KR20170105692A - Asset management system and device of Water Treatment Plant - Google Patents

Abstract

The present invention relates to a water purification facility asset management system for operating a water purification facility efficiently. The water purification facility asset management system comprises: a first system in which a GAP of data for each item is calculated, and the calculated result value is recorded in a database; a second system in which an analytic hierarchy process (AHP) is performed to generate a weight, and the weight is calculated to a service score by a predetermined method; a third system in which the data of the database is calculated by Fourier analysis to the failure probability model of a separation membrane module and the lifetime of the separation membrane module and is recorded in the database; a fourth system in which the data of the database is extracted as an attribute function by Fourier analysis, and the extracted attribute function is input to the database; and a fifth system in which life cycle costs (LCC) analysis and benefit-cost (BC) analysis are performed based on the data input to the third system and the fourth system.

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an asset management system,

The present invention relates to an asset management system for efficiently operating a water purification facility.

Effective management of water purification facilities should minimize disruptions caused by facilities and accidents, and allocate limited budget appropriately so that water purification facilities can continue to function normally. It is necessary to present a standardized model for the maintenance of the facilities, which is called an asset management system.

Asset management is largely classified into service management and facility management. In the present invention, contents of the service level LoS (Level of Service) analysis of service management and the efficient management method of the separation membrane process .

First, the level of service in the service level LoS (Level of Service) analysis is an index of the intangible value formed in the market by converting it into an engineering concept. This series of processes is presented as a standardized analysis model This is equivalent to developing a Benchmark Index BMI (Benchmark Index) for assessing the level of water facility management services. Developing a benchmark index on the LoS analytical techniques of water treatment plant asset management can be seen as a benchmark for asset management analysis techniques as well as benchmarking management indicators used by successful companies.

The evaluation method for the service level of the water purification facility should be reviewed and selected comprehensively in terms of technical, aesthetic, economic and policy aspects. However, when examining citizen satisfaction with tap water, it is easy to find unidirectional cases such as evaluating by technical basis only by supplier or analyzing by user 's own evaluation. However, the service level is an abstract concept that is determined by the mutual communication until the product delivered by the supplier is delivered to the user and evaluated. Therefore, it is necessary to find out the cause of the gap between the two, and to analyze it quantitatively and find a way to improve it. Accordingly, in the present invention, the GAP between the two is settled by setting the operator of the water purification plant providing the tap water as the supplier and the general citizen using the tap water as the user. In addition, since the service level in the water purification facility is not a total score that simply averages the score of a single item but is a complex in which a number of items are closely connected, it is an analysis method considering interaction. (AHP: Analytic Hierachy Process).

In the background of the efficient management method of the separation membrane process, as the system is aged due to the operation of the pressurized separation membrane, the operation efficiency of the separation membrane is decreased and the maintenance frequency is increased. From the present point of view, we have been conducting the replacement of membranes through managerial experience and short-term performance tests. However, in order to optimize the operation and maintenance of the facilities from a long-term point of view, the lifetime of the membrane modules, It is essential to predict. Effective prediction of the lifetime of the membrane module, which is a core part of the system, predicts the deterioration of the facility, so that the costs associated with the risks arising from the operation of the significantly aged facilities and the inefficiency and costly waste of operations due to low efficiency of the system Can be reduced. This means that it is possible to optimize the operation cost of the facility over the long term by operating and maintaining facilities through asset management techniques.

JP 5167667 B2 JP 2015-215759A

The present invention is to provide a facility management method in terms of service management and asset management in order to efficiently operate a water purification facility.

A system for solving the above-

1. A water purification facility asset management system including a database in which a hollow fiber membrane repair history data of a separation membrane of a water purification facility and operation history data of a purification facility are recorded, The questionnaire data for a plurality of items previously set by the group, the second group, and the third group is inputted and stored in the database, and the questionnaire data for each item among the first group and the second group, A first system in which the GAP of the first system is computed and the computed result is output; The third group of questionnaire data recorded in the database is subjected to Analytic Hierarchical Process (AHP) analysis to generate weights, and the generated weights are input to the database. The weights are calculated and output as service scores according to a predetermined method A second system; A third system in which the hollow fiber membrane maintenance history data stored in the database is calculated and output as the failure probability model of the separation membrane module and the lifetime of the separation membrane module by the Poisson model; A fourth system in which operation history data stored in the database is calculated and output as a predetermined property function by Fourier analysis; (TMP), a cost factor per area, a direct maintenance cost per driving cycle (N), and a driving cycle number (N), based on the data calculated in the third system and the fourth system ) LCC (Life Cycle Costs) analysis and BC analysis (Benefit-cost analysis) and indirect cost (LCC) cost per driving time (N) And a fifth system calculated and output by a cost analysis.

In addition, the predetermined number of items preferably include water quality, taste and smell, water pressure, price and service.

In addition, it is preferable that the GAP values calculated in the database are stored in the first system, and then the GAP values are sorted in a descending order.

More specifically, the present invention includes a step of surveying the satisfaction level between the user and the supplier for the relevant water purification facilities for performing the LoS analysis, performing the GAP analysis between the respective objects, constructing the weight through the AHP analysis, , The failure probability model of the module using Fourier analysis and the Fourier analysis and the property function extraction process based on the operation history data and the life prediction method by the BC analysis to efficiently manage the separation membrane process .

1. Performing GAP analysis between each subject and constructing weights through AHP analysis and assigning them to service scoring (first system and second system)

First, the LoS analysis can create a user / supplier / expert questionnaire that scored the specified items and analyze the GAP analysis among the user providers to identify items with a large difference in satisfaction and to supplement them.

Next, AHP analysis is conducted using the information obtained from the expert survey. The weight of each item can be constructed by evaluating the importance of other items to an item recognized by the respondent through a binary comparison of the results.

For example, the results obtained from the AHP analysis are composed of a pair of comparison matrices as follows, and the weights are determined through the average value of the results.

The consistency ratio (CR) examination confirms that the questionnaire is consistent. When the CR ≤ 0.15 is satisfied, the corresponding pair comparison matrix is determined to be sufficiently consistent, and the weight W value obtained from the consistency ratio review process becomes significant (Kinoshita et al., 1999). In this process, only the weight value of the decision maker who passes the consistency check becomes reliable and it is composed of the weight of the final item of the AHP analysis.

SoC, SoT, SoC, SoS, and so on for each item such as water quality, water pressure, taste, smell, price, and service in the item score by using the benchmark index method as shown in FIG. , And the score for each of the five items is added to a total service score (TSS). In the formula, Sq, Sp, St, Sc, and Ss represent the satisfaction level by item. Satisfaction with each item is calculated as a percentage of all conversion formulas with a score of 5.

In order to efficiently manage the separation membrane process, the failure probability model of the module using the Fourier analysis is derived, and Fourier analysis and attribute function extraction process based on the operation history data (third system and fourth system)

2.1 Derivation of Failure Probability Model of Module using Poisson Model (3rd System)

When a failure occurs in the hollow fiber membranes constituting the membrane module, repairs are performed in the direction of closing the broken hollow fiber membrane through the pin repair. In this case, when the ratio of the number of hollow fiber membranes that have undergone pin repair in one module is 0.45% or more, the module is treated as a breakage of the membrane module in actual operation and the module is replaced with a new one. Through this, it is possible to predict the service life until the breakage of the membrane module occurs. Based on this lifetime, the probability of failure of one membrane module over its lifetime can be determined.

It is assumed that the number of hollow fiber membranes constituting one membrane module is Y, the lifetime of one hollow fiber membrane is Θ [month], the failures of the hollow fiber membranes are independent of each other, and have a common failure rate. Since the failure history of the hollow fiber membrane can be obtained through the pin maintenance history information (Data Base) of the separation membrane, the pin maintenance rate h (t) per month can be extracted based on the failure history information (Data Base).

The maintenance rate (h (t)) of the hollow fiber membranes per month is expressed by the relationship between the non-failure probability function (S (t)) and the failure probability density function f (t) Where h (t), S (t), and f (t) have continuous distributions over time.

In addition, the survival function S (t) has the following relationship.

Therefore, the number of hollow fiber membranes, R (t), that has not been broken even after time t is as follows.

The probability of failure (ν) that can occur during the life of the membrane module (τ) may increase over time due to various external factors. In this case, the failure probability (ν) of the membrane module can be assumed to be ν = D (t) / τ by applying the old function (D (t)) considering the deterioration.

This value is calculated as Matrix for the month t (month) and P _x (t, x) for x to obtain the expected value E _m (t) of the failed module per month in the system, The breakage rate can be derived as follows.

The above results are summarized in a matrix form as shown in FIG.

The probability mass function (PMF) and cumulative probability function (CDF) of the system can be derived with the expectation value E _m (t) derived from this matrix. In this case, the derived PMF and CDF are values by the number of modules used or replaced in the system up to t.

That is, to maintain the system, the number of modules to be replaced is equal to the number of modules to be replaced, since it is assumed that a damaged module is replaced with the same new module every t [month]. Therefore, PMF and CDF can be constructed as follows.

Also, h _m (t) derived from this matrix is the ratio of modules that are broken in the system per month: h _m (t) and PMF (t) since it has a value, h _m (t) and PDF (t) means that the approximation is possible with the same thing. Based on this, the failure rate (h _m (t)) of the module per month in the system can be derived as follows.

Here, S _m (t) → 1 reasons are modules survival function (S _m (t)) of the module because the number of breakage arose immediately replaced yirueojyeoseo even if, within the system to survive within a certain period of time the module is the same all the time Can always be approximated by 1.

로 구할 수 있다.From the results above, the damage generation rate of the module per month from the system (h _m (t)) can be treated the same as breakage mass function (P _failure (t)) of the module, the damage the number of modules per month (δ _failure (t)

.

로 다음과 같이 정의 할 수 있다. However, the number of effective breaks per module per month will appear in the phenomenon as actual constants, and the number of breaks

Can be defined as follows.

and at _{k = 0, 0 ≤δ failure (} t) <0.8, → [δ failure (t)] effectiveness = 0,

at k is a natural number, to a _{(k-0.2) ≤δ failure (} t) <(k + 0.8), → [δ failure (t)] effectiveness = k,

Examples of criteria used here are as follows.

The definition of the cumulative number of _failures ( Δ _failure (t) ) of the membrane module from the operating point to the point in time of operation is defined as follows, using the definition of δ _failure (t) _{effectiveness} presented above.

Actually, in the Y water purification plant, when the number of accumulated failure numbers ( Δ _failure (t) ) of the modules in a system of series exceeds 5, since the separation membrane module of the system of the corresponding series is completely replaced, T _{Renewal of System} , which is the life time of the membrane module in the system (T _{Life Soon} ) .

2.2 Fourier Analysis and Attribute Function Extraction Process Based on Operation History Data (4th System)

The following is a method of extracting an attribute function by Fourier analysis.

In a water purification system that operates a pressurized membrane, the property functions Γ (N), R (N), and α (N) can be expressed as the sum of each periodic function based on historical data. Since the operating environment of the system will be repeated in a similar pattern periodically, the resulting attribute functions will be repeated with each periodicity. Therefore, the important factors involved in this repetitive tendency can be derived through the Fourier transform.

The definition of Fourier transform is as follows.

는 신호의 주파수 영역으로 표현되는 스펙트럼밀도함수, G(w)로 다음과 같이 표현될 수 있다.From an engineering point of view, many irregular signals have several frequency components. It is therefore reasonable to represent the signal as a combination of components of different frequencies. The function of the driving domain (N) domain,

Can be expressed as a spectral density function, G (w), expressed in the frequency domain of the signal as follows.

Thus, from a physical point of view, the spectral density function, G (w), represents the frequency range or frequency characteristics of the signal.

This can be extended to irregular functions by performing a more rigorous expansion process using the Fourier series analysis. For example, if a periodic function with a period T _f = 2T, f (N), is expanded with a Fourier series,

)들을 추출하여, 주기함수, f(N)를 재구성할 수 있다. 이때, 불연속점의 영향으로 생긴 오차를 줄이기 위해서, 처음 값과 마지막 값이 같은 구간의 신호들을 추출해 내어 창함수(Window Function)의 형태로 만든 후에 스팩트럼(Spectrum) 분석을 진행하여 유효한 각주파수를 추출하는 것이 좋다.Here, each frequency ( _k ) corresponding to a _k , b _k having a valid size

), And reconstruct the periodic function f (N). In this case, in order to reduce the error caused by the discontinuity point, signals of the same interval between the first value and the last value are extracted to form a window function, and spectrum analysis is performed to extract valid angular frequencies It is good to do.

(1) The attribute function Γ (N) is extracted by the following method through the historical data as shown in FIG.

Generate p (N), which is a value obtained by subtracting the average of Γ (N) from Γ (N) at T with intervals of [l, m], rather than directly extracting the frequency by Fourier transforming Γ (N) , And p (N) in the form of a window function, and extracts samples of interval T. The expansion of p (N) as a Fourier series with respect to these extracted samples is as follows according to the definition of Fourier transform.

From the above p (N), we construct a _k , b _k , d _k as follows.

)만을 추출하고, 복원에서 발생되는 필수적인 위상차Δφ를 고려하여 p(N)을 복원하고, 그 결과로 부터 Γ(N)을 복원한다.Here, the absolute value _{∥a k ∥> 0, ∥b k} ∥> 0, ∥d k ∥> each frequency having a relative effective size to 0 (

), And restores p (N) by considering the necessary phase difference DELTA phi generated in the reconstruction and restores? (N) from the result.

At this time, the phase difference Δφ is derived by the following equation.

이고, 샘플링(Sampling) 한 개수는 M이고, 샘플링(Sampling) 간격(ΔS)은

로 주어진다.Here, the frequency interval? F is

, The number of sampled samples is M, and the sampling interval (? S) is

.

(2) The attribute function R (N) is also extracted in the same manner as Γ (N).

(N), which is a value obtained by subtracting the average of R (N) from T (N) in T having intervals of (L? N? M), rather than extracting frequencies by directly Fourier- (N) into a form of a window function, and extracts the samples of the interval T. [0043] The expansion of q (N) as a Fourier series for the extracted samples is as follows according to the definition of Fourier transform.

)만을 추출하고, 복원에서 발생되는 필수적인 위상차Δφ를 고려하여 q(N)을 복원하고, 그 결과로부터 R(N)을 복원한다.Absolute value _{∥a k ∥> 0, ∥b k} ∥> 0, ∥d k ∥> each frequency having a relative effective size to 0 (

), Extracts q (N) by taking into account the necessary phase difference DELTA phi generated in the reconstruction, and restores R (N) from the result.

은 다음과 같은 방법으로 추출한다. ③ Property function

Is extracted by the following method.

) [L/m2·hr]는 Hegen-poiseuille식에 의해서 온도와 분리막의 특성 및 용액의 특성을 포함하고 있는 유체의 Flux에 대한 물질 전달계수(Kw) [ L/m2·hr·bar]와 막간차압(TMP) [bar]의 농도차에 의해서 발생하는 삼투압(

) [bar]에 의한 식으로 아래와 같이 표현된다.The permeate flux of the fluid passing through the pores of the membrane

) [L / m2 · hr] is the mass transfer coefficient (Kw) [L / m2 · hr · bar] for the flux of the fluid containing the characteristics of the temperature and the membrane and the solution by the Hegen- The osmotic pressure caused by the difference in the concentration of differential pressure (TMP) [bar]

) [bar] is expressed as follows.

은 분리막 공극의 투과성(Porosity)이고, D_pore는 분리막 공극의 지름(Diameter), μ _s는 유체의 점성(Viscosity), δ_m은 분리막의 두께(Thickness)를 말한다.here,

D _pore is the diameter of the separation membrane pore, μ _s is the viscosity of the fluid and δ _m is the thickness of the separation membrane.

)에 의한 TMP 상승 영향이 상대적으로 작기 때문에 삼투압(

)은 무시될 수 있다. 이와 같은 이유로, TMP의 관계식은 아래와 같이 간략히 표현된다.Since the UF and MF separator are separated by the sieving effect by the pore diameter (D _pores ) of the separation membrane, the solute particles are important factors for the separation of the separation membrane. That is, the solute particles accumulate around the pores of the separator to reduce the size of the pores, and an increase in TMP is required to maintain a certain amount of the permeate flux. Therefore, in the case of UF and MF membranes, the osmotic pressure caused by the difference in the solute concentration between the membranes rather than the rate of increase of the TMP due to the decrease of the pore size of the membrane

) Is relatively small, the osmotic pressure

) Can be ignored. For this reason, the relational expression of TMP is briefly expressed as follows.

The above TMP equation is expressed as pure state, physico-chemical deterioration state and fouling state as shown in FIG.

과의 관계를 찾는 과정은 다음과 같다.The influence of the inter-membrane pressure difference TMP on the operation time (t) and the operation speed (N)

The process of finding the relationship with

으로 가정하면,

은 운전회차(N)에서 운전시간(t)에 따라서 선형적으로 증가한다고 볼 수 있다.here,

Respectively,

Can be seen to increase linearly with the driving time (t) in the driving sequence (N).

은 운전회차(N)가 증가함에 따라서 분리막의 공극 지름의 제곱(

)과 공극을 투과하는 유체의 Flux(

)가 변화하면,

이 변화한다는 것을 확인할 수 있다.From the above equation,

The square of the gap diameter of the separation membrane as the operation number N increases

) Of the fluid passing through the gap and the flux

) Changes,

As shown in FIG.

)와 유입원수의 탁도(NTU(N))가 주기적으로 변화할 경우에는

도 주기적으로 변화한다는 것을 의미한다. 즉, 이런 관련성은

의 각주파수 W_k에서 투과Flux(

)의 각주파수(

)와 탁도(NTU(N))의 각주파수(

)와 각각 일치하는 각주파수의 항(terms)들을 추출하여,

을 투과Flux(

)와 탁도(NTU(N))에 관련된 함수로 표현할 수 있다.Therefore, the permeate flux (N) of the separation membrane corresponding to each operation number (N)

) And the turbidity of the incoming water (NTU (N)) change periodically

Also changes periodically. That is,

At each frequency W _k of the transmission flux (

) Of each frequency (

) And turbidity (NTU (N)) at each frequency

And extracts terms of each frequency corresponding to each of the frequencies,

Through Flux (

) And turbidity (NTU (N)).

은 CIP횟수인 운전횟수(N)가 증가함에 따라서, 분리막의 물리화학적 열화로 인해서 점점 증가할 것이고, 투과Flux(

)와 탁도(NTU(N))의 영향인자들을 선형결합하여

을 다음과 같이 표현할 수 있다.Besides,

Will gradually increase due to the physico-chemical deterioration of the membrane as the number of operations (N), which is the number of CIP times, increases, and the permeation flux

) And turbidity (NTU (N)) are linearly combined

Can be expressed as follows.

는 분리막의 특성과 관련된 TMP 상승인자이고,

는 CIP에 의한 물리화학적 열화(Deterioration)로 인해서 발생되는 TMP 상승인자이고,

는 투과Flux(

)의 주기적 변화와 관련된 TMP 상승인자이고,

는 탁도(NTU(N))의 주기적 변화와 관련된 TMP 상승인자이다.here

Lt; / RTI &gt; is a TMP elevation factor associated with the properties of the membrane,

Is a TMP raising factor caused by physical deterioration by CIP,

Lt; RTI ID = 0.0 &gt; Flux (

), &Lt; / RTI &gt; and &lt; RTI ID = 0.0 &gt;

Is the TMP uptake factor associated with the periodic change in turbidity (NTU (N)).

)와 탁도(NTU)의 각주파수(

)가 서로 구분되지 않고, 동일한 대역대의 각주파수 밴드를 현성한다면,

로 투과(Flux)와 탁도(NTU)의 주기적 변화와 관련된 TMP 상승인자로 다음과 같이 나타낼 수 있다.At this time, the angular frequency of the transmission flux (

) And turbine (NTU) frequency

Are not distinguished from each other and each frequency band of the same band band is present,

(TMP) factor associated with periodic changes in permeation (Flux) and turbidity (NTU).

3. Life prediction method by BC analysis (System 5)

In order to analyze LCC during the Lfie Cycle of the separation membrane water purification process, all costs from the installation of the separation membrane to the disposal of the separation membrane are considered. The cost generated in the separation membrane water purification process is expressed as 'operation and maintenance cost (Costs) per operation cycle (N)', as shown in FIG.

All the costs incurred are at a certain time interval from the point of installation. Since costs incurred not at the time of installation are all affected by the discount rate (i), when evaluating the value based on the installation time, the costs incurred should be converted into value considering the influence of the discount rate (i).

Generally, the LCC analysis converts the reference value to the value at the time of installation (t = 0) based on the conversion of the net value method. That is, all future costs from the point in time of installation (t = 0) are discounted by net price method considering the effect of real discount rate (i). Here, the real discount rate (i) is a method that does not consider the change of the discount rate over time and simplifies the mathematical construction of the LCC analysis technique. Here, the actual discount rate (i) is converted according to whether the time unit of the application of the real discount rate is applied as an annual (Yearly) or monthly, daily, or hourly .

However, even if the actual discount rate (i) is applied in the universal lifetime of the membrane module, the rate of decrease in accounting value is considerably small because the current interest rate is very low and the interest rate is expected to remain very low in the future. Therefore, the life cycle cost (LCC) of the membrane module was calculated by excluding the effect of the real discount rate (i) on the assumption that the actual discount rate (i) is 0 in the membrane module itself.

In the LCC expense formula, X is a material variable. When performing LCC analysis of the whole system, it refers to various facilities such as pumps, buildings, and machinery facilities in addition to the separator.

However, in the membrane filtration process system, the most important element of the LCC analysis is the object with the same characteristics as the membrane module, so X = 1 is not taken into account in the formula. T is the time when the lifetime of the facility is terminated and replaced.

C _Build is the cost of construction of separation membrane water system, C _Maintenance is the maintenance cost, C _Renewal is the total _renewal cost of the membrane module, C _total is the accumulated life cycle cost (LCC) It is cost.

① Construction cost of separation membrane water purification system (C _Build ) Calculation

The cost of construction of the system includes design costs, civil engineering and building construction costs, and the cost of purchasing a membrane module, which is a constant investment in the initial construction of the system.

② Maintenance cost (C _Maintenance ) Detailed cost estimate included in

Operational maintenance costs are subdivided into direct maintenance costs and indirect maintenance costs.

C _Direct is the operating costs incurred by the operation of the facility, including C _Operation , C _Water , and CIP costs.

At this time, the critical power cost (C _operation ) in the direct maintenance cost is affected by the magnitude of the inter-membrane pressure difference (TMP) and the operation time (Δt (N)) per driving sequence (N). Therefore, the power cost C _operation per operation N is calculated by multiplying the cost calculation coefficient J _TMP corresponding to the area (Area N) multiplied by the inter-membrane pressure difference TMP and the operation time t (N) can do. Here, the cost calculation coefficient (J _TMP ) is a statistically calculated average value from the historical data.

Indirect maintenance costs (C _Passive ) are risk-based estimating costs that can occur stochastically during operation of the system. They include C _non-ability that can be caused by system inability, And the replacement cost (C _Module ) of the membrane module caused by the breakage of the membrane module. Here, P _Non-ability is the probability of incapability of the system, P _Module is the probability of failure of the _module in the system, and X _Module is the number of membrane modules constituting the system.

Therefore, the LCC cost (C _LCC ) that occurs during the life cycle of the system is as follows, taking into account all the costs of construction, operation and maintenance.

The benefit costs resulting from the operation of the membrane module in the system can be expressed as the benefit per plant N (B _Produce (N)). The direct benefit depends on the production of tap water produced by the operation of the system during the driving cycle (N) and the tap water price. Here, the amount of tap water produced from the system per driving cycle (N) is calculated as the benefit cost by multiplying the benefit coefficient of tap water per ton (Juan, J _Water ). Here, the amount of tap water produced per operation time N is multiplied by the production flow rate (Flow, Q (N)) multiplied by the _membrane separation area (A _membrane ) of the whole system to the permeation flux (ψ (N) DELTA t (N)).

Here, the production flow rate Q (N) is calculated from the historical data, and it is assumed that Q (N) is constant as the average value of the historical data.

Currently, the domestic water treatment plant does not have a management structure that directly produces water through the tap water produced and sold. Therefore, the water benefit coefficient (J _Water ) is calculated to be similar to the production cost per ton of the _water (ton). Since the tap water production cost of the Y water treatment plant is 80 ~ 90 won, the water benefit coefficient (J _Water ) is 0.1 (KRW 1,000).

J _Water = 0.100

In order to calculate the passive benefits generated by tap water, social benefits are enormous and there is no quantitative data for conversion.

Therefore, the total benefit cost (B _LCC ) incurred during the life cycle of the system is as follows, taking into account all the benefits created from construction to operation and maintenance.

The cost benefit (B / C (N)) of the system per driving sequence (N) is as follows.

B / C (N) indicates whether the system has been operating economically per driving sequence (N). That is, it is an index indicating whether the system records the net profit (1 or more) or the net loss (1 or less) of the period for each driving sequence (N). If the value of B / C (N) falls below 1, it indicates that the operation of the system was not economical. Therefore, the system reports that it lost its economic value when the value of B / C (N) lasts less than 1, and up to this point is the economic life of the system's membrane module.

Therefore, in order to extend the economic life of the system, it is necessary to change the B / C (N) value by increasing the water efficiency coefficient (J _Water ) or increasing the economic efficiency of the system by replacing the membrane module constituting the system. Or more. Here, raising the benefit coefficient of tap water (J _Water ) means increasing the selling price of tap water.

The method of representing the net loss (S (N)) or the net profit (S (N)) of the system per driving sequence (N) is also possible by the following method. Here, if S (N) is negative, it is the net loss amount, and if it is positive, it is the net profit amount.

The cost benefit (B / C _LCC ) analysis during the life cycle of the system is as follows.

The B / C _LCC informs you that all the costs economically spent during the life cycle of the system have been recovered through the benefits. This means that the point at which the B / C _LCC becomes 1 reaches the breaking even point (BEP) from an economic point of view. If the system breaks the BEP, it can operate in a financially stable state in the maintenance of the facility, and it is possible to input the cost independently for the potential maintenance costs incurred in the future. This helps ensure that the value of the facility's assets remains sustainable.

In the case of service level management, it is possible to efficiently and quantitatively cope with the operation concept of the service by numerically describing the abstract concept of service through the present invention. In case of a problem in the facility, As a result of the present invention, it is possible to provide a numerical standard and a cost-effective life prediction, so that preventive asset management can be performed.

1 shows a method of calculating a benchmark exponent.
Fig. 2 shows the failure rate of the module in a matrix format.
FIG. 3 shows the TMP equation divided into a pure state, a physico-chemical deterioration state, and a fouling state.
FIG. 4 shows the costs generated in the separation membrane water purification process in terms of 'operation and maintenance cost (Costs) per operation cycle (N)'.
5 is an example of a user-targeted questionnaire.
Figure 6 is an example of a provider-targeted questionnaire.
7 is an example of a questionnaire to be conducted by a specialist.
Figure 8 is a hypothetical result of the GAP analysis for each item.
FIG. 9 shows a summary of virtual surveys.
Fig. 10 shows the total service level score of the virtual water purification plant finally as one index.
11 is a graph showing the number of pin maintenance numbers of the hollow fiber membranes generated per month in the system based on the hollow fiber membrane repair history data.
Fig. 12 shows the pin finishing rate h (t) of the hollow fiber membrane per module by applying the estimation method of the least squares method (SSS).
13 is a process used to determine the parameters a and b of the estimation equation from the least squares method (SSS).
14 shows the results of analyzing the valid data of Γ (N) by changing it to a normal distribution.
Fig. 15 shows that Γ (N) and actual Γ (N) are edited to have a maximum value of 1.5 when Γ (N)> 1.5).
Fig. 16 shows the result of performing Fourier analysis from p (N) in which the form of the window function is completed.
FIG. 17 shows pr (N) obtained by restoring p (N) based on the extracted angular frequencies.
Fig. 18 shows that Γ (N) is reconstructed from the reconstructed pr (N).
FIG. 19 shows the result of calculating the electric power cost per operation cycle N from the historical data of the virtual water purification system.
Figure 20 shows the monthly costs of maintenance items of a virtual water purification system.
Fig. 21 shows the history data of the production flow rate Q (N) on the basis of the permeation flux (? (N)) of the separation membrane per operation number (N) from the history data.
FIG. 22 shows the result of analysis of the LCC cost (C _LCC ) according to the driving sequence (N).
FIG. 23 shows the result of analysis of the benefit cost (B _LCC ) according to the driving sequence (N).
24 shows the results of B / C (N) and B / C _LCC analysis according to the driving sequence N. FIG.
25 is a graph showing that the economic life can be secured by increasing the tap water production cost.

Hereinafter, &quot; the first group &quot; means a user, and more specifically, a general citizen who uses tap water.

"Second group" means the supplier and, in detail, means the operator of the water purification facility.

"Third group" means an expert, and is not limited, in particular, to anyone who can conduct AHP analysis, such as survey reviewers, academic leaders, and project managers.

&Quot; First system &quot; means a system in which survey data performed by a user / supplier / expert is recorded for convenience of explanation, and GAP analysis is performed based on recorded survey data.

&Quot; Second system &quot; means a system that performs AHP analysis based on expert survey data.

The &quot; third system &quot; means a system for performing a failure probability model and life calculation of a membrane module using a Poisson model.

The &quot; fourth system &quot; means a system for extracting attribute functions using Fourier analysis.

&Quot; Fifth System &quot; means a system that performs LCC analysis and BC analysis.

Hereinafter, the present invention will be described in more detail based on the following examples. However, the following examples are intended to illustrate the present invention, but the scope of the present invention is not limited thereto.

1. Performing GAP analysis between each subject and constructing weights through AHP analysis,

First, the user analysis / provider / expert questionnaire is prepared as shown in FIG. 5, FIG. 6, and FIG. 7, and the GAP analysis is performed for each item. The result obtained by the AHP analysis is composed of a binary comparison matrix as follows.

Divide equation (3) by weight W (equation (2)).

The average value of the result of equation (4) is obtained.

To determine if the questionnaire is consistent, determine the consistency ratio (CR) by:

Since the CR ≤ 0.15 is satisfied, the pairwise comparison matrix can be judged to be sufficiently consistent, and the weight W value obtained from the consistency ratio review process becomes significant (Kinoshita et al., 1999). In this process, only the weight value of the decision maker who passes the consistency check becomes reliable and it is composed of the weight of the final item of the AHP analysis.

FIG. 9 is a result of summarizing the above virtual questionnaire.

The final weight W of each item was composed of mean values of 12 weights only for the questionnaire (T1, T2, T3, T4, T9, T16, T17, T18, T19, T21, T22, The final weights for water quality, water pressure, taste, price, and service are as follows: 0.3754, 0.1830, 0.2391, 0.1036, and 0.0988, respectively. The importance of water quality, taste, odor, Price> service.

In FIG. 10, the total service level score of the Y water purification plant is shown as one index and the evaluation grade is shown according to the score. The evaluation level was rated as high as 90 points or more at the service level, very high at the TSS level, 80 points or higher and 90 points or less at the high level, 70 points or more and less than 80 points as the common level, Less than a point was defined as Low, and a value less than 60 was defined as Very Low. As a result, the TSS of the Y water treatment plant was 70.03 points, which corresponds to the normal service level.

2. For efficient management of separation membrane process, we derive module failure probability model using Fourier analysis and extract Fourier analysis and attribute function extraction based on operation history data.

2.1 Derivation of failure probability model of module using Poisson model

The following is an efficient management method of the membrane module.

11 is a graph showing the number of pin maintenance numbers of the hollow fiber membranes generated per month in the three series system based on the hollow fiber membrane repair history data of the membranes of the three series system of the Y water purification plant operating the domestic pressurized membrane system

It can be seen that the number of hollow fiber membrane pin maintenance (ΔG (t)) per month in the 3 series system occurs near the October after the operation. If the cumulative number of pin repair per month is G (t) It can be seen from the graph that the tendency of t increases. From the ΔG (t) and G (t), the pin finishing rate (h (t)) per month of the hollow fiber membrane per module in the system can be derived.

Here, M is the total number of modules constituting the system (42), and k is the number of hollow fiber membranes existing in one module (8900).

Since the hollow fiber membrane module per pin rate of return (h (t)) is a very small value (10-6), in order to derive the estimation of the effective h (t), least square method to a value of h (t) · 10 ⁹ ( ΣSS) is applied. The result of applying this is shown in FIG.

로서, 1개 모듈의 개월당 핀보수률의 초기조건인 h(0)= 0을 만족한다. 최소자승법(ΣSS)으로부터 추정식의 Parameter a와 b를 확정하기 위하여 이용한 과정은 도 13과 같다.The estimation equation used here is

, Which satisfies h (0) = 0, which is an initial condition of a pin maintenance rate per month of one module. The procedure used to determine the parameters a and b of the estimation equation from the least squares method (ΣSS) is shown in FIG.

When sampling the points having the smallest value distribution from the curved surface, the 10 coefficient pairs (a, b) having the smallest sum of the least squares (ΣSS) between the estimation equation and the actual data are as follows.

(0.05, 2.4), (0.01, 2.9), (0.0001, 4.3), (0.00001, 4,9) (0.000005, 5), (0.000001, 5), and the lifetime τ of the module obtained from this is calculated by the following equation.

The life of the module is estimated by comparing the h (t) function with the sum of least squares (ΣSS). The life of the best module is [127,6] months and the sum of least squares (ΣSS) Based on the results of the smallest differences in the Minimum Value, the life span of the module in the membrane is at least [127.6] months and a maximum of [164.4] months. It can be seen that the useful life of the module is about 10 to 14 years.

2.2 Fourier Analysis and Attribute Function Extraction Process Based on Operation History Data

An example of attribute function extraction based on the following operation history data is as follows.

(N), which is the ratio of the operation time (Δt (N)) according to the operation number (N), based on the history data of the membrane of the three series system of the Y water treatment plant operating the domestic pressurized membrane facility. (2, 4.7143), (7, 3.8529), (13, 2.55) from the actual (N, Γ (N)) are very separated values in the above assumed section, see. As a result of analyzing the valid data of Γ (N) except for the special three values, the mean (μ) is [0.8931] and the standard deviation (σ) is [0.4586] ) Are included with a reliability of 68% in the interval of [0.4345 = (μ-σ) ΓΓ (N) ≤ (μ + σ) = 1.3516]. Therefore, the data is edited in a section in which the interval is simplified (0.5? (N)? 1.5), and the reliability of the data is secured at 70%.

Based on this interval, 0.5 is made the minimum value (Γ (N) <0.5), and 1.5 is made the maximum value when (Γ (N)> 1.5). The edited Γ (N) and actual Γ (N) are shown in FIG.

The range of N that makes p (N) into the form of a window function is defined as (2? N? 13) from this edited Γ (N), and this interval is defined as T, (ต) of Γ (N) edited in Fig.

The result of performing the Fourier analysis from p (N) in which the form of the window function is completed is shown in FIG.

From the result of the Fourier analysis, the frequency of each frequency (w _k ) is analyzed. As a result, the most important angular frequencies in d (k) are k [4, 5, 8, 9, 11, 12]. Based on the results of d (k), the angular frequencies of a (k) are relatively valid when k is [4, 5, 8, 11] 4, 5, 8, 9, 12]. If pr (N) is restored based on the extracted angular frequencies, pr (N) is as shown in FIG.

(N) from the restored pr (N) is as shown in the following equation and FIG.

Here, the number of samples M is 12, and the phase difference ?? is [0.5236].

On the other hand, assuming that there is no change in the ratio of the operating time when N = 0 physically, Γr (0) is defined as 1, regardless of the above equation of Γr (N) By definition, it is assumed that Γr (1) also has a value of 1. Therefore, the formula is established from when Γr (N) is (N ≥ 2).

3. Life prediction method by BC analysis

The following is the B / C analysis method.

The LCC analysis and the B / C analysis of the three series system of the domestic Y water purification plant are based on the results of the TMP simulation model developed previously.

① Construction cost of separation membrane water purification system (C _Build ) Calculation

The cost of construction of the system is 7,122,883 (KRW) in total cost of the system of six pressurized membranes of Y water purification plant, including the cost of the design, civil engineering and construction construction, and the purchase cost of the membrane module. 6, which is [1,190,000] (thousand won).

② TMP area cost factor applied to LCC analysis (J _TMP ) Calculation

Figure 19 shows the result of calculating the electric power cost per operation cycle (N) from the historical data of the three series system of domestic Y water purification plant. Here, since the power cost per operation number N is different from the generation amount of the monthly electric power cost in the history data, the power cost per day corresponding to each operation number is calculated and the operation day number? T (N) The power cost per driving cycle (N) was calculated.

Driving Recurrence (N) After calculating the J _TMP (N) by dividing the TMP area of Area (N) for each power costs, N a [1, 7, 8] TMP B is TMP by the high abnormal operation than TMP _R in J _TMP (1), J _TMP (7), and J _TMP (8) represent negative numbers. Also, when N = 14, the value of J _TMP (14) was 6.54, which was significantly larger than other data. Except for these special cases, the average value of J _TMP from the normal operation data of J _TMP (N) is calculated as follows.

③ Direct maintenance cost per driving cycle (N) (C _Direct (N)

The monthly cost of the maintenance items in 2014 from the historical data of the three series system of domestic Y water treatment plant is shown in FIG.

In the three-system system, the average cost per month in 2014 was 2,883 won, the average cost of CIP drug was 261 thousand won per month, the average operating cost was 1,761 thousand won per month, and the labor cost was 3,218 thousand won per month. And an average of 8,122 won per month. Therefore, the total direct maintenance cost is about 4.6 times the monthly electricity cost.

That is, the direct maintenance cost (C _Direct (N)) generated per driving sequence (N) is as follows.

④ indirect maintenance cost per driving cycle (N) (C _Passive (N)

The probability of system failure (P _non-ability ) in a three-system system is related to the failure rate per month among the 42 membrane modules making up the system. The failure rate of membrane per month was derived from the previous Poisson model. In the previous analysis, the failure probability of a module is less than 1 for 100 months (8.2 years), so the probability of system failure is less than 0.0238. Applying this collectively, the system has a P _non-ability of 0.0238 for a period of 8.2 years.

If the system fails, the system cost (C _System ) is calculated as follows.

Here, since Y water purification plant operates at a ratio of production cost to input cost close to 1, it can be approximated as follows, assuming that the monthly benefit cost and the monthly direct maintenance cost are almost the same.

Therefore, risk costs (REB) by the probability (P _Non-ability) the inability of the 3 Series systems experience is substituted for the average of 8,122 per month (thousands) 2014 direct maintenance costs (C _Direct) standards, and P _{Non- ability} = 0.0236 to calculate the risk cost (REB) per driving sequence (N).

In addition, the probability of module breakage (P _module ) is equal to the P _non-ability of system fire, the number of modules (X) in 3 series is 42, the price per module (C _module ) Is calculated at the Y water treatment plant at 500 (KRW), so the risk cost (REB) can be calculated by substituting this into the following equation.

Therefore, the indirect maintenance cost per operation cycle (N) caused by the breakage probability of the membrane module in the system is as follows.

This formula is only valid for up to 100 months (8.3 years) from the initial use of the system. Because, according to the results of the Poisson probability model, the system is considered to perform full replacement in accordance with the maintenance regulations of the Y water purification plant, since the cumulative number of failures exceeds 5 according to the probabilistic analysis at 103 months.

⑤ Estimation of production flow (Q (N)) and benefit cost (B (N)) per driving sequence (N)

From the history data of the three-system system, the permeate flux (? (N)) of the membrane per operation sequence (N) is obtained, and the historical data of the production flow Q (N) is shown in FIG. Here, since the effective permeation area of the membrane module 1 is [75] (m2 / module), the total permeation area (A _membrane ) of the 3 series system composed of 42 membrane module is 3150 (m2).

The average value of Q (N) from N = 1 to N = 13 is as follows.

Therefore, the benefit B (N) obtained from the operation of the system per driving number N is as follows.

⑥ LCC cost (C _LCC ) Analysis result

22, the LCC cost (C _LCC ) increases and the maintenance cost (C _Manintenance (N)) per driving _sequence (N) gradually increases from N = 15 as the driving _sequence (N) The trend can be confirmed. As a result of the LCC analysis, the lifetime cost accumulated over a long period of time due to the operation and maintenance of the facilities exceeds the initial construction cost of the facilities. Therefore, even if the initial cost is increased, it is effective to reduce the LCC cost of the facility by selecting the products with low maintenance cost and durability from the selection of the system components and paying attention to the optimum maintenance of the system. .

⑦ The convenience cost (N) _LCC ) Analysis result

It can be seen from FIG. 23 that as the operating sequence N of the three-system system increases, the benefit cost B _LCC also increases, and the change in the benefit cost B (N) per driving sequence N is significant. This is because B (N) is greatly influenced by the operation time (t (N)) of the system. B (N) shows a gradual increase from N = 15, but shows a very sharp decline since N = 26. These changes serve as a factor in reducing the economics of the system in B / C analysis. Also, the increase in the accumulated benefit cost (B _LCC ) is small compared to the increase in LCC cost. It will take a very long time to recover the initial investment costs of constructing the system.

(8) B / C (N) and B / C _LCC  Analysis

From FIG. 24, B / C (N) tends to decrease as the operating sequence N of the three-system system increases. B / C (N) is converted to a negative net loss (S (N)) starting from N = 17 when J _Water = 0.100 and at the same time B / C (17) do.

Therefore, the three series systems lose their economics from N = 17, reaching an economic life of 0.88 years from October 2014. This is inefficient in terms of efficient use of assets, since the technical life span of the facility, E (N), is much earlier than N = 23, which is less than 0.7. In order to extend the economic life span of a facility to N = 23, which is a technical life span, the benefit cost generated per driving sequence (N) must be increased. The way to increase the benefit cost (B (N)) is to increase J _Water to 0.127. In other words, if the price of produced tap water is increased 27% from 100 won to 127 won, the 3 series system can secure the economic life until the end of the technical life. It can be seen from FIG. 25 that B / C (N) is improved close to 1 up to the technical lifetime N = 23 through increase of J _Water .

While the present invention has been described with reference to exemplary embodiments thereof, it will be understood by those of ordinary skill in the art that various changes in form and details may be made therein without departing from the spirit and scope of the present invention as defined by the following claims. It will be appreciated that embodiments are possible. Accordingly, the scope of protection of the present invention should be determined by the claims.

Claims (3)

1. A water purification facility asset management system including a database in which a hollow fiber membrane repair history data of a separation membrane of a water purification facility and operation history data of a purification facility are recorded,
Questionnaire data on a plurality of items previously set by the first group, the second group and the third group that are set in advance for the LoS (Level of Service) analysis for the water purification facilities are inputted and stored in the database, A first system in which the GAP of the questionnaire data for each item between the first group and the second group is calculated and the calculated result value is output;
The third group of questionnaire data recorded in the database is subjected to Analytic Hierarchical Process (AHP) analysis to generate weights, and the generated weights are input to the database. The weights are calculated and output as service scores according to a predetermined method A second system;
A third system in which the hollow fiber membrane maintenance history data stored in the database is calculated and output as the failure probability model of the separation membrane module and the lifetime of the separation membrane module by the Poisson model;
A fourth system in which operation history data stored in the database is calculated and output as a predetermined property function by Fourier analysis; And
(TMP), a cost per unit area (TMP), a direct maintenance cost per driving cycle (N), a driving cycle (N) based on the data calculated in the third system and the fourth system, (LCC) cost and driving cost (N) cost per LCC (Life Cycle Costs) analysis and BC analysis (Benefit-Cost) and a fifth system that is computed and output by an analysis,
Water purification facility asset management system.
The method according to claim 1,
Wherein the predetermined number of items includes at least one of water quality, taste and odor, water pressure,
Water purification facility asset management system.
The method according to claim 1,
Wherein the GAP values calculated in the database are stored in the first system,
Water purification facility asset management system.

Images