CN110472753B

CN110472753B - Equipment facility unit evaluation method and device based on deep learning

Info

Publication number: CN110472753B
Application number: CN201910776664.XA
Authority: CN
Inventors: 苏宝炜; 李薇薇; 苏宝昕
Original assignee: Individual
Current assignee: Individual
Priority date: 2019-08-22
Filing date: 2019-08-22
Publication date: 2023-04-18
Anticipated expiration: 2039-08-22
Also published as: CN110472753A

Abstract

The invention discloses a device facility unit evaluation method and a device based on deep learning, wherein the method comprises the following steps: acquiring evaluation item parameters corresponding to the equipment units; the number of the equipment units is N, N is a natural number, and N is more than or equal to 1; assigning and scoring each evaluation project parameter to obtain a score corresponding to each evaluation project parameter; establishing an assignability problem mathematical model for all evaluation project parameters, and solving a total evaluation quantitative index function; and calculating the quantitative probability of each evaluation item parameter by using a Bayesian inference strategy. The method realizes the evaluation of the items to be optimized, reduces the labor cost and improves the working efficiency.

Description

Equipment facility unit evaluation method and device based on deep learning

Technical Field

The invention relates to the technical field of equipment and facility maintenance management, in particular to an equipment and facility unit evaluation method and device based on deep learning.

Background

At present, the modern property asset operation and maintenance management service industry safely, reliably and prospectively predicts possible faults and risks of engineering equipment in advance, and is a basic guarantee for exerting the advantages of advanced technical equipment and realizing that enterprises provide high-quality services with high efficiency, low consumption, flexibility and on time. With the technical progress of the facility and equipment system of the modern property building engineering, the construction of the engineering system of the building is developed towards systematization, automation and technology-intensive direction, the facility and equipment engineering system is more complex and the function is more powerful; at the same time, the time and space dimensions of daily use and maintenance management of equipment facilities are rapidly changing, and these changes are not guaranteed by a long-lasting matched equipment maintenance mechanism.

The equipment maintenance management refers to a general term of management activities performed to ensure the normal operation of the equipment or facility or maintain the proper state and function of the equipment or facility, and the management contents may include: the daily guarantee of the equipment, zhou Bao, lubrication, inspection, spot inspection, major repair, intermediate repair, minor repair and the like. Existing device Maintenance management technologies include time-limited preventative Maintenance PM (predictive Maintenance), condition-limited predictive Maintenance PdM (predictive Maintenance), reliability-centric Maintenance RCM (Reliability centered Maintenance), and state-limited device Maintenance CBM (Condition-Based Maintenance).

The preventive maintenance refers to scheduled regular equipment maintenance and spare and accessory part replacement, and comprises several types of maintenance, regular use inspection, regular function detection, regular overhaul, regular replacement and the like; generally, the three-stage maintenance system is a daily maintenance system, which can be referred to as a first-stage maintenance system, and is generally performed by an operator of the equipment, and the main purpose of the maintenance system is to hope that the operator finds obvious faults or symptoms, such as performing daily cleaning, checking operation sound, checking whether normal starting is possible, and checking that the basic parameter, air pressure and temperature are correct. The second-level maintenance is performed every XX hours, generally completed by an operator and an equipment management department, and the overhaul (third-level maintenance) is performed every YY hours, generally completed by an equipment manufacturer if necessary by the equipment department.

The predictive maintenance is to collect, count, analyze and judge the degradation trend, fault position and reason of the equipment by using various means and professional judgment of people, predict the change development and provide precautionary measures to prevent and control possible faults. In this sense, maintenance schedules may be inconsistent between multiple machines for the same type of equipment; the maintenance schedule for the same equipment may not be consistent from year to year. Like preventive maintenance, predictive maintenance is also prospective, and the greatest advantage of applying predictive maintenance is to control the costs incurred by emergency maintenance such as repair or replacement due to equipment damage, while reducing operational disruption of the equipment.

Regarding the RCM with limited reliability as the center, in order to avoid the occurrence of the situations of excessive PM and PdM, the RCM has more scientific and reasonable maintenance work, and the RCM performs function and fault analysis on the system to determine the consequences of each fault in the system; determining preventive countermeasures for the consequences of each fault by using a normalized logic decision program; on the premise of ensuring safety and integrity by means of field fault data statistics, professional evaluation, quantitative modeling and the like, the RCM optimizes the maintenance strategy of the system by taking the minimum maintenance shutdown loss and the minimum maintenance resource consumption as targets.

With respect to the state-limited equipment maintenance CBM, in actual work, the reliability requires that the system must normally operate within a time period of 0-T, while the requirement for availability is relatively low, the system can be out of order and then repaired within a time period of 0-T, and the availability can still be counted as long as the system can normally operate after being repaired; thus, the availability is greater than or equal to the reliability. For this reason, abnormality of the equipment is judged based on equipment status information provided by advanced status monitoring and diagnostic techniques, a failure of the equipment is predicted, and a maintenance and repair plan is arranged reasonably based on the predicted failure information, that is, based on the health status of the equipment, and a mode of performing equipment maintenance is called status maintenance or visual maintenance. Because the different working environments of the system cause that the damage curve and the service life of the same type of system are different when the same type of system runs in different environments, the state-based equipment maintenance mode is used, the health state of the system is not monitored, managed and maintained in real time, the maintenance cost can be greatly saved, unnecessary maintenance expenditure is reduced, and the running availability of the whole system is provided.

Because the above technologies for managing and maintaining the equipment and facilities have their advantages and limitations, in the process of transforming and upgrading the traditional property management to the modern property asset operation and maintenance management service, it is necessary to prevent the poor tendency of insufficient execution or excessive prevention, which is easy to occur in the practical link. The method has the advantages that overall continuous management and control are lacked, project parameters needing to be optimized cannot be accurately obtained, so that the optimized projects needing service are unclear, projects not needing service can be included in a service range, a large amount of labor cost is needed, the working efficiency is low, waste work which does not produce direct value-added effects and does not belong to auxiliary value-added is directly caused to be increased gradually, and the labor cost is increased; in the past, the device is difficult to self-pull due to the fact that the device is trapped in a low water average balance trap of the service, the working efficiency is low, and the service quality is poor.

Disclosure of Invention

The invention aims to provide an equipment facility unit evaluation method and device based on deep learning, so as to realize evaluation and obtain items needing to be optimized, reduce labor cost and improve working efficiency.

In order to solve the above technical problem, the present invention provides an equipment and facility unit evaluation method based on deep learning, including:

acquiring evaluation item parameters corresponding to the equipment units; the number of the equipment units is N, N is a natural number, and N is more than or equal to 1;

assigning and scoring each evaluation item parameter to obtain a score corresponding to each evaluation item parameter;

establishing an assignability problem mathematical model for all evaluation project parameters, and solving an overall evaluation quantitative index function;

and calculating the quantitative probability of each evaluation item parameter by using a Bayesian inference strategy.

Preferably, the method further comprises:

and selecting the evaluation item parameters to be optimized according to the quantization probability of each evaluation item parameter.

Preferably, the number of the equipment units is 5, and the expression of the assigned problem mathematical model is as follows:

the expression of the overall evaluation quantization index function is as follows:

wherein, C _ij An evaluation score representing that j question occurs corresponding to the ith equipment unit; the first constraint is: each equipment unit can only correspond to one critical problem; the expression of the first constraint is as follows:

the second constraint is: each question of importanceThe question can only appear once correspondingly; the expression of the second constraint condition is as follows: />

Preferably, the values of each evaluation item parameter are assigned and scored by adopting a Borda program counting method.

The invention also provides a device for evaluating equipment facilities based on deep learning, which is used for realizing the method and comprises the following steps:

the acquisition module is used for acquiring evaluation item parameters corresponding to the equipment units; the number of the equipment units is N, N is a natural number, and N is more than or equal to 1;

the scoring module is used for assigning and scoring each evaluation project parameter to obtain a score corresponding to each evaluation project parameter;

the solving module is used for establishing an assignability problem mathematical model for all evaluation project parameters and solving a total evaluation quantitative index function;

and the calculation module is used for calculating the quantitative probability of each evaluation item parameter by using a Bayesian inference strategy.

Preferably, the apparatus further comprises:

and the selection module is used for selecting the evaluation item parameters to be optimized according to the quantization probability of each evaluation item parameter.

wherein, C _ij An evaluation score representing that j question occurs corresponding to the ith equipment unit;the first constraint is: each equipment unit can only correspond to one critical problem; the expression of the first constraint is as follows:

the second constraint is: each important problem can only appear once correspondingly; the expression of the second constraint condition is as follows:

preferably, the scoring module is specifically configured to assign a score to each evaluation item parameter by using a bolda program counting method to obtain a score corresponding to each evaluation item parameter.

The invention provides a method and a device for evaluating equipment facility units based on deep learning, which are used for obtaining evaluation item parameters corresponding to the equipment units; the number of the equipment units is N, N is a natural number, and N is more than or equal to 1; assigning and scoring each evaluation item parameter to obtain a score corresponding to each evaluation item parameter; establishing an assignability problem mathematical model for all evaluation project parameters, and solving a total evaluation quantitative index function; and calculating the quantitative probability of each evaluation item parameter by using a Bayesian inference strategy. Therefore, assignment scoring, mathematical model establishment and function solution are carried out, the quantization probability of the parameters of the evaluation items is finally obtained, deep learning is carried out, the quantization probability of each final evaluation item is obtained through calculation, the items to be optimized can be directly determined through the probability, the target of the items to be optimized is clear, the items to be optimized are served, waste of a large amount of manpower and material resources in the items not to be optimized is avoided, the labor cost is reduced, and as the items to be optimized are clear, useless tedious work is avoided, and the work efficiency and the service quality are improved.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the provided drawings without creative efforts.

FIG. 1 is a flow chart of a deep learning-based equipment facility unit assessment method according to the present invention;

FIG. 2 is a schematic diagram showing the comparison of the cost of the optimization group and the cost of the control group;

FIG. 3 is a graphical visualization of the results of solving a likelihood function formula using a function;

fig. 4 is a schematic structural diagram of an equipment and facility unit evaluation apparatus based on deep learning according to the present invention.

Detailed Description

The core of the invention is to provide an equipment facility unit evaluation method and device based on deep learning, so as to realize evaluation and obtain items to be optimized, reduce labor cost and improve working efficiency.

In order to make the technical solutions of the present invention better understood, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

Referring to fig. 1, fig. 1 is a flowchart of a deep learning-based equipment facility unit evaluation method provided in the present invention, the method including the following steps:

s11: acquiring evaluation item parameters corresponding to the equipment units;

wherein the number of the equipment units is N, N is a natural number, and N is more than or equal to 1;

s12: assigning and scoring each evaluation item parameter to obtain a score corresponding to each evaluation item parameter;

s13: establishing an assignability problem mathematical model for all evaluation project parameters, and solving a total evaluation quantitative index function;

s14: and calculating the quantitative probability of each evaluation item parameter by using a Bayesian inference strategy.

Therefore, the method includes the steps of assigning and scoring, establishing a mathematical model, solving a function, finally obtaining the quantization probability of the parameters of the evaluation items, carrying out deep learning in the way, obtaining the quantization probability of each final evaluation item through calculation, directly determining the items to be optimized through the probabilities, serving the items to be optimized, determining the target of the items to be optimized, avoiding wasting a large amount of manpower and material resources in the items not to be optimized, reducing the labor cost, avoiding useless and tedious work due to the fact that the items to be optimized are determined, and improving the working efficiency and the service quality.

Based on the method, the equipment unit is specifically a water supply equipment facility unit. The item evaluation parameter herein represents a specific category of the item. The evaluation project parameters comprise a long running state, personnel in-out conditions, daily customer repair, daily customer complaints and running performance conditions. Further, evaluating the project parameters further comprises: physical environment operation, spot check of good conditions, active inspection maintenance, routine maintenance conditions, and timely maintenance conditions.

Further, after step S14, the method further includes the following steps:

s15: and selecting the evaluation item parameters to be optimized according to the quantitative probability of each evaluation item parameter.

Further, the number of the equipment units is 5. Taking a building A as an example, the building A is provided with No. 1-5 water supply equipment facility units.

Wherein, a service quality SERVQUAL model evaluation method is adopted, specifically, a new service quality evaluation system is provided in the service industry according to the SERVQUAL theory and the theory core is a service quality gap model, namely: the quality of service depends on the degree of difference between the service level perceived by the user and the service level desired by the user, and is therefore also referred to as a "expectation-perception" model, the user's expectation is a prerequisite for the development of a premium service, the key to providing a premium service being in excess of the user's expectation.

In step S12, each evaluation item parameter is assigned and scored by using a bolda program counting method. Counting by using a Borda program, specifically, respectively adopting a plurality of evaluation detailed items according to a SERVQUAL model, and evaluating the comprehensive condition of each group of equipment units, wherein the evaluation detailed items are respectively as follows: the system comprises a running long state, personnel in-and-out conditions, daily customer repair, daily customer complaints, running performance conditions, running physical environment, spot check good conditions, active inspection maintenance, routine maintenance conditions, maintenance timeliness and the like. And then, scoring by utilizing the values assigned by the Borda program, and basically: according to the professional opinion and self judgment of an evaluator, each detail item is scored, the score is from 0 to 5.0, the evaluation result is shown in table 1, table 1 is a score table of evaluation categories corresponding to the five water supply equipment facility units, and the evaluated value is the score of the evaluation category. The evaluation category is the parameter category of the evaluation item.

TABLE 1

Then, the key point probability is calculated, i.e., step S13 and step S14 are performed. In step S13, an assignment problem mathematical model needs to be established, and according to the requirement of the theory for identifying the key equipment unit, the mathematical model is now converted into an "assignment problem" for relating to the deformation in the operational research, that is, firstly, the problem assignment, secondly, the optimal screening is performed, and then, importance ranking is performed on different problems occurring in different equipment.

Wherein, the expression of the mathematical model of the assignability problem is as follows:

the expression of the overall evaluation quantization index function is:

wherein, C _ij An evaluation score representing that j problems occur corresponding to the ith equipment unit; the first constraint is: each equipment unit can only correspond to one critical problem; the expression of the first constraint is:

the second constraint is: each important problem can only appear once correspondingly; the expression of the second constraint is:

wherein Z is a target, the target is a total evaluation quantization index, the positive correlation is minimum and is expressed as min Z, namely, the minimum value of the total evaluation quantization index Z is solved, namely, the value of a mathematical model expression is solved, so that Z is minimum, and the constraint condition for solving the minimum value of Z is as follows: each equipment unit can only correspond to one critical problem and each important problem can only correspondingly appear once. x is the number of _ij Represents the value of an assignability problem mathematical model, i.e., a mathematical model that evaluates a project parameter. j is a natural number, and j is more than or equal to 1 and less than or equal to 5. The value of i is a natural number, and i is more than or equal to 1 and less than or equal to 5.C _ij In particular scores for the evaluation categories corresponding to the five water supply equipment utility units. j =1 indicates that the long state of the operation fails, j =2 indicates that the entering and exiting condition of the personnel fails, j =3 indicates that the repair of the daily client fails, j =4 indicates that the complaint of the daily client fails, and j =5 indicates that the performance condition of the operation fails. The first constraint condition indicates that each equipment unit can only correspond to one fault in five equipment units, namely, one equipment unit can only correspond to one fault problem and cannot simultaneously correspond to a plurality of fault problems. The second constraint condition indicates that each fault problem can only correspond to one fault in five fault problems, that is, one fault problem can only correspond to one equipment unit, and cannot simultaneously correspond to a plurality of equipment units. According to the first constraint condition and the second constraint condition, each equipment unit can only correspond to different fault problems respectively, and the fault problems of each equipment unit are all out of phaseSimilarly, five equipment units correspond to five fault problems respectively. The critical problem and the important problem both represent fault problems and have the same meaning.

Specifically, for the solution of the formula (2), the Excel spreadsheet high-order planning is used for solving, the solved result is shown in table 2, table 2 is a numerical result table of the mathematical models of the evaluation categories corresponding to the five water supply equipment facility units, and the evaluated value is the value of the mathematical model of the evaluation category.

TABLE 2

Then, bayesian inference is used to calculate the probability, i.e. step S14 is performed. Specifically, modern property asset management operation and maintenance service enterprises cannot realize development transition of management services only by improving standards, strengthening requirements and optimizing systems, and because mechanical replication cannot bring more unexpected surprise to people, service organizations need a preposed accelerator for service promotion, namely Bayesian inference (Bayesian inference) and a famous Bayesian formula provided by the Bayesian inference, wherein the formula is as follows:

P(A|B)＝P(B|A)*P(A)/P(B) (3)

assuming that a first condition and a second condition are known, the first condition is a random variable M (Machine) of the state of the water supply equipment unit, the M has two states { working, broken }, namely, normal operation and fault occurrence, the probability P (M = working) =0.99 of normal operation, and the probability P (M = broken) =0.0 of fault occurrence. The second condition is Quality Q (Quality), which represents the effect of the operation of the water supply equipment unit, and the effect may be good or not good, so Q includes two states { good, bad }.

According to the historical operation record of the water supply equipment unit, on the premise that the equipment normally operates, the probability P of good quality (Q = good | M = working) =0.99, which indicates that 99% of the probability operation effect is qualified, and the probability P of bad quality (Q = bad | M = working) =0.01, which indicates that 1% of the probability operation effect is unqualified. On the premise of abnormal failure operation, the probability P of good quality (Q = good | M = broken) =0.6 indicates that the operation effect is acceptable at a probability of 60%, and the probability P of bad quality (Q = bad | M = broken) =0.4 indicates that the operation effect is not acceptable at a probability of 40%.

The quantization probabilities of the evaluation item parameters are calculated according to the bayesian formula P (M | Q) = P (M) × P (Q | M)/P (Q), and as shown in table 3, table 3 is a calculation result table of the quantization probabilities of the evaluation categories corresponding to the five water supply equipment facility units, and the evaluated values are the values of the quantization probabilities of the evaluation categories.

TABLE 3

Referring to table 2 and table 3, it can be known from the quantitative probability calculation results of the equipment unit a, that the integrated joint correction optimization strategy should be specifically executed for the customer service repair issue No. 3 (47.06%) and the customer complaint issue No. 4 (44.69%), and specific continuous improvement and optimization should be performed for the corresponding equipment maintenance management.

In addition, resource customized improvement needs to be integrated, and continuous improvement refers to continuous improvement and perfection of different fields or working positions of an enterprise, and the literal meaning of the continuous improvement is as follows: by modification.

The specific steps comprise the following specific contents:

step one, selecting an important work task and clarifying the reason for selecting the project or the work task;

step two, collecting data, and before starting, the essence of the current situation of the project must be clarified and analyzed;

step three, deeply analyzing the collected data so as to be capable of clearing the real background and reason of the things;

step four, researching the countermeasure on the basis of analysis.

Step five, importing and executing a strategy;

step six, observing and recording the influence after adopting the countermeasures;

step seven, modifying or re-formulating the standard to avoid the recurrence of similar problems;

step eight, checking the whole process from the step one to the step seven so as to introduce the next action.

In addition, the visualization and active communication of the problem solving process and the establishment of efficient record documentation are helpful for the promotion of continuous improvement activities. Recording service information data of an 'optimization group' and a 'comparison group' in each month, counting 2018 year-round data, and serving each department of the center according to property assets projects, wherein the service information data comprises the following steps: human resources (providing wage and welfare data), engineering technologies (providing hydroelectric energy data) and financial purchasing (providing maintenance consumable data), and integrated information results provided on the basis are compared in two groups. Referring to fig. 2, fig. 2 is a schematic diagram showing the comparison between the cost of the optimization group and the cost of the comparison group, in order to illustrate the effect of the method, the optimization group and the comparison group are divided, and finally, the difference actually generated in the equipment maintenance cost of the two groups in the same period of one year is compared with each other, so as to verify the correctness of the forecast. The optimization group means the cost consumed after determining the project to be optimized by the method and optimizing the project, and the comparison group means the cost consumed after obtaining the project to be optimized without a method and optimizing. As can be seen from fig. 2, although the two groups differ from each other in the initial month of the year, the cost of the optimization group is saved by about 1 ten thousand yuan compared with the control group because of the adoption of the identification emphasis targeted improvement measures in the optimization group.

It should be noted that, the emphasis here is on the advanced modern property asset operation and maintenance management service concept, which plays a role in improving the value of the intelligent building; but relative to the abstract concept, the property asset price and cost which are similar and familiar to the public can be more easily understood and approved. As proved by market examination, people can find that the office building with high survival rate is not necessarily low in rent but is necessarily good in service. More and more customers have realized that using advanced asset management concepts can maximize the profitability of real estate assets, helping to increase and maintain real estate values. Therefore, the value added to the concerned assets can be reflected from the price of the property in the practical verification stage; it is particularly noteworthy that this does not mean that the value is simply equivalent to the price.

In the beginning of planning, the performance improvement practice activity tries to avoid the influence of human and some accidental factors, and by means of two groups of data of independent sources of an 'optimization group' and a 'comparison group', a conclusion is drawn with a very high confidence coefficient through scientific statistical calculation and comparative analysis on a large amount of service record data generated by the property asset project in one year, namely the performance improvement practice of an innovation system driven by the SIM is believed to be really effective for the improvement of the property asset value.

The method is particularly an SIM (Service Information Modeling) deep learning intelligent evolution algorithm which is based on customer requirement prejudging management Service practice and can effectively prolong the Service life of equipment. According to the theoretical elliptic orbit model, equipment maintenance management in the traditional sense is converted into collection, arrangement, evaluation and classification based on Service Information by utilizing a deep Information learning mode drive of an SIM (subscriber identity module) Service through an integrated connected thinking cognition mode, then analysis and judgment are carried out by means of planning solution and a Bayesian algorithm, corresponding optimization measures are taken for future services, and finally the scientific effectiveness of the method is verified. The novel technical method can provide differentiated experience for customers, and meet the current dominance and future potential service requirements of the customers; meanwhile, the system can help to optimize building maintenance and efficient operation service. The science effectively links "human books", "assets" and "services". The SIM intelligent evolution algorithm links the three with the physical world objects, helps to enhance comprehensive competitiveness of property asset operation and maintenance service enterprises in a complicated and complicated business environment characterized by VUCA (unstable), uncertain, complex and fuzzy) in a new economic era of large service, and finally realizes intelligent flexible management with property asset appreciation effect and optimal service practice of prospective customization! The method is based on comprehensive demand information of a service process, adopts an SIM intelligent evolution technical algorithm taking the increase of resources as a core, namely, the provision of property services in the traditional sense is changed into the collection, arrangement, evaluation and classification based on service information, then, the analysis and the judgment are carried out by means of planning solution and a Bayesian algorithm, and corresponding optimization measures are taken for future services and the effectiveness of the method is checked; and further extends beyond the management and maintenance of equipment facilities, and can even extend to wider enterprise management fields.

The real service process is much more complicated than the case, and the quantitative evaluation value of each equipment unit is quite complicated and close to the constitution, in this case, the optimum experimental result of the property value increment without the SIM drive is difficult to achieve the scientific and reasonable effect. In addition, on the basis, deep value mining is carried out, and continuous upgrading of knowledge management in property projects is realized.

First, knowledge management and experience summarization are required. On the way of searching the modern asset service value origin, the traditional experience has a lineage logical reasoning, and a rational deduction for the future innovation development is needed; at the same time, practitioners have also found limitations in the above methods related to equipment maintenance. From the aspect of time, the service process focuses on not only the current but also the problem needs to be analyzed and solved from the perspective of sustainability; from the aspect of space, the whole service process is not limited to one or more property assets, and the final aim is to realize the extension of the property assets in the time and space dimensions; in addition to earning profits, there is a need to create new value to customers; while pursuing economic benefits, the enterprise can only realize long-term sustainable development by taking the triple bottom line principle of social benefits and environmental benefits into consideration.

The failure to pay enough attention to the existing method can not reflect the asset value-added effect if the equipment management level is maintained in a low water average balance state for a long time, so attention needs to be paid to the summary and promotion of the experienced screening judgment. This process, like the current machine learning, evolves spontaneously and completely, and we turn it into "deep training". And (3) performing importance identification and sequencing through the service information big data, judging and improving an optimization method, and improving the complete spiral management efficiency.

Secondly, the maximum likelihood estimation method is adopted for assistance. Before the efficiency of the innovation system driven by the SIM is improved and optimized, the management layer does not explicitly inform operators of the specific equipment composition of the 'optimization group' and the 'comparison group' to adopt 'back-to-back' operation, and the routine maintenance management of the group can be normally carried out; that is, the objective is to create two separate sets of independent sources of information that are validated through scientific process statistics methods and process data statistics that require less human assumption to affect the optimization results. After eliminating the influence of some accidental factors, it is believed that if the comprehensive cost (converted into cost) of the statistical "optimization group" after one year is actually reduced compared with the comprehensive cost (converted into cost) of the "comparison group", it can be said that the efficiency of the innovative system under the drive of the SIM is improved and optimized, and is really effective for improving the property value.

The 'integrated join service space-time theory' takes the property asset value as the core, so that one more step is needed to be added for knowledge management; in 12 months of a year, optimization and adjustment measures taken by a device maintenance management team in the month need to be analyzed and mastered most effectively according to actual operation and maintenance data of property projects, and the service team can achieve the management service level, wherein the maximum likelihood is needed, and the maximum likelihood is also an important subject in machine learning.

Project data accumulated for 12 months in a year are: the desired excellent probability θ is a model parameter, and the maintenance solution model can be assumed to be a "binomial distribution"; wherein, the month with the lifting rate of more than 2% is determined to be excellent, refer to table 4, and table 4 is a record table of the improved and optimized effect of the statistical annual adjustment.

TABLE 4

01 month

Month 02

03 month

04 month

05 month

06 months

Month number 07

08 month

09 month

Month of 10

11 month

12 month

In general terms

Excellent

In general terms

Excellent

In general terms

The likelihood function is formulated as: f (X, theta) = theta ⁷ (1-θ) ⁵ . The solution is performed by using binomdst function in Excel spreadsheet, that is, when θ =0.5833=60%, the likelihood function takes maximum value, that is, the "binomial distribution" reaches maximum value at 7 th time. Referring to fig. 3, fig. 3 is a visualization graph of the result of solving the likelihood function formula by using the function, and as can be seen from fig. 3, the maximum value is at 7 months, that is, the adjustment measures are optimized all the year round, and the adjustment scheme adopted in the month should become a property asset project, that is, a part of the knowledge management accumulation of the property enterprise.

Furthermore, extended applications for performance enhancement are needed. If the economic index is taken as the measurement standard, the work payment without direct or indirect value is found, namely the waste phenomenon is very serious. This waste may manifest as, for example: the method has the advantages of inappropriate control of important matters caused by comprehensive factors, insufficient anticipation of things which may happen in the future, unfavorable adoption of targeted optimization measures and the like. Therefore, in the process of continuous improvement, the work needs to be carried out aiming at the internal requirements of the points. The best practice of the research of the property asset value-added system driven by the SIM and the improvement management of the equipment maintenance efficiency is deep excavation and attempt aiming at the problem; taking the operation data of the sample item in 2018 as an example, a scene is assumed to illustrate the role of scientific and reasonable equipment maintenance management in the operation of a certain office building.

According to the cost of 5 ten thousand of centrifugal water pumps and the calculation of 6 years of normal use, if a proprietor (namely a client of a property enterprise) normally uses the centrifugal water pumps, the cost allocated to each year is 0.833 ten thousand. If the owner does not have due maintenance and has to change it in year 3, then the amortization is 6 years and the costs will include: 0.833 ten thousand each year in the first 3 years and 0.919 ten thousand each year in the last 3 years. The life of the newly replaced water pump is still 6 years (the replacement cost of the water pump in the actual market is increased by 5% per year), the total cost is increased to 5.256 ten thousand, and the replacement cost is increased by 0.256 thousand compared with the original plan.

The owner of the water pump can also choose better, if the equipment is maintained properly and certain necessary system components are replaced regularly, the service life of the water pump can be extended to 9 years. The 5 ten thousand initial investment costs are amortized to 6 years, i.e. 0.833 ten thousand per year. Since the annual pump replacement costs are increasing at 5% per year, the replacement cost at year 9 is 6.381 ten thousand. However, after the replacement delay, the actual cost of the new pump between the 7 th year and the 9 th year is 0. In total, the cost of each year in the first 6 years is 0.83 ten thousand, and the cost of the last 3 years is 0.

Pump replacement is postponed for 3 years due to additional maintenance, which saves the owner 3.191 million savings compared to 6 years for replacing a pump whose life has expired. Of course, the owner should also take into account the cost of replacing the necessary system components during this period; even if the owner invests 1 year in the service cycle and is universal to the maintenance and management of prospective customized equipment, the return on investment can be improved by using a method of postponing the replacement of the pump body. Only by taking the case project water supply system pump body as an example, the expenditure of 1 pump is saved by 26.75%, namely 2.191 ten thousands, 5 pumps are used for each pump, the project ABCDE has 5 seats, and the total saving is about 54.780 ten thousands by taking 9 years as a calculation period.

If the achievement of the best practice of SIM-driven property value-added system research and equipment maintenance efficiency improvement management is expanded to more fields of modern property asset operation and maintenance management service processes, the excellent service required by high-end property assets can be realized by integrating integrated service information; namely, the human-oriented, hardware equipment and service process are well connected; the SIM is used as a driving force to search the original value source, and the goal of keeping and developing property assets and increasing value is realized.

Referring to fig. 4, fig. 4 is a schematic structural diagram of an apparatus for evaluating a device facility unit based on deep learning according to the present invention, the apparatus being configured to implement the method described above, and the apparatus including:

an obtaining module 101, configured to obtain an evaluation item parameter corresponding to an equipment unit; the number of the equipment units is N, N is a natural number, and N is more than or equal to 1;

the scoring module 102 is configured to assign and score each evaluation item parameter to obtain a score corresponding to each evaluation item parameter;

the solving module 103 is used for establishing an assignability problem mathematical model for all evaluation project parameters and solving a total evaluation quantitative index function;

and the calculation module 104 is used for calculating the quantitative probability of each evaluation item parameter by using a Bayesian inference strategy.

Obviously, in the device, the scoring module assigns and scores, the solving module establishes a mathematical model and solves functions, the calculating module calculates the quantization probability, deep learning is carried out, the quantization probability of the evaluation project parameters is finally obtained, the quantization probability of each evaluation project is finally obtained, the projects needing to be optimized can be directly determined through the probability, the projects needing to be optimized are served, the project target needing to be optimized is clear, waste of a large amount of manpower and material resources in the projects needing no optimization is avoided, the labor cost is reduced, and the work efficiency and the service quality are improved.

Based on the above apparatus, further, the apparatus further includes: and the selection module is used for selecting the evaluation item parameters to be optimized according to the quantization probability of each evaluation item parameter.

Further, the number of the equipment units is 5, and the expression of the assignment problem mathematical model is as follows:

wherein, C _ij An evaluation score representing that j question occurs corresponding to the ith equipment unit; the first constraint is: each equipment unit can only correspond to one critical problem; the expression of the first constraint is:

further, the scoring module is specifically configured to assign and score each evaluation item parameter by using a bolda program counting method to obtain a score corresponding to each evaluation item parameter.

In the present specification, the embodiments are described in a progressive manner, and each embodiment focuses on differences from other embodiments, and the same or similar parts between the embodiments are referred to each other. The device disclosed by the embodiment corresponds to the method disclosed by the embodiment, so that the description is simple, and the relevant points can be referred to the method part for description.

Those of skill would further appreciate that the various illustrative components and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both, and that the components and steps of the various examples have been described above generally in terms of their functionality in order to clearly illustrate this interchangeability of hardware and software. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the implementation. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.

The steps of a method or algorithm described in connection with the embodiments disclosed herein may be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. A software module may reside in Random Access Memory (RAM), memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

The method and the device for evaluating the equipment and facility units based on deep learning provided by the invention are described in detail above. The principles and embodiments of the present invention have been described herein using specific examples, which are presented only to assist in understanding the method and its core concepts of the present invention. It should be noted that, for those skilled in the art, it is possible to make various improvements and modifications to the present invention without departing from the principle of the present invention, and those improvements and modifications also fall within the scope of the claims of the present invention.

Claims

1. A deep learning-based equipment facility unit evaluation method is characterized by comprising the following steps:

calculating the quantitative probability of each evaluation item parameter by using a Bayesian inference strategy;

the method further comprises the following steps:

selecting the evaluation item parameters to be optimized according to the quantitative probability of each evaluation item parameter;

the number of the equipment units is 5, and the expression of the assignment problem mathematical model is as follows:

wherein, C _ij Representing the ith device unit pairAn evaluation score for which j questions should appear; the first constraint is: each equipment unit can only correspond to one critical problem; the expression of the first constraint is as follows:

the second constraint is: each important problem can only appear once correspondingly; the expression of the second constraint is as follows:

2. the method of claim 1, wherein each assessment item parameter is assigned a score using a bolda program counting method.

3. An equipment facility assessment apparatus based on deep learning, for implementing the method of any one of claims 1 to 2, comprising:

the scoring module is used for assigning and scoring the evaluation project parameters to obtain a score corresponding to each evaluation project parameter;

the solving module is used for establishing an assignable problem mathematical model for all evaluation project parameters and solving a total evaluation quantitative index function;

4. The apparatus of claim 3, further comprising:

5. The apparatus of claim 4, wherein the number of equipment units is 5, and the expression of the assigned problem mathematical model is:

wherein, C _ij An evaluation score representing that j problems occur corresponding to the ith equipment unit; the first constraint is: each equipment unit can only correspond to one critical problem; the expression of the first constraint is as follows:

6. the apparatus according to claim 5, wherein the scoring module is specifically configured to score each evaluation item parameter by using a bolda program counting method to obtain a score corresponding to each evaluation item parameter.