)}^2+{\left(\frac{x_{12}-{\hat{x}}_{12}}{{\hat{x}}_{12}}\right)}^2+{\left(\frac{x_{21}-{\hat{x}}_{21}}{{\hat{x}}_{21}}\right)}^2+{\left(\frac{x_{22}-{\hat{x}}_{22}}{{\hat{x}}_{22}}\right)}^2\end{array}$

${\mathrm{\&alpha;}}^{\left(k\right)}=-\frac{\frac{d_1^{\left(k\right)}(x_1-{\hat{x}}_1)}{{\hat{x}}_1}+\frac{d_2^{\left(k\right)}(x_2-{\hat{x}}_2)}{{\hat{x}}_2}+\frac{d_4^{\left(k\right)}(x_{11}-{\hat{x}}_{11})}{{\hat{x}}_{11}}+\frac{d_4^{\left(k\right)}(x_{12}-{\hat{x}}_{12})}{{\hat{x}}_{12}}+\frac{d_5^{\left(k\right)}(x_{21}-{\hat{x}}_{21})}{{\hat{x}}_{21}}+\frac{d_6^{\left(k\right)}(x_{22}-{\hat{x}}_{22})}{{\hat{x}}_{22}}}{{\left(\frac{d_1^{\left(k\right)}}{{\hat{x}}_1}\right)}^2+{\left(\frac{d_2^{\left(k\right)}}{{\hat{x}}_2}\right)}^2+{\left(\frac{d_3^{\left(k\right)}}{{\hat{x}}_{11}}\right)}^2+{\left(\frac{d_4^{\left(k\right)}}{{\hat{x}}_{12}}\right)}^2+{\left(\frac{d_5^{\left(k\right)}}{{\hat{x}}_{21}}\right)}^2+{\left(\frac{d_6^{\left(k\right)}}{{\hat{x}}_{22}}\right)}^2}$ Time, f (α ^(k)) minimum.

5.2: for α ^(k)∈ [0, α _max] situation

5.2.1: if

${\mathrm{\&alpha;}}^{\left(k\right)}=-\frac{\frac{d_1^{\left(k\right)}(x_1-{\hat{x}}_1)}{{\hat{x}}_1}+\frac{d_2^{\left(k\right)}(x_2-{\hat{x}}_2)}{{\hat{x}}_2}+\frac{d_4^{\left(k\right)}(x_{11}-{\hat{x}}_{11})}{{\hat{x}}_{11}}+\frac{d_4^{\left(k\right)}(x_{12}-{\hat{x}}_{12})}{{\hat{x}}_{12}}+\frac{d_5^{\left(k\right)}(x_{21}-{\hat{x}}_{21})}{{\hat{x}}_{21}}+\frac{d_6^{\left(k\right)}(x_{22}-{\hat{x}}_{22})}{{\hat{x}}_{22}}}{{\left(\frac{d_1^{\left(k\right)}}{{\hat{x}}_1}\right)}^2+{\left(\frac{d_2^{\left(k\right)}}{{\hat{x}}_2}\right)}^2+{\left(\frac{d_3^{\left(k\right)}}{{\hat{x}}_{11}}\right)}^2+{\left(\frac{d_4^{\left(k\right)}}{{\hat{x}}_{12}}\right)}^2+{\left(\frac{d_5^{\left(k\right)}}{{\hat{x}}_{21}}\right)}^2+{\left(\frac{d_6^{\left(k\right)}}{{\hat{x}}_{22}}\right)}^2}<0,$

Then α ^(k)when=0, f (α ^(k)) minimum, now x ^(k+1)=x ^(k);

5.2.2: if

$0\&le;-\frac{\frac{d_1^{\left(k\right)}(x_1-{\hat{x}}_1)}{{\hat{x}}_1}+\frac{d_2^{\left(k\right)}(x_2-{\hat{x}}_2)}{{\hat{x}}_2}+\frac{d_4^{\left(k\right)}(x_{11}-{\hat{x}}_{11})}{{\hat{x}}_{11}}+\frac{d_4^{\left(k\right)}(x_{12}-{\hat{x}}_{12})}{{\hat{x}}_{12}}+\frac{d_5^{\left(k\right)}(x_{21}-{\hat{x}}_{21})}{{\hat{x}}_{21}}+\frac{d_6^{\left(k\right)}(x_{22}-{\hat{x}}_{22})}{{\hat{x}}_{22}}}{{\left(\frac{d_1^{\left(k\right)}}{{\hat{x}}_1}\right)}^2+{\left(\frac{d_2^{\left(k\right)}}{{\hat{x}}_2}\right)}^2+{\left(\frac{d_3^{\left(k\right)}}{{\hat{x}}_{11}}\right)}^2+{\left(\frac{d_4^{\left(k\right)}}{{\hat{x}}_{12}}\right)}^2+{\left(\frac{d_5^{\left(k\right)}}{{\hat{x}}_{21}}\right)}^2+{\left(\frac{d_6^{\left(k\right)}}{{\hat{x}}_{22}}\right)}^2}\&le;{\mathrm{\&alpha;}}_{\max },$

If then

${\mathrm{\&alpha;}}^{\left(k\right)}=-\frac{\frac{d_1^{\left(k\right)}(x_1-{\hat{x}}_1)}{{\hat{x}}_1}+\frac{d_2^{\left(k\right)}(x_2-{\hat{x}}_2)}{{\hat{x}}_2}+\frac{d_4^{\left(k\right)}(x_{11}-{\hat{x}}_{11})}{{\hat{x}}_{11}}+\frac{d_4^{\left(k\right)}(x_{12}-{\hat{x}}_{12})}{{\hat{x}}_{12}}+\frac{d_5^{\left(k\right)}(x_{21}-{\hat{x}}_{21})}{{\hat{x}}_{21}}+\frac{d_6^{\left(k\right)}(x_{22}-{\hat{x}}_{22})}{{\hat{x}}_{22}}}{{\left(\frac{d_1^{\left(k\right)}}{{\hat{x}}_1}\right)}^2+{\left(\frac{d_2^{\left(k\right)}}{{\hat{x}}_2}\right)}^2+{\left(\frac{d_3^{\left(k\right)}}{{\hat{x}}_{11}}\right)}^2+{\left(\frac{d_4^{\left(k\right)}}{{\hat{x}}_{12}}\right)}^2+{\left(\frac{d_5^{\left(k\right)}}{{\hat{x}}_{21}}\right)}^2+{\left(\frac{d_6^{\left(k\right)}}{{\hat{x}}_{22}}\right)}^2}$ Time, f (α ^(k)) minimum, now x ^(k+1)=x ^(k)+ α ^(k)d ^(k);

5.2.3: if

$-\frac{\frac{d_1^{\left(k\right)}(x_1-{\hat{x}}_1)}{{\hat{x}}_1}+\frac{d_2^{\left(k\right)}(x_2-{\hat{x}}_2)}{{\hat{x}}_2}+\frac{d_4^{\left(k\right)}(x_{11}-{\hat{x}}_{11})}{{\hat{x}}_{11}}+\frac{d_4^{\left(k\right)}(x_{12}-{\hat{x}}_{12})}{{\hat{x}}_{12}}+\frac{d_5^{\left(k\right)}(x_{21}-{\hat{x}}_{21})}{{\hat{x}}_{21}}+\frac{d_6^{\left(k\right)}(x_{22}-{\hat{x}}_{22})}{{\hat{x}}_{22}}}{{\left(\frac{d_1^{\left(k\right)}}{{\hat{x}}_1}\right)}^2+{\left(\frac{d_2^{\left(k\right)}}{{\hat{x}}_2}\right)}^2+{\left(\frac{d_3^{\left(k\right)}}{{\hat{x}}_{11}}\right)}^2+{\left(\frac{d_4^{\left(k\right)}}{{\hat{x}}_{12}}\right)}^2+{\left(\frac{d_5^{\left(k\right)}}{{\hat{x}}_{21}}\right)}^2+{\left(\frac{d_6^{\left(k\right)}}{{\hat{x}}_{22}}\right)}^2}>{\mathrm{\&alpha;}}_{\max },$

Then α ^(k)=α _maxtime, f (α ^(k)) minimum, now x ^(k+1)=x ^(k)+ α _maxd ^(k);

Wherein, for the downward gradient of objective function is at the component of different directions, namely raw ore lead revises grade, raw ore zinc correction grade, concentrate lead revise grade, concentrate zinc correction grade, mine tailing lead revise grade, mine tailing zinc correction grade, concentrate calculates productive rate, mine tailing calculates each component of productive rate downward gradient component; α ^(k)for the optimum stepsize of optimizing iterative computation.

Step6: if || x ^(k+1)-x ^(k)||≤ε, then x ^(k+1)for approximate optimal solution, iteration ends; Otherwise, order ${\mathrm{\&epsiv;}}_k=\left\{\begin{array}{c}{\mathrm{\&epsiv;}}_k,{\mathrm{\&epsiv;}}_k\&le;-{\mathrm{\&sigma;}}_k\\ \frac{{\mathrm{\&epsiv;}}_k}{2},{\mathrm{\&epsiv;}}_k>{-\mathrm{\&sigma;}}_k\end{array}\right.,k=k+1,$ Turn Step2.

Described productive rate is optimized in computation process, because the domain of variation of the feasible zone and metal content assay that calculate metal productive rate has direct relation.For the true relation that metal content and optimization after accurate description assay metal content, adjustment calculate between metal productive rate, according to the inherent feature of process production technique and the experimental knowledge of operation slip-stick artist, set up metal productive rate and optimize computation model rule base.Model rule-based knowledge base mainly comprises metal productive rate and product metal content constraint rule, equilibrium condition constraint rule, external constraint rule and knowledge base update are regular.

Described metal balance calculates in During Process of Long-term Operation, there is situation inconsistent in analytical test data Sum fanction knowledge base unavoidably, be consistent with actual for ensureing to calculate concentrate amount, metal productive rate in metal productive rate and mineral processing production form, the computing information according to grade regulation rule, metal productive rate adjusts rule base.

Production report is in conjunction with production data information (treatment capacity, water of productive use amount etc. come from the data of supervisory system) according to the metal productive rate coordinating to calculate, the ore dressing overall targets such as the process high concentrate amount of the class of extrapolating, day, the moon, low concentrate amount, metal productive rate, for flow operations and production management provide decision support.

(1) grade rule is adjusted

After the adjustment correction of each measuring point of flow process, metal content and analysis grade should drop in certain domain of variation, and this domain of variation is defined by preparation engineer and provides.The metal content of general assay also should fall in this domain of variation, and when the metal content of assay is positioned at outside domain of variation, then amendment adjustment Grade change territory, makes it to meet above-mentioned condition.

(2) equilibrium condition rule

To in the balance set up according to production procedure, its raw ore metal content should lower than the metal content of concentrate after enrichment, and raw ore metal content should higher than the metal content of underflow simultaneously.Reverse situation because sampling, sample preparation and analytical test cause often can occur in production, and namely head grade is lower than underflow grade or the head grade grade higher than floating concentrate.For ensureing the calculating accuracy of software, if violate one of them condition, then according to the anti-grade pushing away another product of material balance of setting; If the situation that two conditions are all violated is extremely rare, then directly carry out exchanging rear calculating, and provide to calculate and report to the police, the flow process balance point position residing for record.

(3) priori productive rate rule

If solve the stage productive rate drawn drop on outside corresponding priori productive rate domain of variation according to global optimization, then direct priori productive rate rule to be modified, and provide and extremely calculate warning.

(4) external condition rule

If in production run, according to the results of calculation obtained between actual raw ore output and concentrate output, metal balance computing system judges whether the constraint rule violating outside productive rate, is directly modified by the productive rate of correspondence.

The algorithm of a kind of many metal balances productive rate of the present invention, its beneficial effect is: 1. the present invention adopts non-linear constrain derivation algorithm, and the correction mechanism of sampling, analytical test error is introduced computation model, and calculation of yield precision is high; 2., in this quick calculation method, productive rate and grade corrected parameter carry out iterative computation simultaneously, and calculated amount is little, and speed is fast; 3. highly versatile of the present invention, be applicable to the material calculation of yield of various flow process, be skillfully constructed, accurate and practical, cost performance is high, is conducive to large-scale promotion application.

Accompanying drawing explanation

In order to be illustrated more clearly in the embodiment of the present invention or technical scheme of the prior art, below the accompanying drawing used required in describing embodiment is briefly described, apparently, accompanying drawing in the following describes is only one of them embodiment of the present invention, for those of ordinary skill in the art, under the prerequisite not paying creative work, other accompanying drawing can also be obtained according to these accompanying drawings.

Accompanying drawing 1 is basic hardware structure of the present invention;

1-model calculates host computer, the 2-network switch, 3-in-line analyzer, 4-off-line data typing terminal, 5-production run Control Server, 6-communication line;

Accompanying drawing 2 is algorithm model functional structure of the present invention;

Accompanying drawing 3 is embodiment of the present invention flow process measuring point distribution schematic diagram;

S1-mono-section of raw ore, S2-bis-sections of raw ores, S3-mono-section of concentrate, S4-bis-sections of concentrate, S5-height concentrate, the low concentrate of S6-, the total mine tailing of S7-;

Accompanying drawing 4 is embodiment of the present invention algorithm model calculation flow chart.

Embodiment

Below in conjunction with the accompanying drawing in the embodiment of the present invention, be clearly and completely described the technical scheme in the embodiment of the present invention, obviously, described embodiment is only one of them embodiment of the present invention, instead of whole embodiments.Based on the embodiment in the present invention, those of ordinary skill in the art, not making the every other embodiment obtained under creative work prerequisite, belong to the scope of protection of the invention.

Embodiment 1

The quick calculation method of a kind of many metal balances productive rate of the present invention, its model algorithm process is as follows: according to metal balance and productive rate balance, can obtain following relational expression:

One essence is+two former=and mono-former

Metal balance: $\begin{array}{c}{\hat{x}}_{21}{\mathrm{\&gamma;}}_2+{\hat{x}}_{11}{\mathrm{\&gamma;}}_1={\hat{x}}_1\\ {\hat{x}}_{22}{\mathrm{\&gamma;}}_2+{\hat{x}}_{12}{\mathrm{\&gamma;}}_1={\hat{x}}_2\end{array}$

Productive rate balances: γ ₂+ γ ₁=1

One essence+two essence=high-precision

Metal balance: $\begin{array}{c}{\hat{x}}_{21}{\mathrm{\&gamma;}}_2+{\hat{x}}_{33}{\mathrm{\&gamma;}}_3={\hat{x}}_{43}{\mathrm{\&gamma;}}_4\\ {\hat{x}}_{22}{\mathrm{\&gamma;}}_2+{\hat{x}}_{34}{\mathrm{\&gamma;}}_3={\hat{x}}_{44}{\mathrm{\&gamma;}}_4\end{array}$

Productive rate balances: γ ₂+ γ ₃=γ ₄

Two essence+low essences+total tail=bis-are former

Metal balance: $\begin{array}{c}{\hat{x}}_{33}{\mathrm{\&gamma;}}_3+{\hat{x}}_{55}{\mathrm{\&gamma;}}_5+{\hat{x}}_{66}{\mathrm{\&gamma;}}_6={\hat{x}}_{11}{\mathrm{\&gamma;}}_1\\ {\hat{x}}_{34}{\mathrm{\&gamma;}}_3+{\hat{x}}_{56}{\mathrm{\&gamma;}}_5+{\hat{x}}_{65}{\mathrm{\&gamma;}}_6={\hat{x}}_2{\mathrm{\&gamma;}}_1\end{array}$

Productive rate balances: γ ₃+ γ ₅+ γ ₆=γ ₁

According to theoretical metal balance and the productive rate EQUILIBRIUM CALCULATION FOR PROCESS formula of flow process measuring point, can draw the following conclusions:

(1) according to the plumbous grade x of one section of raw ore ₁, one section of raw ore zinc grade x ₂, two sections of plumbous grade x of raw ore ₁₁, two sections of raw ore zinc grade x ₁₂, one section of plumbous grade x of concentrate ₂₁with one section of concentrate zinc grade x ₂₂, two sections of raw ore γ can be calculated ₁with the productive rate γ of one section of concentrate ₂, one section of raw ore lead adjustment grade one section of raw ore zinc adjustment grade two sections of raw ore lead adjustment grades two sections of raw ore zinc adjustment grades one section of concentrate lead adjustment grade with one section of concentrate zinc adjustment grade

(2) according to one section of concentrate lead adjustment grade with one section of concentrate zinc adjustment grade and the productive rate γ of one section of concentrate ₂, two sections of plumbous grade x of concentrate ₃₃, two sections of concentrate zinc grade x ₃₄, the plumbous grade x of high concentrate ₄₃, high concentrate zinc grade x ₄₄, two sections of concentrate γ can be calculated ₃with the productive rate γ of high concentrate ₂, two sections of concentrate lead adjustment grades two sections of concentrate adjustment zinc grades high concentrate lead adjustment grade high concentrate zinc adjustment grade

(3) according to two sections of raw ore lead adjustment grades two sections of raw ore zinc adjustment grades and two sections of raw ore productive rate γ ₁, two sections of concentrate lead adjustment grades two sections of concentrate adjustment zinc grades and two sections of concentrate γ ₃, the plumbous grade x of low concentrate ₅₅, low concentrate zinc grade x ₅₆, the plumbous grade x of total mine tailing ₆₆, total mine tailing zinc grade x ₆₅, low concentrate γ can be calculated ₅with the productive rate γ of total mine tailing ₆, low concentrate lead adjustment grade low concentrate zinc grade the plumbous grade of total mine tailing total mine tailing zinc grade

As having 7 flow process measuring points in accompanying drawing 3, needing altogether to solve 6 metal productive rates, revising the metal content of adjustment 7 measuring points.The characteristic of foundation production technology and on-the-spot real medium, raw 13 domain of variation constraint of this Solution of Nonlinear Optimal Problem common property, obtains 6 equality constraints altogether according to material balance in floatation process and metal balance.

If carry out correction adjustment and calculation of yield to the measuring point grade of flow process simultaneously, it is obviously comparatively large that it optimizes scale, and algorithm computation complexity increases.Find out that this optimization problem can be decomposed into A according to material balance and the equality constraint of metal balance and the domain of variation condition of feasible solution, B, the sub-optimization problem that C tri-is different, calculate A respectively, B, unknown productive rate and correction metal content in C model, and the rolling optimization calculating of next stage is participated in as known terms.

$\begin{array}{c}\min {\left(\frac{x_1-{\hat{x}}_1}{{\hat{x}}_1}\right)}^2+{\left(\frac{x_2-{\hat{x}}_2}{{\hat{x}}_2}\right)}^2+{\left(\frac{x_{11}+{\hat{x}}_{11}}{{\hat{x}}_{11}}\right)}^2+{\left(\frac{x_{12}-{\hat{x}}_{12}}{{\hat{x}}_{12}}\right)}^2\\ +{\left(\frac{x_{21}-{\hat{x}}_{21}}{{\hat{x}}_{21}}\right)}^2+{\left(\frac{x_{22}-{\hat{x}}_{22}}{{\hat{x}}_{22}}\right)}^2\end{array}$

In formula, γ ₁two original rates;

γ ₂it is a smart productive rate;

X ₁₁it is the plumbous grade of two former middle reality;

two former plumbous grades after adjustment;

X ₁₂two former middle actual zinc grades;

two former zinc grades after adjustment;

X ₂₁it is actual plumbous grade in an essence;

it is the lead bullion grade after adjustment;

X ₂₂it is actual zinc grade in an essence;

it is the smart zinc grade after adjustment;

X ₁it is the plumbous grade of a former middle reality;

it is the former plumbous grade after adjustment;

X ₂it is a former middle actual zinc grade;

it is the former zinc grade after adjustment;

C ₁, C ₂, C ₃, C ₄, C ₅, C ₆be that mineral processing production flow process one is former, two former, limit D-value between plumbous in a concentrate, zinc grade;

A ₁, B ₁, A ₂, B ₂be respectively the domain of variation bound of two former and smart productive rates;

A ₁, a ₂be respectively the lower limit set value of a former lead, zinc adjustment grade;

B ₁, b ₂be respectively the upper limit set value of a former lead, zinc adjustment grade;

A ₁₁, a ₂₁be respectively the lower limit set value of two former lead, zinc adjustment grade;

B ₁₁, b ₂₁be respectively the upper limit set value of two former lead, zinc adjustment grade;

A ₁₂, a ₂₂be respectively the lower limit set value of a lead bullion, zinc adjustment grade;

B ₁₂, b ₂₂be respectively the upper limit set value of a lead bullion, zinc adjustment grade.

$\begin{array}{c}\min {\left(\frac{x_{21}-{\hat{x}}_{21}}{{\hat{x}}_{21}}\right)}^2+{\left(\frac{x_{22}-{\hat{x}}_{22}}{{\hat{x}}_{22}}\right)}^2+{\left(\frac{x_{33}+{\hat{x}}_{33}}{{\hat{x}}_{33}}\right)}^2+{\left(\frac{x_{34}-{\hat{x}}_{34}}{{\hat{x}}_{34}}\right)}^2\\ +{\left(\frac{x_{43}-{\hat{x}}_{43}}{{\hat{x}}_{43}}\right)}^2+{\left(\frac{x_{44}-{\hat{x}}_{44}}{{\hat{x}}_{44}}\right)}^2\end{array}$

In formula, solve according to A model algorithm the known terms drawn;

γ ₂it is the known smart productive rate obtained by A algorithm model;

γ ₃two smart productive rates;

γ ₄it is high-precision productive rate;

X ₃₃it is actual plumbous grade in two concentrate;

it is the plumbous grade of two concentrate after adjustment;

X ₃₄it is actual zinc grade in two concentrate;

two concentrate zinc grades after adjustment;

X ₄₃it is actual plumbous grade in high concentrate;

it is the plumbous grade of high concentrate after adjustment;

X ₄₄it is actual zinc grade in high concentrate;

it is the high concentrate zinc grade after adjustment;

X ₂₁it is actual plumbous grade in a concentrate;

it is the plumbous grade of known adjustment obtained by A algorithm model;

X ₂₂it is actual zinc grade in a concentrate;

it is the known adjustment zinc grade obtained by A algorithm model;

C ₆, C ₂₁, C ₃₁, C ₄₁, C ₅₁, C ₆₁it is the limit D-value in mineral processing production flow process one essence, two essences, high concentrate between lead, zinc grade;

A ₃, B ₃, A ₄, B ₄be respectively the domain of variation bound of two essences and high-precision productive rate;

A ₁₂, a ₂₂be respectively the lower limit set value that a concentrate is plumbous, zinc adjusts grade;

B ₁₂, b ₂₂be respectively the upper limit set value that a concentrate is plumbous, zinc adjusts grade;

A ₃₃, a ₄₃be respectively the lower limit set value that two concentrate are plumbous, zinc adjusts grade;

B ₃₃, b ₄₃be respectively the upper limit set value that two concentrate are plumbous, zinc adjusts grade;

A ₃₄, a ₄₄be respectively the lower limit set value that high concentrate is plumbous, zinc adjusts grade;

B ₃₄, b ₄₄be respectively the upper limit set value that high concentrate is plumbous, zinc adjusts grade.

$\begin{array}{c}\min {\left(\frac{x_{11}-{\hat{x}}_{11}}{{\hat{x}}_{11}}\right)}^2+{\left(\frac{x_{12}-{\hat{x}}_{12}}{{\hat{x}}_{12}}\right)}^2+{\left(\frac{x_{33}+{\hat{x}}_{33}}{{\hat{x}}_{33}}\right)}^2+{\left(\frac{x_{34}-{\hat{x}}_{34}}{{\hat{x}}_{34}}\right)}^2\\ +{\left(\frac{x_{55}-{\hat{x}}_{55}}{{\hat{x}}_{55}}\right)}^2+{\left(\frac{x_{56}-{\hat{x}}_{56}}{{\hat{x}}_{56}}\right)}^2+{\left(\frac{x_{66}-{\hat{x}}_{66}}{{\hat{x}}_{66}}\right)}^2+{\left(\frac{x_{65}-{\hat{x}}_{65}}{{\hat{x}}_{65}}\right)}^2\end{array}$

In formula, ${\left(\frac{x_{11}-{\hat{x}}_{11}}{{\hat{x}}_{11}}\right)}^2+{\left(\frac{x_{12}-{\hat{x}}_{12}}{{\hat{x}}_{12}}\right)}^2+{\left(\frac{x_{33}-{\hat{x}}_{33}}{{\hat{x}}_{33}}\right)}^2+{\left(\frac{x_{34}-{\hat{x}}_{34}}{{\hat{x}}_{34}}\right)}^2$ Solve according to model algorithm A and B the known terms drawn;

γ ₁the known two original rates obtained by A algorithm model;

γ ₃the known two smart productive rates obtained by B algorithm model;

γ ₅it is high-precision productive rate;

γ ₆it is total tail productive rate;

X ₅₅it is actual plumbous grade in low concentrate;

it is the plumbous grade of low concentrate after adjustment;

X ₅₆it is actual zinc grade in low concentrate;

it is the low concentrate zinc grade after adjustment;

X ₆₆it is actual plumbous grade in total mine tailing;

it is the plumbous grade of total mine tailing after adjustment;

X ₆₅it is actual zinc grade in total mine tailing;

it is the total mine tailing zinc grade after adjustment;

X ₁₁it is actual plumbous grade in a raw ore;

it is the plumbous grade of a known former adjustment obtained by A algorithm model;

X ₁₂it is actual zinc grade in a raw ore;

it is the known former adjustment zinc grade obtained by A algorithm model;

X ₃₃it is actual plumbous grade in two concentrate;

that known two concentrate obtained by B algorithm model adjust plumbous grade;

X ₃₄it is actual zinc grade in two concentrate;

that known two concentrate obtained by B algorithm model adjust zinc grade;

C ₅₃, C ₅₂, C ₅₄, C ₅₆, C ₆₅, C ₆₆be that mineral processing production flow process two is former, two essences, low essence, limit D-value between plumbous in total mine tailing, zinc grade;

A ₅, B ₅, A ₆, B ₆be respectively the domain of variation bound of low essence and total tail productive rate;

A ₅₁, a ₅₂be respectively the lower limit set value that low concentrate is plumbous, zinc adjusts grade;

B ₅₁, b ₅₂be respectively the upper limit set value that low concentrate is plumbous, zinc adjusts grade;

A ₅₃, a ₅₄be respectively the lower limit set value that total mine tailing is plumbous, zinc adjusts grade;

B ₅₃, b ₅₄be respectively the upper limit set value that total mine tailing is plumbous, zinc adjusts grade.

The quick calculation method of a kind of many metals productive rate of the present invention, operative installations as shown in Figure 1, comprises 1 model and calculates host computer, 2 network switchs, 3 in-line analyzers, 4 off-line data typing terminals, 5 production run Control Servers and 6 communication lines.It is SQLServer2005 that model calculates stored data base in host computer, and development environment is VS2008.A in embodiment, B, C tri-sections of rolling optimization computation model input/output arguments are stored in SQL Server2005, model algorithm, remote terminal typing interface and production process data OPC Client communication interface are developed under VS2005, and be deployed in model and calculate in host computer 1, the print preview function of the trend display of support model parameter modification, result of calculation, inquiry and supplement production form; The network switch 2, in-line analyzer 3, off-line data typing terminal 4, production run Control Server 5 and model calculate communication link and the data interaction function of host computer 1.In in-line analyzer 3 pairs of accompanying drawing 3 beneficiation flowsheets, the metal content of measuring point S1 ~ S7 carries out on-line analysis, and analysis result is sent to model calculating host computer 1 in accompanying drawing 1; Off-line data typing terminal 4 is responsible for processing Analytical Laboratory Results, and is uploaded to model calculating host computer 1.Production run Control Server 5 procedure parameters such as treatment capacity in production run is uploaded in real time model to calculate host computer 1_.Concrete computation process is as follows:

1. input former, two former, an essence, two smart, high-precision, low essences, always tailings grade and the borders parameter to product yield in the rolling optimization models calculating of three stages.

2. computation model starts in-line analyzer 3 or off-line data typing terminal 4 in interface routine connection accompanying drawing 1, reads current measuring point and analyzes data.

3. check former, two former, accurate measurement point acquisition data message integralities, start A stage model and calculate.

3.1 according to step1 computing method, and solve the initial feasible solution of A stage model, namely iteration starting point, proceeds to step2;

3.2 according to step2, calculates the constraint set matrix of A stage model with the downward gradient of objective function proceed to step3;

3.3 stopping criterion for iteration set according to step3, check whether and satisfy condition.Stop iteration if meet, draw optimum solution; Otherwise, proceed to step4 and calculate;

3.4, according to step4 linear programming, solve optimizing factor lambda _kwith direction of search d ^(k);

3.5 according to α ^(k), direction of search d ^(k), solve approximate optimal solution x ^(k+1);

3.6 check approximate optimal solution x ^(k+1)if, || x ^(k+1)-x ^(k)|| < ε _k, stop; Otherwise, proceed to step2.Result of calculation after iteration ends is adjustment metal contents that are former, two former and essences, the calculating productive rate of two former and essences;

4., according to the whole rear metal content of an accurate adjustment after A stage algorithm model optimizing and calculating productive rate, simultaneous two essence and the actual grade of high-precision measuring point, carry out the calculating of B-stage algorithm model.Computation process as indicated at 3.Result of calculation after iteration ends is two essences and high-precision adjustment metal content, calculates productive rate.

5. according to metal content after two former adjustment after A stage algorithm model optimizing and calculating productive rate, the whole rear metal content of two accurate adjustments after B-stage algorithm model optimizing and calculating productive rate, the low essence of simultaneous and the actual grade of total tail measuring point, carry out C stage algorithm model and calculate.Computation process is as shown in 3 and 4.Result of calculation after iteration ends is adjustment metal content, the calculating productive rate of low essence and total tail.

The foregoing is only preferred embodiment of the present invention, not in order to limit the present invention, within the spirit and principles in the present invention all, any amendment done, equivalent replacement, improvement etc., all should be included within protection scope of the present invention.

Claims (3)

1. an algorithm for the productive rate of metal balance more than, is characterized in that: comprise the following steps:

1) equation γ is listed according to the total input quality of actual mining processing industry flow process is equal with total output quality ₁+ γ ₂=1

${\hat{x}}_{11}{\mathrm{\&gamma;}}_1+{\hat{x}}_{12}{\mathrm{\&gamma;}}_2={\hat{x}}_1$

${\hat{x}}_{21}{\mathrm{\&gamma;}}_1+{\hat{x}}_{22}{\mathrm{\&gamma;}}_2={\hat{x}}_2;$

In formula, γ ₁it is concentrate yield;

γ ₂it is mine tailing productive rate;

it is plumbous adjustment grade in concentrate;

it is plumbous adjustment grade in mine tailing;

it is zinc adjustment grade in concentrate;

it is zinc adjustment grade in mine tailing;

it is plumbous adjustment grade in raw ore;

it is zinc adjustment grade in raw ore;

2) according to priori as constraint condition, by equations turned be optimization problem:

$\begin{array}{cc}\min & {\left(\frac{x_1-{\hat{x}}_1}{{\hat{x}}_1}\right)}^2+{\left(\frac{x_2-{\hat{x}}_2}{{\hat{x}}_2}\right)}^2+{\left(\frac{x_{11}-{\hat{x}}_{11}}{{\hat{x}}_{11}}\right)}^2+{\left(\frac{x_{12}-{\hat{x}}_{12}}{{\hat{x}}_{12}}\right)}^2\\ & +{\left(\frac{x_{21}-{\hat{x}}_{21}}{{\hat{x}}_{21}}\right)}^2+{\left(\frac{x_{22}-{\hat{x}}_{22}}{{\hat{x}}_{22}}\right)}^2\end{array}$

$s.t.\left\{\begin{array}{c}{\mathrm{\&gamma;}}_1+{\mathrm{\&gamma;}}_2-1\&GreaterEqual;0\\ -{\mathrm{\&gamma;}}_1-{\mathrm{\&gamma;}}_2+1\&GreaterEqual;0\\ x_{11}-x_1-C_1>0\\ x_{21}-x_2-C_2>0\\ x_{12}-x_1+C_3>0\\ x_{22}-x_2+C_4>0\\ x_{11}-x_{12}-C_5>0\\ x_{21}-x_{22}-C_6>0\\ x_1-a_1\&GreaterEqual;0\\ -x_1+b_1\&GreaterEqual;0\\ x_2-a_2\&GreaterEqual;0\\ {-x}_2+b_2\&GreaterEqual;0\\ x_{11}-a_{11}\&GreaterEqual;0\\ {-x}_{11}+b_{11}\&GreaterEqual;0\\ x_{12}-a_{12}\&GreaterEqual;0\\ {-x}_{12}+b_{12}\&GreaterEqual;0\\ x_{21}-a_{21}\&GreaterEqual;0\\ {-x}_{21}+b_{21}\&GreaterEqual;0\\ x_{22}-a_{22}\&GreaterEqual;0\\ {-x}_{22}+b_{22}\&GreaterEqual;0\\ {\mathrm{\&gamma;}}_1-A_1\&GreaterEqual;0\\ {-\mathrm{\&gamma;}}_1+B_1\&GreaterEqual;0\\ {\mathrm{\&gamma;}}_2-A_2\&GreaterEqual;0\\ -{\mathrm{\&gamma;}}_2-B_2\&GreaterEqual;0\end{array}\right.;$

In formula, γ ₁it is concentrate yield;

γ ₂it is mine tailing productive rate;

X ₁₁it is actual plumbous grade in concentrate;

it is the plumbous grade of concentrate after adjustment;

X ₁₂it is actual plumbous grade in mine tailing;

it is the plumbous grade of mine tailing after adjustment;

X ₂₁it is actual zinc grade in concentrate;

it is the concentrate zinc grade after adjustment;

X ₂₂it is actual zinc grade in mine tailing;

it is the mine tailing zinc grade after adjustment;

X ₁it is actual plumbous grade in raw ore;

it is the plumbous grade of raw ore after adjustment;

X ₂it is actual zinc grade in raw ore;

it is the raw ore zinc grade after adjustment;

C ₁, C ₂, C ₃, C ₄, C ₅, C ₆it is the limit D-value between mineral processing production flow process raw ore, concentrate, mine tailing lead, zinc grade;

A ₁, B ₁, A ₂, B ₂be respectively the domain of variation bound of concentrate and tailings productive rate;

A ₁, a ₂be respectively the lower limit set value of raw ore lead, zinc adjustment grade;

B ₁, b ₂be respectively the upper limit set value of raw ore lead, zinc adjustment grade;

A ₁₁, a ₂₁be respectively the lower limit set value of concentrate lead, zinc adjustment grade;

B ₁₁, b ₂₁be respectively the upper limit set value of concentrate lead, zinc adjustment grade;

A ₁₂, a ₂₂be respectively the lower limit set value of mine tailing lead, zinc adjustment grade;

B ₁₂, b ₂₂be respectively the upper limit set value of mine tailing lead, zinc adjustment grade;

3) solution procedure 2) in nonlinear optimization equation, what obtain measures as measurable modified value, and the required amount obtained as the estimated value of required amount, and then obtains the algorithm of many metal balances productive rate.

2. the algorithm of a kind of many metal balances productive rate as claimed in claim 1, is characterized in that: according to priori as constraint condition, by equations turned be optimization problem:

$\begin{array}{cc}\min & {\left(\frac{x_1-{\hat{x}}_1}{{\hat{x}}_1}\right)}^2+{\left(\frac{x_2-{\hat{x}}_2}{{\hat{x}}_2}\right)}^2+{\left(\frac{x_{11}-{\hat{x}}_{11}}{{\hat{x}}_{11}}\right)}^2+{\left(\frac{x_{12}-{\hat{x}}_{12}}{{\hat{x}}_{12}}\right)}^2\\ & +{\left(\frac{x_{21}-{\hat{x}}_{21}}{{\hat{x}}_{21}}\right)}^2+{\left(\frac{x_{22}-{\hat{x}}_{22}}{{\hat{x}}_{22}}\right)}^2\end{array}$

$s.t.\left\{\begin{array}{c}{\mathrm{\&gamma;}}_1+{\mathrm{\&gamma;}}_2-1\&GreaterEqual;0\\ -{\mathrm{\&gamma;}}_1-{\mathrm{\&gamma;}}_2+1\&GreaterEqual;0\\ x_{11}-x_1-C_1>0\\ x_{21}-x_2-C_2>0\\ x_{12}-x_1+C_3>0\\ x_{22}-x_2+C_4>0\\ x_{11}-x_{12}-C_5>0\\ x_{21}-x_{22}-C_6>0\\ x_1-a_1\&GreaterEqual;0\\ -x_1+b_1\&GreaterEqual;0\\ x_2-a_2\&GreaterEqual;0\\ {-x}_2+b_2\&GreaterEqual;0\\ x_{11}-a_{11}\&GreaterEqual;0\\ {-x}_{11}+b_{11}\&GreaterEqual;0\\ x_{12}-a_{12}\&GreaterEqual;0\\ {-x}_{12}+b_{12}\&GreaterEqual;0\\ x_{21}-a_{21}\&GreaterEqual;0\\ {-x}_{21}+b_{21}\&GreaterEqual;0\\ x_{22}-a_{22}\&GreaterEqual;0\\ {-x}_{22}+b_{22}\&GreaterEqual;0\\ {\mathrm{\&gamma;}}_1-A_1\&GreaterEqual;0\\ {-\mathrm{\&gamma;}}_1+B_1\&GreaterEqual;0\\ {\mathrm{\&gamma;}}_2-A_2\&GreaterEqual;0\\ -{\mathrm{\&gamma;}}_2-B_2\&GreaterEqual;0\end{array}\right.---\left(22\right)$

In formula, γ ₁two original rates;

γ ₂it is a smart productive rate;

X ₁₁it is the plumbous grade of two former middle reality;

two former plumbous grades after adjustment;

X ₁₂two former middle actual zinc grades;

two former zinc grades after adjustment;

X ₂₁it is actual plumbous grade in an essence;

it is the lead bullion grade after adjustment;

X ₂₂it is actual zinc grade in an essence;

it is the smart zinc grade after adjustment;

X ₁it is the plumbous grade of a former middle reality;

it is the former plumbous grade after adjustment;

X ₂it is a former middle actual zinc grade;

it is the former zinc grade after adjustment;

C ₁, C ₂, C ₃, C ₄, C ₅, C ₆be that mineral processing production flow process one is former, two former, limit D-value between plumbous in a concentrate, zinc grade;

A ₁, B ₁, A ₂, B ₂be respectively the domain of variation bound of two former and smart productive rates;

A ₁, a ₂be respectively the lower limit set value of a former lead, zinc adjustment grade;

B ₁, b ₂be respectively the upper limit set value of a former lead, zinc adjustment grade;

A ₁₁, a ₂₁be respectively the lower limit set value of two former lead, zinc adjustment grade;

B ₁₁, b ₂₁be respectively the upper limit set value of two former lead, zinc adjustment grade;

A ₁₂, a ₂₂be respectively the lower limit set value of a lead bullion, zinc adjustment grade;

B ₁₂, b ₂₂be respectively the upper limit set value of a lead bullion, zinc adjustment grade.

$\begin{array}{cc}\min & {\left(\frac{x_{21}-{\hat{x}}_{21}}{{\hat{x}}_{21}}\right)}^2+{\left(\frac{x_{22}-{\hat{x}}_{22}}{{\hat{x}}_{22}}\right)}^2+{\left(\frac{x_{33}-{\hat{x}}_{33}}{{\hat{x}}_{33}}\right)}^2+{\left(\frac{x_{34}-{\hat{x}}_{34}}{{\hat{x}}_{34}}\right)}^2\\ & +{\left(\frac{x_{43}-{\hat{x}}_{43}}{{\hat{x}}_{43}}\right)}^2+{\left(\frac{x_{44}-{\hat{x}}_{44}}{{\hat{x}}_{44}}\right)}^2\end{array}$

B： $s.t.\left\{\begin{array}{c}{\mathrm{\&gamma;}}_3+{\mathrm{\&gamma;}}_4-{\mathrm{\&gamma;}}_2\&GreaterEqual;0\\ -{\mathrm{\&gamma;}}_3-{\mathrm{\&gamma;}}_4+{\mathrm{\&gamma;}}_2\&GreaterEqual;0\\ x_{21}-x_{33}-C_6>0\\ x_{22}-x_{34}-C_2>0\\ x_{21}-x_{43}-C_{31}>0\\ x_{22}-x_{44}+C_{41}>0\\ x_{43}-x_{33}-C_{51}>0\\ x_{44}-x_{34}-C_{61}>0\\ x_{21}-a_{12}\&GreaterEqual;0\\ -x_{21}+b_{12}\&GreaterEqual;0\\ x_{22}-a_{22}\&GreaterEqual;0\\ {-x}_{22}+b_{22}\&GreaterEqual;0\\ x_{33}-a_{33}\&GreaterEqual;0\\ {-x}_{33}+b_{33}\&GreaterEqual;0\\ x_{34}-a_{34}\&GreaterEqual;0\\ {-x}_{34}+b_{34}\&GreaterEqual;0\\ x_{43}-a_{43}\&GreaterEqual;0\\ {-x}_{43}+b_{43}\&GreaterEqual;0\\ x_{44}-a_{44}\&GreaterEqual;0\\ {-x}_{44}+b_{44}\&GreaterEqual;0\\ {\mathrm{\&gamma;\&gamma;}}_3-A_3\&GreaterEqual;0\\ {-\mathrm{\&gamma;}}_3+B_3\&GreaterEqual;0\\ {\mathrm{\&gamma;}}_4-A_4\&GreaterEqual;0\\ -{\mathrm{\&gamma;}}_4-B_4\&GreaterEqual;0\end{array}\right.---\left(23\right)$

In formula, solve according to A model algorithm the known terms drawn;

γ ₂it is the known smart productive rate obtained by A algorithm model;

γ ₃two smart productive rates;

γ ₄it is high-precision productive rate;

X ₃₃it is actual plumbous grade in two concentrate;

it is the plumbous grade of two concentrate after adjustment;

X ₃₄it is actual zinc grade in two concentrate;

two concentrate zinc grades after adjustment;

X ₄₃it is actual plumbous grade in high concentrate;

it is the plumbous grade of high concentrate after adjustment;

X ₄₄it is actual zinc grade in high concentrate;

it is the high concentrate zinc grade after adjustment;

X ₂₁it is actual plumbous grade in a concentrate;

it is the plumbous grade of known adjustment obtained by A algorithm model;

X ₂₂it is actual zinc grade in a concentrate;

it is the known adjustment zinc grade obtained by A algorithm model;

C ₆, C ₂₁, C ₃₁, C ₄₁, C ₅₁, C ₆₁it is the limit D-value in mineral processing production flow process one essence, two essences, high concentrate between lead, zinc grade;

A ₃, B ₃, A ₄, B ₄be respectively the domain of variation bound of two essences and high-precision productive rate;

A ₁₂, a ₂₂be respectively the lower limit set value that a concentrate is plumbous, zinc adjusts grade;

B ₁₂, b ₂₂be respectively the upper limit set value that a concentrate is plumbous, zinc adjusts grade;

A ₃₃, a ₄₃be respectively the lower limit set value that two concentrate are plumbous, zinc adjusts grade;

B ₃₃, b ₄₃be respectively the upper limit set value that two concentrate are plumbous, zinc adjusts grade;

A ₃₄, a ₄₄be respectively the lower limit set value that high concentrate is plumbous, zinc adjusts grade;

B ₃₄, b ₄₄be respectively the upper limit set value that high concentrate is plumbous, zinc adjusts grade.

3. the algorithm of a kind of many metal balances productive rate as claimed in claim 1, is characterized in that: according to priori as constraint condition, by equations turned be optimization problem:

$\begin{array}{cc}\min & {\left(\frac{x_{11}-{\hat{x}}_{11}}{{\hat{x}}_{11}}\right)}^2+{\left(\frac{x_{12}-{\hat{x}}_{12}}{{\hat{x}}_{12}}\right)}^2+{\left(\frac{x_{33}-{\hat{x}}_{33}}{{\hat{x}}_{33}}\right)}^2+{\left(\frac{x_{34}-{\hat{x}}_{34}}{{\hat{x}}_{34}}\right)}^2\\ & +{\left(\frac{x_{55}-{\hat{x}}_{55}}{{\hat{x}}_{55}}\right)}^2+{\left(\frac{x_{56}-{\hat{x}}_{56}}{{\hat{x}}_{56}}\right)}^2+{\left(\frac{x_{66}-{\hat{x}}_{66}}{{\hat{x}}_{66}}\right)}^2+{\left(\frac{x_{65}-{\hat{x}}_{65}}{{\hat{x}}_{65}}\right)}^2\end{array}$

C： $\left\{\begin{array}{c}{\mathrm{\&gamma;}}_5+{\mathrm{\&gamma;}}_6-{\mathrm{\&gamma;}}_1-{\mathrm{\&gamma;}}_3\&GreaterEqual;0\\ -{\mathrm{\&gamma;}}_5-{\mathrm{\&gamma;}}_6+{\mathrm{\&gamma;}}_1+{\mathrm{\&gamma;}}_3\&GreaterEqual;0\\ x_{33}-x_{55}-C_{53}>0\\ x_{11}-x_{55}-C_{52}>0\\ x_{34}-x_{56}-C_{54}>0\\ x_{12}-x_{56}-C_{56}>0\\ x_{55}-x_{66}-C_{65}>0\\ x_{56}-x_{65}-C_{66}>0\\ x_{55}-a_{51}\&GreaterEqual;0\\ {-x}_{55}+b_{51}\&GreaterEqual;0\\ x_{56}-a_{52}\&GreaterEqual;0\\ {-x}_{56}+b_{52}\&GreaterEqual;0\\ x_{66}-a_{53}\&GreaterEqual;0\\ {-x}_{66}-b_{53}\&GreaterEqual;0\\ x_{65}-a_{54}\&GreaterEqual;0\\ {-x}_{65}+b_{54}\&GreaterEqual;0\\ {\mathrm{\&gamma;}}_5-A_5\&GreaterEqual;0\\ -{\mathrm{\&gamma;}}_5+B_5\&GreaterEqual;0\\ {\mathrm{\&gamma;}}_6-A_6\&GreaterEqual;0\\ {-\mathrm{\&gamma;}}_6-B_6\&GreaterEqual;0\end{array}\right.---\left(24\right)$

In formula, ${\left(\frac{x_{11}-{\hat{x}}_{11}}{{\hat{x}}_{11}}\right)}^2+{\left(\frac{x_{12}-{\hat{x}}_{12}}{{\hat{x}}_{12}}\right)}^2+{\left(\frac{x_{33}-{\hat{x}}_{33}}{{\hat{x}}_{33}}\right)}^2+\left(\frac{x_{34}-{\hat{x}}_{34}}{{\hat{x}}_{34}}\right)$ Solve according to model algorithm A and B the known terms drawn;

γ ₁the known two original rates obtained by A algorithm model;

γ ₃the known two smart productive rates obtained by B algorithm model;

γ ₅it is high-precision productive rate;

γ ₆it is total tail productive rate;

X ₅₅it is actual plumbous grade in low concentrate;

it is the plumbous grade of low concentrate after adjustment;

X ₅₆it is actual zinc grade in low concentrate;

it is the low concentrate zinc grade after adjustment;

X ₆₆it is actual plumbous grade in total mine tailing;

it is the plumbous grade of total mine tailing after adjustment;

X ₆₅it is actual zinc grade in total mine tailing;

it is the total mine tailing zinc grade after adjustment;

X ₁₁it is actual plumbous grade in a raw ore;

it is the plumbous grade of a known former adjustment obtained by A algorithm model;

X ₁₂it is actual zinc grade in a raw ore;

it is the known former adjustment zinc grade obtained by A algorithm model;

X ₃₃it is actual plumbous grade in two concentrate;

that known two concentrate obtained by B algorithm model adjust plumbous grade;

X ₃₄it is actual zinc grade in two concentrate;

that known two concentrate obtained by B algorithm model adjust zinc grade;

C ₅₃, C ₅₂, C ₅₄, C ₅₆, C ₆₅, C ₆₆be that mineral processing production flow process two is former, two essences, low essence, limit D-value between plumbous in total mine tailing, zinc grade;

A ₅, B ₅, A ₆, B ₆be respectively the domain of variation bound of low essence and total tail productive rate;

A ₅₁, a ₅₂be respectively the lower limit set value that low concentrate is plumbous, zinc adjusts grade;

B ₅₁, b ₅₂be respectively the upper limit set value that low concentrate is plumbous, zinc adjusts grade;

A ₅₃, a ₅₄be respectively the lower limit set value that total mine tailing is plumbous, zinc adjusts grade;

B ₅₃, b ₅₄be respectively the upper limit set value that total mine tailing is plumbous, zinc adjusts grade.